v alue in health regional issues volume 2 number 2 september

VALUE IN HEALTH REG

ION

AL ISSUES VO

LUME 2 N

UMBER 2 SEPTEM

BER /OCTO

BER 2013 PAG

ES 169–334 ELSEVIER

Volume 2 Number 2 September/October 2013 ISSN 2212-1099

EDITORIAL BOARDCo-Editor-in-Chief (For CEEWAA Issues)

Dan Greenberg, PhDBen-Gurion University of the NegevBeer-Sheva, [email protected]

Co-Editor-in-Chief (For Asia Issues)

Bong-Min Yang, PhDSeoul National UniversitySeoul, [email protected]

Co-Editor-in-Chief (For Latin American Issues)

Federico Augustovski, MD, MSc, PhDInstitute for Clinical Effectiveness and HealthPolicy (IECS); University of Buenos Aires;Hospital Italiano of Buenos AiresBuenos Aires, [email protected]

For complete listing of the Editorial Advisory Board, please visit http://www.ispor.org/publications/VIHRI/Editorial-Advisory-Board.asp

CO-EDITORS

Mary Geitona, BSc, PhDUniversity of the PeloponneseDamaskinou & Kolokotroni, [email protected]

Dominik Golicki, MSc, MD, PhDHealthQuest Warsaw, [email protected]

László Gulácsi, MSc, MD, PhDCorvinus University of BudapestBudapest, [email protected]

Mihajlo B. Jakovljevic, MD, PhDUniversity of KragujevacKragujevac, [email protected]

Zoltán Kaló, MSc, MD, PhDEötvös Loránd University (ELTE) and Syreon

Research InstituteBudapest, [email protected]

Hakan Ergün, MD, PhDAnkara UniversityAnkara, [email protected]

Nathorn Chaiyakunapruk, PharmD, PhDMonash University, Sunway CampusSelangor, [email protected]

Jeff Jianfei Guo, MS, PhDUniversity of Cincinnati Health Academic CenterCincinnati, Ohio, [email protected]

Kenneth KC Lee, JP, MPhil, PhD Reg Pharm (HK)Monash UniversityKuala Lumpur, [email protected]

Asia

J. Jaime Caro, MDCM, FRCPC, FACPUnited BioSourceLexington, MA, [email protected]

Marcos Bosi Ferraz, MD, MSc, PhDFederal University of São PauloSão Paulo, BrazilMarcos.Ferraz@fl eury.com.br

Victor Zarate, MD, MScUniversity of YorkSantiago, [email protected]

Latin America

Imre Boncz, MD, MSc, PhD, HabilUniversity of PécsPécs, [email protected]

Mohamed Izham b. Mohamed Ibrahim, PhDCollege of Pharmacy, Qatar UniversityDoha, [email protected]

Central & Eastern Europe, Western Asia & Africa

Central & Eastern Europe, Western Asia & Africa EDITORIAL ADVISORY BOARD

Jan Busschbach, PhD (Chair)Erasmus UniversityRotterdam, [email protected]

Maarten J. IJzerman, PhD University of Twente Enschede, The Netherlands [email protected]

Donald Patrick, PhD, MSPHUniversity of WashingtonSeattle, WA, [email protected]

Managing EditorStephen L. [email protected]

Editorial AssistantNancy [email protected]

Malgorzata (Gosia) Juszczak-Punwaney, MA [email protected]

MANAGEMENT ADVISORY BOARD EDITORIAL OFFICE

VOLUME 2 NUMBER 2 SEPTEMBER/OCTOBER 2013

TABLE OF CONTENTS

EDITORIAL

169 Further Steps in the Development of Pharmacoeconomics, Outcomes Research, and Health Technology Assessment in Central and Eastern Europe, Western Asia, and Africa

Dan Greenberg , Imre Boncz , Zolt á n Kal ó , and Mohamed Izham B. Mohamed Ibrahim

ECONOMIC EVALUATION

171 Cost-Effectiveness Analysis of Aripiprazole Augmentation Treatment of Patients with Major Depressive Disorder Compared to Olanzapine and Quetiapine Augmentation in Turkey: A Microsimulation Approach

Mete Saylan , M.J. Treur , R. Postema , N. Dilbaz , H. Savas , B.M. Heeg , and P.B. Drost

181 Cost-Utility Analysis of Depot Atypical Antipsychotics for Chronic Schizophrenia in Croatia Vlado Jukic , Miro Jakovljevic , Igor Filipcic , Miroslav Herceg , Ante Silic , Tatjana Tomljanovic , Roman Zilbershtein , Rasmus C.D. Jensen , Michiel E.H. Hemels , and Thomas R. Einarson

189 Cost-Utility Analysis of Pharmaceutical Care Intervention Versus Usual Care in Management of Nigerian Patients with Type 2 Diabetes

Maxwell O. Adibe , Cletus N. Aguwa , and Chinwe V. Ukwe

199 Economic Burden of Cardiovascular Diseases in the Russian Federation Anna Kontsevaya , Anna Kalinina , and Rafael Oganov

205 Cost for Treatment of Chronic Lymphocytic Leukemia in Specialized Institutions of Ukraine Olena Mandrik , Isaac Corro Ramos , Olga Zalis ’ ka , Andriy Gaisenko , and Johan L. Severens

210 Costs of Medically Attended Acute Gastrointestinal Infections: The Polish Prospective Healthcare Utilization Survey Marcin Czech , Magdalena Rosinska , Justyna Rogalska , Ewa Staszewska , and Pawel Stefanoff

218 Radiology Services Costs and Utilization Patterns Estimates in Southeastern Europe — A Retrospective Analysis from Serbia

Mihajlo Jakovljevi c , Ana Rankovi c , Nemanja Ran c i c , Mirjana Jovanovi c , Milo š Ivanovi c , Olgica Gajovi c , and Zorica Lazi c

226 Children Hospitalized for Varicella: Complications and Cost Burden Ozden Turel , Mustafa Bakir , Ismail Gonen , Nevin Hatipoglu , Cigdem Aydogmus , Emine Hosaf , and Rengin Siraneci

PATIENT-REPORTED OUTCOMES

231 The Methodological Challenges for the Estimation of Quality of Life in Children for Use in Economic Evaluation in Low-Income Countries

Travor Mabugu , Paul Revill , and Bernard van den Berg

240 The Impact of Pharmaceutical Care Intervention on the Quality of Life of Nigerian Patients Receiving Treatment for Type 2 Diabetes

Maxwell O. Adibe , Chinwe V. Ukwe , and Cletus N. Aguwa

TABLE OF CONTENTS - continued

248 An Audit of Diabetes-Dependent Quality of Life (ADDQOL) in Older Patients with Diabetes Mellitus Type 2 in Slovenia

Eva Turk , Valentina Prevolnik Rupel , Alojz Tapajner , Stephen Leyshon , and Arja Isola

254 Patient-Reported Quality of Life During Antiretroviral Therapy in a Nigerian Hospital Azuka C. Oparah , Jeffrey S. Soni , Herbert I. Arinze , and Ifeanyi E. Chiazor

CLINICAL OUTCOMES STUDIES

259 Clinical Burden of Invasive Pneumococcal Disease in Selected Developing CountriesNamaitijiang Maimaiti, Zafar Ahmed, Zaleha Md Isa, Hasanain Faisal Ghazi, and Syed Aljunid

HEALTH POLICY ANALYSIS

264 Capacity Building for HTA Implementation in Middle-Income Countries: The Case of HungaryZoltán Kaló, József Bodrogi, Imre Boncz, Csaba Dózsa, Gabriella Jóna, Rita Kövi, Zsolt Pásztélyi, and Balázs Sinkovits, on behalf of ISPOR Hungary Chapter

267 What Infl uences Recommendations Issued by the Agency for Health Technology Assessment in Poland? A Glimpse Into Decision Makers’ Preferences

Maciej Niewada, Małgorzata Polkowska, Michał Jakubczyk, and Dominik Golicki

273 A Framework for Applying Health Technology Assessment in Cyprus: Thoughts, Success Stories, and Recommendations

Panagiotis Petrou and Michalis A. Talias

279 Systematic Review of Economic Evaluation Literature in Ghana: Is Health Technology Assessment the Future?Emmanuel Ankrah Odame

284 Dossier System as a Practical Tool for Compiling Reimbursement ListsMaria V. Sura and Vitaly V. Omelyanovskiy

290 Impact of the Pharma Economic Act on Diffusion of Innovation and Reduction of Costs in the Hungarian Prescription Drug Market (2007–2010)

Rok Hren

300 Performance Assessment of Ga District Mutual Health Insurance Scheme, Greater Accra Region, GhanaEric Nsiah-Boateng and Moses Aikins

306 The Process of Privatization of Health Care Provision in PolandKrzysztof Kaczmarek, Hannah Flynn, Edyta Letka-Paralusz, Krzysztof Krajewski-Siuda, and Christian A. Gericke

312 Transforming Public Servants’ Health Care Organization in Greece through the Implementation of an Electronic Referral Project

Kyriakos Souliotis, Vasiliki Mantzana, and Manto Papageorgiou

CONCEPTUAL PAPER

319 Recommendations for Reporting Pharmacoeconomic Evaluations in EgyptGihan H. Elsisi, Zoltán Kaló, Randa Eldessouki, Mahmoud D. Elmahdawy, Amr Saad, Samah Ragab, Amr M. Elshalakani, and Sherif Abaza

TABLE OF CONTENTS - continued

328 Note from the Editors

329 Erratum

LETTERS TO THE EDITORS

331 Response to “Potential Regulatory and Commercial Environment for Biosimilars in Latin America” by Azevedo et al.

333 Response to Letter from Dr. Jorge Revilla Beltri dated 22nd March, 2013

I Guide for Authors

Editorial Offi ce. Value in Health Regional Issues, ISPOR, 505 Lawrence Square Blvd. South, Lawrenceville, NJ 08648.

ISPOR Offi ce. Marilyn Dix Smith, RPh, PhD, Executive Director, 505 Lawrence Square Blvd. South, Lawrenceville, NJ 08648. Tel: (609) 586-4981, Fax: (609) 586-4982, E-mail: [email protected], Web site: http://www.ispor.org/publications/VIHRI/VIHRImain.asp

Value in Health Regional Issues (ISSN 2212-1099) is published 3 times a year on behalf of the International Society for Pharmacoeconomics and Outcomes Research by Elsevier Inc, 360 Park Avenue South, New York, NY 10010-1710.

POSTMASTER: Send address changes to, Value in Health Regional Issues, Elsevier Health Sciences Division, Subscription Customer Service, 3251 Riverport Lane, Maryland Heights, MO 63043.

CUSTOMER SERVICE (orders, claims, online, change of address): ElsevierHealth Sciences Division, Subscription Customer Service, 3251 Riverport Lane,Maryland Heights, MO 63043. Tel: (800) 654-2452 (U.S. and Canada); (314)447-8871 (outside U.S. and Canada). Fax: (314) 447-8029. E-mail:[email protected] (for print support);[email protected] (for online support). Address changesmust be submitted four weeks in advance.

YEARLY SUBSCRIPTION RATES: All subscribers to the journal Value in Health automatically receive Value in Health Regional Issues. Value in Health Regional Issues is not available for separate subscription purchase. Single issues of Value in Health Regional Issues may be purchased via Elsevier Customer Service, using the contact details above.

Further information on this journal is available from the Publisher or from thisjournal’s website (http://www.elsevier.com/locate/vhri). Information on otherElsevier products is available through Elsevier’s website (http://www.elsevier.com).

Author inquiriesFor inquiries relating to the submission of articles (including electronicsubmission where available), visit http:/www.elsevier.com/authors. The site alsoprovides the facility to track accepted articles and set up e-mail alerts to informyou of when an article’s status has changed, as well as detailed artwork guidelines, copyright information, frequently asked questions, and more. Pleasesee Information for Authors for individual journal. Contact details for questionsarising after acceptance of an article, especially those relating to proofs, areprovided after registration of an article for publication.

English language help service: Upon request, Elsevier will direct authors toan agent who can check and improve the English of their paper (beforesubmission). Please contact [email protected] for further information.

Reprints. For queries about author offprints, e-mail [email protected]. To order 100 or more reprints for educational, commercial, orpromotional use, contact the Commercial Reprints Department, Elsevier Inc.,360 Park Avenue South, New York, NY 10010-1710. Fax: (212) 462-1935; e-mail [email protected]. Reprints of single articles available online may beobtained by purchasing Pay-Per-View access for $36 per article on the journalwebsite http://www.elsevier.com/locate/vhri.

Copyright © 2013, International Society for Pharmacoeconomics and OutcomesResearch (ISPOR). Published by Elsevier Inc.

This journal and the individual contributions contained in it are protected undercopyright by International Society for Pharmacoeconomics and OutcomesResearch, and the following terms and conditions apply to their use:

PhotocopyingSingle photocopies of single articles may be made for personal use as allowedby national copyright laws. Permission of the Publisher and payment of a fee isrequired for all other photocopying, including multiple or systematic copying,copying for advertising or promotional purposes, resale, and all forms ofdocument delivery. Special rates are available for educational institutions thatwish to make photocopies for non-profi t educational classroom use.Permissions may be sought directly from Elsevier’s Rights Department inOxford, UK: phone +44 (0) 1865 843830, fax +44 (0) 1865 853333. Requests may also be completed online via the Elsevier homepage (http://www.elsevier.com/authors/obtaining-permission-to-re-use-elsevier-material).In the USA, users may clear permissions and make payments through theCopyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923,USA; phone: (978) 750-8400, fax: (978) 750-4744, and in the UK through theCopyright Licensing Agency Rapid Clearance Service (CLARCS), 90 TottenhamCourt Road, London W1P 0LP, UK; phone: (+44) 20 7631 5555; fax: (+44) 207631 5500. Other countries may have a local reprographic rights agency forpayments.

Derivative WorksSubscribers may reproduce tables of contents or prepare lists of articlesincluding abstracts for internal circulation within their institutions. Permission ofthe Publisher is required for resale or distribution outside the institution.Permission of the Publisher is required for all other derivative works, includingcompilations and translations.

Electronic Storage or UsagePermission of the Publisher is required to store or use electronically anymaterial contained in this journal, including any article or part of an article.Except as outlined above, no part of this publication may be reproduced, storedin a retrieval system or transmitted in any form or by any means, electronic,mechanical, photocopying, recording or otherwise, without prior writtenpermission of the Publisher.Address permissions requests to: Elsevier Rights Department, at the fax ande-mail addresses noted above.

Abstracting and IndexingValue in Health Regional Issues is indexed in SciVerse Scopus.

NoticeNo responsibility is assumed by the Publisher or the International Society forPharmacoeconomics and Outcomes Research for any injury and/or damage topersons or property as a matter of products liability, negligence or otherwise,or from any use or operation of any methods, products, instructions or ideascontained in the material herein. Because of rapid advances in the medicalsciences, in particular, independent verifi cation of diagnoses and drug dosagesshould be made.Although all advertising material is expected to conform to ethical (medical)standards, inclusion in this publication does not constitute a guarantee orendorsement of the quality or value of such product or of the claims made of itby its manufacturer.

Avai lable onl ine at www.sc iencedirect .com

journal homepage: www.elsevier .com/ locate /vhr i

EDITORIAL

Further Steps in the Development of Pharmacoeconomics, OutcomesResearch, and Health Technology Assessment in Central and EasternEurope, Western Asia, and Africa

During the organizational and functional development of theInternational Society for Pharmacoeconomics and OutcomesResearch (ISPOR), more attention was paid to developing regions.In addition to the traditional annual meetings in Western Europeand North-America, the 1st Asia-Pacific Conference and the 1stLatin America Conference were organized in 2003 and 2007, respec-tively. In addition to Value in Health, its well-established journal,ISPOR introduced Value in Health Regional Issues (ViHRI), its newindependent, official scientific journal in 2012. Its 1st volume,published in 2012, consisted of two issues covering the regions ofAsia [1,2] and Latin America [3,4]. As of 2013, a special issue hasbeen devoted to the regions of Central and Eastern Europe, WesternAsia, and Africa (CEEWAA). Countries eligible from Central andEastern Europe include Albania, Belarus, Bosnia and Herzegovina,Bulgaria, Croatia, Czech Republic, Estonia, Greece, Hungary, Latvia,Lithuania, Montenegro, Poland, Republic of Moldova, Romania,Russian Federation, Serbia, Slovakia, Slovenia, Ukraine, and TheFormer Yugoslav Republic of Macedonia. Countries eligible fromWestern Asia are Armenia, Azerbaijan, Bahrain, Cyprus, Georgia,Iraq, Israel, Jordan, Kuwait, Lebanon, Occupied Palestinian Terri-tory, Oman, Qatar, Saudi Arabia, Syrian Arab Republic, Turkey,United Arab Emirates, and Yemen. All African countries are eligible.

The criteria of an article to be considered in ViHRI CEEWAA areas follow: at least one of the authors of an article submitted to thejournal must reside in the region; when an article reporting on anempirical study is submitted, it must include subjects frompopulation(s) in the region. For the first CEEWAA issue of ViHRI,54 manuscripts were submitted of which 24 have been acceptedfor publication following a thorough review process.

The introduction of a special issue for the CEEWAA region ofISPOR provides a great opportunity for scholars from these coun-tries to publish their research findings and health policy reports inan international scientific journal. Yet, it is very challenging toinclude in one issue articles from a variety of regions. One shouldbear in mind that countries of the CEEWAA region represent aheterogeneous region with substantial differences and challengesrelating to country’s wealth (i.e., gross domestic product percapita), political environment, population’s health status, healthcare affordability, and spending. For example, the life expectancyat birth for males is 72 years for the World Health Organization(WHO) European Region, 67 years for the Eastern MediterraneanRegion, and 55 years for the African Region, while for females, thevalues are 79, 70, and 58 years, respectively. The number ofphysicians per 10,000 populations is 33.3 for the WHO EuropeanRegion, 10.8 for the Eastern Mediterranean Region, and 2.5 for the

African Region. The per-capita total expenditure on health (pur-chasing power parity international $) is 2282 for the WHOEuropean Region, 326 for the Eastern Mediterranean Region, and154 for the African Region [5].

The diversity of challenges in these countries could result in abroader spectrum of published articles. While in some developedcountries, cutting of hospital beds or narrowing the healthinsurance basic benefit package represent key health policyobjectives, other countries try to establish a hospital system orintroduce a basic benefit package for larger parts of their popula-tion. Whatever is a current leading health policy issue in either anationwide or a local health care system, decision making shouldrely on strong scientific evidence. In this decision-making process,health-economics and outcomes research must play an importantrole by informing decision makers on the costs and benefits ofalternative medical interventions. Although publications from theCEEWAA relating to pharmacoeconomics and outcomes researchare limited, current research topics cover health-economic ana-lyses [6,7], coverage policy of new medical technologies [8,9],pharmaceutical market analyses [10,11], burden of disease studies[12], and budget constraints issues [13].

The current issue of ViHRI features articles from 15 differentcountries and include economic analyses and patient-reportedoutcomes on various disease areas such as cardiovascular dis-eases, diabetes, cancer, and psychiatric conditions, as well asclinical outcomes studies and health policy analyses. We hopethat both readers and policymakers will find this issue informa-tive and enriching.

Finally, we thank the ISPOR staff for initiating and supportingthis new journal and encourage scholars from ISPOR’s CEEWAAcountries to submit their research findings to ViHRI.

Imre Boncz, MD, MSc, PhD, HabilInstitute for Health Insurance, Faculty of Health Sciences,

Institute for Health Insurance, University of Pécs, Pécs, Hungary

Zoltán Kaló, MSc, MD, PhD, HabilDepartment of Health Policy and Health, Health Economic Research

Center, Economics Eötvös Loránd University (ELTE), and SyreonResearch Institute, Budapest, Hungary

Mohamed Izham B. Mohamed Ibrahim, PhD,College of Pharmacy, Qatar University, Doha, Qatar

Conflicts of interest: The authors have indicated that they have no conflicts of interest with regard to the content of this article.Source of financial support: The authors have no other financial relationships to disclose.

V A L U E I N H E A L T H R E G I O N A L I S S U E S 2 ( 2 0 1 3 ) 1 6 9 – 1 7 0

dx.doi.org/10.1016/j.vhri.2013.07.007

Dan Greenberg, PhD, MSc

Department of Health Systems Management,Ben-Gurion University of the Negev, Beer-Sheva, Israel

2212-1099/$36.00 – see front matter Copyright & 2013,International Society for Pharmacoeconomics and Outcomes

Research (ISPOR). Published by Elsevier Inc.http://dx.doi.org/10.1016/j.vhri.2013.07.007

R E F E R E N C E S

[1] Khonputsa P, Veerman LJ, Bertram M, et al. Generalized cost-effectiveness analysis of pharmaceutical interventions for primaryprevention of cardiovascular disease in Thailand. Value HealthRegional 2012;1:15–22.

[2] Kim BRM, Lee TJ, Lee HJ, et al. Cost-effectiveness of sertindole amongatypical antipsychotics in the treatment of schizophrenia in SouthKorea. Value Health Regional 2012;1:59–65.

[3] Janusz CB, Jauregui B, Sinha A, et al. Performing country-ledeconomic evaluations to inform immunization policy: ProVac

experiences in Latin America and the Caribbean. Value Health Regional2012;1:248–53.

[4] Augustovski F, Rojas JAD, Ferraz MB, et al. Status update of thereimbursement review environment in the public sector across fourLatin American countries. Value Health Regional 2012;1:223–7.

[5] World Health Organization. World Health Statistics 2013. Geneva,Switzerland: World Health Organization, 2013.

[6] Vokó Z, Nagyjánosi L, Margitai B, et al. Modeling cost-effectiveness ofcervical cancer screening in Hungary. Value Health 2012;15:39–45.

[7] Shmueli A, Fraifeld S, Peretz T, et al. Cost effectiveness of baseline low-dose CT screening for lung cancer: the Israeli experience. Value Health2013;16:922–31.

[8] Inotai A, Pékli M, Jóna G, et al. Attempt to increase the transparency offourth hurdle implementation in Central-Eastern European middleincome countries: publication of the critical appraisal methodology.BMC Health Serv Res 2012;12:332.

[9] Abuzar A, Yassir AH, Mohamed Izham MI. Do Saudi communitypharmacists know how to use MDIs? J Pharm Prac Res 2012;42:77.

[10] Greenberg D, Siebzehner M, Pliskin JS. The process of updating theNational List of Health Services in Israel: is it legitimate? is it fair? Int JTechnol Assess Health Care 2009;25:255–71.

[11] Abdulkareem MAS, Mohamed Izham MI, Ahmed AR. The qualityof prescriptions with antibiotics in Yemen. J Clin Diagnostic Res2011;5:808–12.

[12] Boncz I, Brodszky V, Péntek M, et al. The disease burden of colorectalcancer in Hungary. Eur J Health Econ 2010;10(S1):S35–40.

[13] Boncz I, Sebestyen A. Financial deficits in the health services of the UKand Hungary. Lancet 2006;368:917–8.


dx.doi.org/10.1016/j.vhri.2013.07.007

dx.doi.org/10.1016/j.vhri.2013.07.007

dx.doi.org/10.1016/j.vhri.2013.07.007

http://refhub.elsevier.com/S2212-1099(13)00086-1/sbref1







http://refhub.elsevier.com/S2212-1099(13)00086-1/othref0005
































ECONOMIC EVALUATION

Cost-Effectiveness Analysis of Aripiprazole Augmentation Treatment ofPatients with Major Depressive Disorder Compared to Olanzapine andQuetiapine Augmentation in Turkey: A Microsimulation ApproachMete Saylan, MD1,�, M.J. Treur2, R. Postema2, N. Dilbaz3, H. Savas4, B.M. Heeg2, P.B. Drost5

1Market Access Department, Bristol Myers Squibb, Istanbul, Turkey; 2Pharmerit, Rotterdam, The Netherlands; 3Psychiatry Department, Ankara Numune Researchand Training Hospital, Ankara, Turkey; 4Psychiatry Department, Gaziantep University Medical School, Gaziantep, Turkey; 5Bristol Myers Squibb, Paris, France

A B S T R A C T

Objectives: Major depressive disorder (MDD) is a chronic illnessassociated with a major burden on quality of life (QOL) and healthcare resources. Aripiprazole augmentation to antidepressant treat-ment was recently approved for patients with MDD respondinginsufficiently to antidepressant treatment in Turkey. The objectivewas to estimate the cost-effectiveness of aripiprazole augmentationin this indication compared with olanzapine and quetiapine augmen-tation from a payer perspective. Methods: A lifetime economic modelwas built simulating transitions of patients with MDD between majordepressive episodes (MDEs) and remission. During MDEs, patientswere treated with adjunctive aripiprazole, quetiapine, or olanzapine.Patients who did not respond switched to subsequent treatment lines.Comparative effectiveness between adjunctive aripiprazole, quetia-pine, and olanzapine was estimated by using an indirect comparison.Resource utilization and costs were obtained from Turkish studies.Results: Over a lifetime horizon, patients treated with aripiprazolespent less time in MDEs than did patients treated with quetiapine (−11weeks) and olanzapine (−7 weeks). On average, patients treated with

aripiprazole showed improvement in QOL compared with patientstreated with quetiapine (þ0.054 quality-adjusted life-years [QALYs])and olanzapine (þ0.039 QALYs) combined with cost saving of 593Turkish lira (TL) versus quetiapine and 485 TL versus olanzapine. Theprobability that adjunctive aripiprazole would be cost-effective amongthe three strategies ranged between 74% and 75% for willingness-to-pay values between 0 TL and 100,000 TL per QALY gained. Conclu-sions: This is the first lifetime health-economic model in Turkey thattakes patient heterogeneity into account when assessing QOL andcosts of different adjunctive strategies in MDD. The results indicatethat adjunctive treatment with aripiprazole provides health benefitsat lower costs in patients with MDD when compared with quetiapineand olanzapine augmentation.Keywords: antipsychotics, aripiprazole, cost-effectiveness analysis,depression, discrete probability distribution, major depressivedisorder, olanzapine, quetiapine, simulation model, Turkey.

Copyright & 2013, International Society for Pharmacoeconomics andOutcomes Research (ISPOR). Published by Elsevier Inc.

Introduction

Mood disorders represent a major health problem. Depression is afrequent and severe illness with a substantial impact on personaland familial suffering. Several surveys such as the NationalComorbidity Survey Replication in the United States have showna lifetime prevalence of mood disorders of more than 20% inadults [1]. Most of this prevalence was associated with majordepression, which had a lifetime prevalence of 16.6%. In theWorldHealth Organization’s World Mental Health Survey Initiative, theprojected lifetime prevalence of any mood disorder was 31.4% inthe United States [2]. In the European Study of the Epidemiology ofMental Disorders, 13% of the individuals reported a history ofmajor depression, with a 12-month prevalence of 4% [3]. InTurkey, the prevalence of depression was estimated to be 21% in2004 [4]. Depression is a highly recurrent disease; 80% of the

patients with a history of two episodes will have another recur-rence during their lifetime [5]. Because of the high risk of suicide(6.3% annually [6]), depression can be a life-threatening illness.

According to the World Health Organization, major depressionis currently ranked as the leading cause of disability in middle-and high-income countries. At an international level, 4.1% of thetotal global burden of disease is due to major depression [7].Depression, being an important source of impaired health-relatedquality of life (HRQOL) of patients [8,9], was also the fourthleading cause of disease burden in Turkey [4]. Depression pri-marily impacts the usual activities, pain and discomfort, andanxiety and depression domains on the EuroQol five-dimensionalquestionnaire [10]. Reported utility values for depressive episodeswere between 0.09 and 0.47 [10–14]. Total cost for depression wasestimated at $267 million in Turkey in 2004, primarily related tohospital-based treatment (93%) [15].

2212-1099/$36.00 – see front matter Copyright & 2013, International Society for Pharmacoeconomics and Outcomes Research (ISPOR).

Published by Elsevier Inc.

http://dx.doi.org/10.1016/j.vhri.2013.06.004

Conflicts of interest: The authors have indicated that they have no conflicts of interest with regard to the content of this article.

E-mail: [email protected].

�Address correspondence to: Mete Saylan, Market Access Department, Bristol Myers Squibb, Istanbul, Turkey.


dx.doi.org/10.1016/j.vhri.2013.06.004

dx.doi.org/10.1016/j.vhri.2013.06.004

dx.doi.org/10.1016/j.vhri.2013.06.004

mailto:[email protected]

dx.doi.org/10.1016/j.vhri.2013.06.004

Today, the ultimate goal in the treatment of major depressionis remission, that is, a full symptomatic recovery with a return topremorbid functioning. Indeed, partial remission is associatedwith a greater risk of relapse and recurrence, decreased quality oflife, a poorer psychosocial functioning, a higher mortality risk,and increased cost of illness. A Swedish study has shown thatpatients who are not in remission use 1.6 times more medicalresources than do those in remission [16].

In the Sequenced Treatment Alternatives to Relieve Depres-sion (STAR*D) study, less than 30% of the patients reachedremission with first-step antidepressant treatment within 14weeks of starting treatment [17,18]. Another recent study per-formed in primary care also reported very low remission rateswith antidepressant treatment: 28.3% according to the cliniciansand 17.1% according to the patients [19]. For these insufficientresponders to antidepressant treatment, one may considerincreasing the dose or switching to another antidepressant,depending on the level of initial response. Alternatively, thetreatment of patients with an insufficient response to an anti-depressant may be augmented with an atypical antipsychotic.Turkey was the first country in Europe to approve aripiprazoleaugmentation for the treatment of major depressive episodes inpatients who showed inadequate response after at least oneantidepressant treatment [20]. For reimbursement decisions, it isimportant to consider the value for money of this strategycompared with other alternatives. Quetiapine augmentation isalso approved for this indication in Turkey [20], and olanzapineaugmentation is used off-label (it is not officially approved inTurkey but has a US license as combination with fluoxetine[http://www.accessdata.fda.gov/drugsatfda_docs/label/2011/020592s060s061,021086s038s039lbl.pdf]). The short-term use of theseregimens has been compared in a recent cost-effectivenessanalysis in the United States [21]. A Turkish cost-effectivenessassessment, however, is still missing. As such, this article aimedto assess the cost-effectiveness of aripiprazole augmentationcompared with that of quetiapine and olanzapine augmentationfor the treatment of major depressive disorder (MDD) in Turkeyfrom a payer perspective.

Methods

Model Structure

A patient-level simulation model was built structuring the evi-dence on clinical and economic outcomes of treating patientssuffering from MDD with adjunctive aripiprazole compared withadjunctive quetiapine and adjunctive olanzapine. The model wasbuilt in Microsoft Excel and Visual Basic for applications. A totalof 50,000 patients were simulated to reach stable results.A microsimulation approach was deemed most appropriate inthis indication, due to the heterogeneity of the patient populationand the strong association between a patient’s history and his orher future disease course. To represent this with a Markov modelwould require too many health states. A schematic overview ofthe simulation model structure is presented in Figure 1, repre-senting the modeled health states and possible transitions. Thedepressive episode is the initial health state of a patient. Durationwas simulated to determine the time at which a patient would

move to the remission state. Once there, the time until a nextdepressive episode was simulated, specifying the length of stayin remission. If that period was longer than 9 months, a patientspent the remaining time in the “between episodes” state,incurred fewer costs and experienced further quality of lifeimprovements. Back in the depressive episode state, the proce-dure was repeated until a patient died. Time of death wassimulated at model entrance (based on age and gender) andcould be shortened if a patient committed suicide, which waspossible only during a depressive episode. During each depres-sive episode it was simulated whether a patient had committedsuicide. It was assumed that this would take place in the middleof the episode. Further model details are provided in the follow-ing sections.

Patient Population Simulated

The characteristics of the patients that were simulated at modelentrance resemble the populations enrolled in the double-blindrandomization phases of the three clinical trials assessing theefficacy of aripiprazole augmentation [22–24]. The patients inthese trials suffered from a major depressive episode and had aninsufficient response to at least two prior antidepressant thera-pies prior to trial entry. Their characteristics and the distributionsused for simulating them in the model are provided in Table 1.

Clinical Data

The time a patient spent in the depressive episode statedepended on the remission rate of the therapy. Remission rateswith aripiprazole augmentation were based on the three clinicaltrials assessing the efficacy of aripiprazole as adjunctive therapyin MDD [22–24]. During a 6-week treatment period, 28.8% of thepatients reached remission (see Table 2). A Bernoulli distributionwith a probability of 0.288 was used in the model to simulatewhether a patient would respond to aripiprazole augmentationwithin 6 weeks. This discrete probability distribution takes avalue of 1 (response) with a probability of 28.8% and a value of0 (no response) with a probability of 71.2%. A remitting patientwould move to the remission state after 6 weeks. Patients notreaching remission after 6 weeks remained in the depressivestate and were switched to a subsequent treatment line (seeFig. 2). Comparative 6-week remission rates of the other adjunc-tive strategies were based on a formal indirect comparison due toa lack of direct comparable data in this indication. To estimatethe efficacy of other adjunctive strategies, a systematic reviewwas conducted identifying head-to-head or placebo controlledstudies (PCSs) of antidepressant augmentation with aripiprazole,

Depressiveepisode

Death

BetweenEpisodes

Remission

Fig. 1 – Schematic model representation.

Table 1 – Baseline patient characteristics [13–15] and corresponding distributions used for simulating.

Characteristic Mean � SD Distribution Parameter (s)

Age (y) 45.1 � 4.4 Normal m ¼ 45.1, s ¼ 4.4Gender (% males) 68.0 Bernoulli P ¼ 0.68Number of prior episodes 6 � 5.2 Geometric P ¼ 0.17


quetiapine, olanzapine, lithium, and triiodothyronine in atreatment-resistant depression adult population. A literaturesearch was conducted in MEDLINE covering studies publishedbetween January 1980 and June 2010. Three PCSs for aripiprazole[22–24], two PCSs for adjunctive quetiapine [25,26], three PCSs foradjunctive olanzapine [27–29], one PCS for lithium [30], and onePCS for T3 [31] were identified. Remission rates of the activetreatments and the control group were extracted. Study charac-teristics and remission numbers are summarized in Table 3.Subsequently, a fixed effects Bayesian meta-analysis using non-informative priors implemented in WinBUGS [32] was conductedon the remission rates to obtain indirect odds ratios of eachtreatment compared with aripiprazole [33]. The indirect oddsratios were applied to the aripiprazole remission rate to obtainremission rates for the other treatments (Table 2). Again, aBernoulli distribution was used to simulate whether a patientwould respond to a subsequent treatment line.

A patient failing three consecutive adjunctive therapies wasassumed to receive so-called best supportive care (BSC). In reality,

at this stage, medical professionals will try different therapiesincluding different treatment combinations. It is difficult toformally implement all these different treatment options in ahealth economic model structure because of the complexity ofdecisions and lack of published data to substantiate correspond-ing efficacy. Therefore, BSC grouped all these treatment alter-natives into one for which efficacy was based on the STAR*D trial,thereby mimicking real-life practice [34]. It was anticipated thatthis simplification would not affect incremental results. STAR*Dfound that 13% of the patients responded to treatment within 9.2weeks. This was implemented in the model by drawing aduration on BSC from an exponential distribution with a weeklyhazard rate of 0.015 (− ln(1 − 0.13)/9.2).

A patient who developed a new major depressive episodereceived the same treatment as he or she had previouslyresponded to. No efficacy data are available for patients with aprior response to adjunctive treatment. However, it is highlylikely that this subgroup of patients will have an increasedprobability of remission, reflected by assuming a 6-week proba-bility of remission of 90%. This value was applied to all therapiesexcept for BSC, for which the duration was drawn from the sameexponential distribution as for the first treatment episode. Theimpact of the remission probability value for prior responderswas tested in a sensitivity analysis.

The assumed risk of developing a new major depressiveepisode reflected the naturalistic data observed by Solomonet al. [35]. They followed 318 patients suffering from MDD overa period of 10 years and found a decrease in mean time until thenext major depressive episode with an increase in the amount ofprior episodes. Weibull curves were fitted on their data (seeFig. 3), and time until the next major depressive episode wassimulated from the appropriate distribution reflecting the num-ber of prior episodes a patient had experienced. Weibull curveswere selected on the basis of a least square difference between fitand actual data. Patients in the remission and time betweenepisodes health states continued their antidepressant butstopped their augmentation strategy. To date, there is no evi-dence surrounding the impact of long-term atypical augmenta-tion on the risk of developing a depressive episode.

Table 2 – Health economic model input parameters.�

Remission probability (%) at 6 wkAripiprazole augmentation 28.8 (25.0–33.0)Quetiapine augmentation 24.5 (16.7–32.2)Olanzapine fluoxetine combination 25.4 (18.2–33.2)Lithium augmentation 16.0 (1.00–44.6)Triiodothyronine augmentation 67.9 (14.5–89.8)Best supportive care (yearly hazard) 0.78

Remission probability prior responders 90.0 (50.0–99.0)

Depressive episode Remission Between episodesUtility values 0.46 (0.34–0.58) 0.81 (0.76– 0.86) 0.86 (0.84–0.88)

Costs ($/TL�)Aripiprazole augmentation 221 (108–327) 119 (48–186) 4Quetiapine augmentation 200 (87–306) 119 (48–186) 4Olanzapine fluoxetine combination 210 (96, 316) 119 (48–186) 4Lithium augmentation 188 (74–294) 119 (48–186) 4Triiodothyronine augmentation 187 (74–293) 119 (48–186) 4Best supportive care 187 (61–307) 119 (48–186) 4

Suicide attempt cost 1269 (801–1763)

Note. 95% Confidence intervals reported between parentheses. All parameters showing confidence intervals in this table were varied in theprobabilistic sensitivity analysis.TL, Turkish lira.�Please note that the purchasing power parity between Turkey and the United States is 1.0 (2010 figure, http://stats.oecd.org).

AripiprazoleAugmenta�on

Que�apineAugmenta�on

OlanzapineAugmenta�on

Lithium Augmenta�on

T3Augmenta�on

Best Suppor�veCare

Strategy 1 Strategy 2 Strategy 3

Fig. 2 – Treatment sequence strategies. T3, triiodothyronine.

V A L U E I N H E A L T H R E G I O N A L I S S U E S 2 ( 2 0 1 3 ) 1 7 1 – 1 8 0 173

http://www.nice.org.uk/media/B52/A7/TAMethodsGuideUpdatedJune2008.pdf

Table 3 – Characteristics of studies used in indirect comparison.

Study Dur. (wk) Country Participants Methods Remission definition Placebo Active arm

N n N n

Aripiprazole Berman et al. [22] 6 US Age: 18–65 y; MDE accordingto DSM-IV-TR; HAM-D-17score ≥ 18 at end of thescreening phase; 1–3historical ADTs of 4 6 wk

7–28-d screening phase; 8-wkprospective treatment phasewith open-label ADT plussingle-blind adjunctiveplacebo. Patients withincomplete responsecontinued ADT and enteredthe randomization phase.

≥ 50% decrease inMADRS score þMADRS score ≤ 10

172 27 181 47

Marcus et al. [24] 6 US Same as above Same as above Same as above 184 28 185 47Berman et al. [23] 6 US Same as above Same as above Same as above 169 32 174 64

Quetiapine McIntyre et al. [26] 8 CA Age: 18–65 y; MDE accordingto DSM-IV; HAM-D-17 score≥ 18. One historical ADT of≥ 6 wk.

Patients treated for currentepisode with single AD attherapeutic dose for ≥ 6 wkand meeting study criteriafor residual depressive andcomorbid anxiety symptomswere randomized.

HAM-D-17 score ≤ 7 29 5 29 9

Bauer et al. [25] 6 AU, CA,EU,andZA

Outpatients; age: 18–65 y;MDD according to DSM-IV-TR; HAM-D-17 score≥ 20; HAM-D item 1 ≥ 2.One historical ADT of≥ 6 wk.

Eligible patients with aninadequate response to anADT during their currentepisode were randomized to6- wk double-blindquetiapine extended releaseor placebo adjunctive toongoing ADT.

MADRS score ≤ 10 160 50 327 134

Olanzapine Shelton et al. [27] 8 US andCA

Age: 18–65 y; DSM-IVunipolar, nonpsychoticMDD; MADRS score ≥ 20 atthe beginning and the endof the screening phase.One historical ADT (SSRI ≥4 wk at a therapeutic dose).

2–7-d screening/washoutphase; 7-wk lead-in phase ofnortriptyline to demonstratetreatment failure to a TCA, 8-wk randomized, double-blind phase: olanzapine/fluoxetine combination,olanzapine, fluoxetine, ornortriptyline

2 consecutive MADRStotal scores of ≤ 8

142 19 146 25

VALUE

INH

EALTH

REGIO

NAL

ISSUES

2(2

013)171–180

174

Corya et al. [29] 12 Various Age: ≥ 18 y. DSM-IV MDD,single or recurrent,unipolar, withoutpsychotic features; CGI-Sscore ≥ 4. One historicalADT (SSRI ≥ 6 wk at atherapeutic dose)

2–7-d screening phase; 7-wkopen-label venlafaxine lead-in phase; patients witho 30% improvement onMADRS score proceeded to5–9-d double-blind taperphase before therandomization phase.

2 consecutive MADRStotal scores of ≤ 8

56 10 230 69

Thase et al. [28]� 8 US andCA

Age: 18–65 y; DSM-IV MDD,recurrent, withoutpsychotic features; HAM-D-17 score ≥ 22. Onehistorical ADT ≥ 6 wk

3–14-d screening phase; 8-wkopen-label lead-in phase toestablish fluoxetineresistance; patients witho 25% decrease in HAM-D-17 score and HAM-D-17 score≥ 18 and ≤ 15% decreasebetween week 7 and 8 oflead-in entered

MADRS score ≤ 10 atend point

102 18 101 24

Thase et al. [28]� 8 US andCA

Same as above Same as above Same as above 101 16 97 30

Lithium Nierenberg et al.[30]

6 US Age: 18–70 y. DSM-III-R MDD,HAM-D-17 score ≥ 18; 1–5historical adequate ADcourses. One prospectiveADT

1–5 adequate AD coursesfailed; 6-wk prospectiveopen-label nortriptyline

≥ 50% decrease inHAM-D-17

17 3 18 2

T3 Joffe et al. [31] 2 CA Mean age 37.4 y; RDCunipolar, nonpsychoticMD; HAM-D score ≥ 16after 5 wk of desipramineor imipramine

Subjects were randomlyassigned to receive 2-wkliothyronine, lithium, orplacebo in addition todesipramine or imipramine

≥ 50% decrease andHAM-D score ≤ 7

16 2 17 7

AD, antidepressant; ADT, antidepressant trial/therapy; AU, Australia; CA, Canada; CGI-S, Clinical Global Impression – Severity; Dur., duration; DSM-III-R, Diagnostic and Statistical Manual of MentalDisorders, Revised Third Edition; DSM-IV, Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition; DSM-IV-TR, Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, Text Revised;EU, Europe; HAM-D-17, Hamilton rating scale for depression; MADRS, Montgomery Åsberg Depression Rating Scale; MD, major depression; MDD, major depressive disorder; MDE, majordepressive episode; N, patient population size; n, number of remitted patients; RDC, Research Diagnostic Criteria; SSRI, selective serotonin reuptake inhibitor; TCA, tricyclic antidepressant, US,United States; ZA, South Africa.�Two identical studies reported in one article.

VALUE

INH

EALTH

REGIO

NAL

ISSUES

2(2

013)171–180

175

Time of death was simulated at model entry on the basis ofgeneral mortality statistics in Turkey taken from the WorldHealth Organization (http://apps.who.int/whosis/database/life_tables/life_tables.cfm, accessed April 2010). The risk of suicideduring major depressive episodes was based on information fromBernal et al. [36] who found a 3.91-fold (95% confidence interval2.74–5.6) higher suicide risk for patients with MDD relative to thegeneral population. According to Devrimci-Ozguven and Sayil[37], suicide risk in the general Turkish population was 112.11 per100,000 inhabitants in 2001. This gives an estimated suicideattempt probability (rate) of 0.4% (0.044) per year for patientswith MDD in Turkey. A constant hazard rate was assumed fordetermining the probability of attempting suicide, based on thepatient’s actual depressive episode duration. The model imputedan 8.3% probability that an attempted suicide resulted in death,which was based on Bilici et al. [38] who found 2.6 completedsuicides out of 31.5 attempts per 100,000 person-years.

Utility Weights

No specific utility values for Turkish patients with MDD wereavailable at the time of this research. A Swedish study by Sobockiet al. [10,16], however, reported utility values for patients withmoderate depression (Clinical Global Impression – Severity scoreof 4) and remission, based on a naturalistic longitudinal obser-vational study of 447 patients in primary care. Utilities werederived from patients’ EuroQol five-dimensional questionnairehealth status questionnaires, applying UK national tariff in theabsence of specific social tariffs for Sweden at the time of theirstudy [10]. Utility values were applied to the depressive episodeand remission health states in the model. It was assumed thatthe quality of life of patients in the between episodes health stateresembled the quality of life in the general population [39] (seealso Table 2).

Cost Data

Drug prices were extracted from the Turkish Ministry of Health,Directorate of Pharmaceuticals Official Web Page on February 2,2010 [40], and adjusted for mandatory discounts to obtain

reimbursed prices. It was assumed that all patients would incurthe same background antidepressant cost during the entiremodel horizon. Background antidepressant costs were calculatedas a weighted average cost of the antidepressants used atrandomization in the clinical trials of aripiprazole augmentation[22–24]. Augmentation treatment costs were incurred duringdepressive episodes. Health care resource use during a depressiveepisode is generally higher than during remission periods.Turkish-specific data about resource use is available only forpatients with MDD in general without stratification between bothstates [41]. A Swedish study, however, has shown that patients ina depressive episode and a remission episode have 1.24 and0.8 times the medical resource consumption of an averagepatient with MDD, respectively [16]. The number of psychiatristvisits and the average number of hospitalization days wereobtained from Karamustafalioglu et al. [41] and multiplied withcorresponding Turkish unit costs [42] to obtain health careresource costs for an average patient with MDD. The ratios fromSobocki et al. [16] were multiplied with this number to obtaindepressive and remission cost estimates for Turkey. Costs duringthe between episodes state were assumed to reflect only theantidepressant use. The cost of suicide attempts was based onKaramustafalioglu et al. [41], including only the costs of carereceived. Total weekly costs per health state and adjunctivetreatment are outlined in Table 2. Please note that remissioncost and between episode cost are identical for all treatmentsbecause there is no differentiation in treatment costs opposed tothe depressive episode phase. Costs and quality-adjusted life-years (QALYs) were discounted with 3.5% per annum.

Probabilistic Sensitivity Analysis

As with all health-economic models, the input parameters aresubject to uncertainty. To describe the influence of the uncer-tainty in model parameters on the incremental model outcomes,a probabilistic sensitivity analysis (PSA) was conducted. The jointuncertainty surrounding incremental costs and effects followingfrom the uncertainty around all model input parameters wasaddressed by generating 1000 random sets of input parameters,using probability distributions for each parameter that reflect its

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 5

Cum

ula�

ve p

erce

ntag

e of

pa�

ents

feer

of n

ew d

epre

ssiv

e ep

isod

e

Time since end of previous depressive episode (years)

1 prior episode2 prior episodes3 prior episodes4 prior episodes5 prior episodes

Fig. 3 – Time until next episode per number of prior episodes. Data points are from Solomon et al. [35]. Lines represent Weibulsurvival curve fits.


uncertainty. Model outputs were subsequently recorded. The 95%confidence intervals for each parameter are presented in Table 2.A major source of the uncertainty in incremental model outcomesis the uncertainty surrounding the difference in remission ratesbetween adjunctive aripiprazole and adjunctive quetiapine andolanzapine as represented by their respective odds ratios. Theuncertainty surrounding these inputs was based on the jointposterior distribution obtained from the indirect comparison.The corresponding uncertainty surrounding the remission ratesof adjunctive quetapine and olanzapine is presented in Table 2.Utility values were varied by using beta –distributions, and thecosts inputs were varied by using gamma distributions. A scatterplot was generated, showing the 1000 combinations of incremen-tal costs and effects generated from the PSA. Corresponding cost-effectiveness acceptability curves were drawn, showing the prob-ability that adjunctive aripiprazole augmentation is cost-effectivecompared with adjunctive quetiapine and olanzapine at variouslevels of willingness to pay (WTP) per QALY gained.

Results

Base Case

Results for the comparison of adjunctive aripiprazole with adjunc-tive quetiapine and adjunctive olanzapine are presented in Table 4.Costs are presented in Turkish lira (TL), which has a purchasingpower parity of 1.0 compared with the US dollar (2010 figure, http://stats.oecd.org). The average life expectancy of a patient in themodelis 31.6 years, with an average starting age of 45 years, implying thatpatients, on average, live to be 76 years old. The minor difference in

life expectancy between the different treatment sequences isexplained by a difference in completed suicides. Aripiprazoleaugmentation has fewer attempted and thereby fewer committedsuicides compared with adjunctive quetiapine and olanzapine. Thisdifference is explained by the lesser time spent in major depressiveepisodes with aripiprazole, during which patients are exposed to therisk of suicide (11 and 7 weeks less compared with quetiapine andolanzapine augmentation). Because of a higher remission rate withadjunctive aripiprazole, on average, patients treated with adjunctivearipiprazole spend less time in the depression state than do thosetreated with adjunctive quetiapine and olanzapine. Because thedepressive episode is associated with diminished quality of life,patients starting with aripiprazole augmentation gain 0.054 and0.039 QALYs on average than do those starting with quetiapineaugmentation and olanzapine augmentation, respectively.

A major cost driver in the model is hospitalization costs, whichcomprise approximately 71% of total costs. Because patients in thedepressive episode are more likely to be hospitalized and patientsin the aripiprazole augmentation arm spend less time in thatepisode, hospitalization costs are saved compared with quetiapineaugmentation (712 TL) and olanzapine augmentation (554 TL). Asimilar pattern is observed for psychiatrist visit costs (savings of240 TL and 174 TL, respectively). Augmentation costs for thearipiprazole augmentation arm are higher than for quetiapineand olanzapine augmentation arms due to the higher drug acquis-ition costs for aripiprazole. Because of the savings on hospital-ization and psychiatrist visit costs, however, aripiprazoleaugmentation saves, on average, 593 TL and 485 TL per patientthan do quetiapine and olanzapine augmentation, respectively.

Dominance of aripiprazole augmentation was independent ofthe remission probability value for prior responders.

Table 4 – Health economic model results.

Outcome Aripiprazoleaugmentation

Quetiapineaugmentation

Olanzapine-augmentation

Differencearipiprazole withquetiapine

Differencearipiprazole witholanzapine

Life expectancy (y) 31.62 31.61 31.62 0.012 0.010Time spent in

depressiveepisodes (y)

9.44 9.64 9.58 −0.21 −0.14

Time spent inremission (y)

8.88 8.81 8.83 0.08 0.06

Time spent betweenepisodes (y)

13.31 13.17 13.22 0.14 0.09

Percentage ofpatientsattempting suicide

4.30 4.30 4.33 −0.01 −0.04

Percentage ofpatientscompleting suicide

0.30 0.33 0.34 −0.03 −0.04

QALYs (discounted) 13.62 13.56 13.58 0.054 (−0.038 to 0.213) 0.039 (−0.048 to 0.171)

Total costs ($/TL)� 84,800 85,393 85,285 −593 (−3780 to 619) −485 (−3132 to 757)Antidepressant 3,840 3,839 3,841 1 1Augmentation 580 222 367 359 212Psychiatrist visits 20,265 20,504 20,439 −240 −174Hospitalization 60,086 60,798 60,604 −712 −554Suicide 29 29 29 1 1

QALY, quality-adjusted life-years; TL, Turkish lira.�Please note that the purchasing power parity between Turkey and the United States is 1.0 (2010 figure, http://stats.oecd.org). Numbers in thetable may not add up because of rounding errors.


Probabilistic Sensitivity Analysis

The scatter plot illustrated in Figure 4 represents the jointuncertainty surrounding incremental QALYs and incrementalcosts for both comparisons, based on 1000 Monte Carlosimulations. This is a result of the combined uncertainty sur-rounding all model parameters, including efficacy, quality oflife, and cost inputs. Scatter points in the lower right quad-rant represent instances in which aripiprazole augmentationimproves quality of life and saves costs. This occurs in 85% and86% of the cases for the comparison with quetiapine andolanzapine, respectively. The top left quadrant represents instan-ces in which aripiprazole augmentation decreases qualityof life and is more expensive. This occurs in 13% of the cases

for both comparisons. The uncertainty surrounding results ismainly driven by the uncertainty around the remissionprobabilities for aripiprazole, quetiapine, and olanzapineaugmentation.

Figure 5 presents the cost-effectiveness acceptability curvesfor the three treatment strategies showing the probability thateach of them is cost-effective at various WTP thresholds. If one isinterested only in health gains at no additional costs, aripiprazoleaugmentation has a probability of being cost-effective of 74%,whereas quetiapine and olanzapine each have a probability of13% to be cost-effective. The acceptability lines remain relativelystable at various WTP levels. This is because only 2% of thescatter points for both comparisons are in the northeast andsouthwest quadrants, where an actual trade-off is made between

-$ \ TL 8,000

-$ \ TL 6,000

-$ \ TL 4,000

-$ \ TL 2,000

$ \ TL 0

$ \ TL 2,000

$ \ TL 4,000

-0.30 -0.20 -0.10 0.00 0.10 0.20 0.30 0.40 0.50

Incr

emen

tal C

osts

Incremental QALYs

Comparison Aripiprazole - Que�apine

Comparison Aripiprazole - Olanzapine

13%13%

1%1%

1%1%

85%86%

Fig. 4 – Scatter plot of comparison adjunctive aripiprazole with adjunctive quetiapine and adjunctive olanzapine.QALY, quality-adjusted life-year; TL, Turkish lira.

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

$ \ TL 0 $ \ TL 20,000 $ \ TL 40,000 $ \ TL 60,000 $ \ TL 80,000 $ \ TL 100,000

Prob

abili

ty o

f bei

ng c

ost-

effec

�ve

Willingness to Pay per QALY gained

Aripiprazole

Que�apine

Olanzapine

Fig. 5 – Cost-effectiveness acceptability curves. QALY, quality-adjusted life-year; TL, Turkish lira.


incremental QALYs and incremental costs. As such, at WTPvalues ranging from 0 to 100,000 TL per QALY gained, theprobability that aripiprazole is cost-effective among all threestrategies varies between 74% and 75%.

Discussion

The objective of this research was to assess the cost-effectivenessof adjunctive aripiprazole treatment compared with adjunctivequetiapine and olanzapine treatment in patients with MDD whohad an insufficient response to antidepressant treatment inTurkey.

A patient-level simulation model was developed simulatingthe chronic and deteriorating course of the disease, as well as thesequence of treatment steps.

The model results showed that adjunctive aripiprazole isdominant compared with quetiapine and olanzapine. Patientsstarting with adjunctive aripiprazole spend 11 weeks and 7 weeksless in major depressive episodes than do patients starting withquetiapine and olanzapine, respectively. This translates to 0.054and 0.039 QALY gains, respectively. Despite the higher drugacquisition cost, on average, the total cost for a patient startingwith adjunctive aripiprazole is 593 TL and 485 TL less than thosefor a patient starting with adjunctive quetiapine and olanzapine,respectively. These savings are mainly explained by less hospital-ization costs and fewer psychiatrist visits. The increased remis-sion rate of adjunctive aripiprazole compared with quetiapineand olanzapine underlies the health gains and cost savings. Thedifference in remission rates between adjunctive aripiprazole,quetiapine, and olanzapine were obtained from an indirectcomparison, combining the available evidence reported in thepublished literature by using 10 identified PCSs [22–31]. Theinfluence of the uncertainty in the estimated differences inremission rates between the three adjunctive atypical antipsy-chotic treatments resulting from this indirect comparison wasincorporated in a PSA. In addition, the uncertainty surroundingthe other model inputs was incorporated in the PSA. Theprobability that adjunctive aripiprazole would be cost-effectiveamong the three strategies ranged between 74% and 75% for WTPvalues between 0 TL and 100,000 TL per QALY gained.

To ensure that the model resembled Turkish clinical practice,resource use, unit cost, and mortality data were based on Turkishsources. To the best of our knowledge, no lifetime cost-effectiveness models in this indication have been published.Previously published cost-effectiveness models in MDD are lim-ited. Only two lifetime horizon models have been published sofar, both by Revicki et al. in 1995 and 1997 [43,44]. Economicguidelines by the National Institute for Health and ClinicalExcellence [45] and Drummond and Jefferson [46] recommend alifetime model horizon in the case of chronic illnesses such asdepression. Also, the deteriorating course of depression shouldideally be captured in the model. This is, however, possible onlywhen the model is able to take the history of a patient intoaccount. Until now, the only published model in which patientscould experience multiple recurrences was a Markov modelpublished by Sobocki et al. in 2006 [47].

A recent cost-effectiveness analysis in the United States alsocompared adjunctive aripiprazole, quetiapine, and olanzapine foracute treatment of MDD [21]. It was concluded that the cost peradditional responder was lower for aripiprazole than for quetia-pine or olanzapine/fluoxetine [21]. In the present analysis, aripi-prazole is not only more effective but also cost saving. Thisdifference is likely due to the short horizon length of the USanalysis (6 weeks only) that cannot capture future cost savings.

Limitations

Because of a lack of Turkish evidence, Swedish-specific utilityvalues were used. Different utility values, however, would impactonly the magnitude of QALYs gained. As such, the base-case andPSA conclusions are independent of country-specific utilities. Clin-ical trial data were largely based on US populations; however, it isnot expected that these would be different in a Turkish population.

The only data available for adjunctive aripiprazole in MDDconsider the treatment of acute depressive episodes for a durationof 6 weeks. This is also the case for the comparators. The impact onrelapse prevention of adjunctive aripiprazole is yet to be inves-tigated. As such, the differences in outcomes between the threeadjunctive atypical antipsychotic treatments are solely based on thedifferences observed during the acute depressive episode treat-ment. When additional data about relapse prevention becomeavailable, however, the model can be adjusted to incorporate this.The relative efficacies of the three treatments are major drivers ofuncertainty in the incremental outcomes. Preferably, these shouldbe based on direct comparable data; however, such trials have notbeen conducted so far, which is why an indirect comparison of PCSswas used for this research. If direct comparable evidence wouldbecome available in the future, this can be imputed in the model.

Adverse events are not included in the model. This is becauseaugmentation treatment is given only for a short period of timeand most important side effects for these drugs will developwhen taking the drug for a longer duration. Aripiprazole aug-mentation treatment is associated with akathisia. This side effectwas not implemented in the model. This is because akathasiawill be present only in the 6 weeks of treatment and willtherefore have a minor influence on the total model outcomes.However, side effects commonly associated with quetiapine andolanzapine (such as extrapyrimidal syndrome, weight gain, anddiabetes) are also not incorporated. Because of the relativelyshort amount of time in which the drugs are given in the modeland the already high disability of the disease itself, it is expectedthat incorporating side effects would have only a minor influenceon model outcomes and omitting them is expected to be aconservative approach for aripiprazole.

Costs associated with monitoring tests during atypical anti-psychotic use were not incorporated. Because monitoring wouldbe required for all three atypical augmentation regimes com-pared, this would not affect incremental costs.

When implementing long-term data on the effects of aripi-prazole augmentation in the model, side effects should be takeninto account. In the long run, the chronic use of antipsychoticmedication is a risk factor for developing diabetes mellitus type 2and extensive weight gain. There is evidence, however, substan-tiating that aripiprazole is an exception and will have less impacton these side effects than, for example, quetiapine or olanzapine.

Indirect costs associated with work productivity losses, whichare not uncommon in patients with MDD, are not considered bythe Turkish payer, which is why these were excluded from theanalysis. Including indirect costs, however, would likely result infurther cost savings for aripipazole augmentation. Less time isspent in depressive episodes with this treatment, during whichpatients have a higher chance of incurring work productivitylosses than during phases of remission.

Conclusions

The cost-effectiveness model described here showed that aripi-prazole augmentation dominates quetiapine augmentation andolanzapine augmentation. Taking into account the uncertainty inall model input parameters, the probability that adjunctivearipiprazole would be cost-effective among the three strategies


ranged between 74% and 75% for WTP values between 0 TL and100,000 TL per QALY gained. Although atypical antipsychoticshave the same reimbursement status for augmentation treat-ment of MDD in Turkey, these economic findings may helpinform clinicians in their choice of antipsychotic augmentation.

Source of financial support: This work was sponsored byBristol Myers Squibb pharmaceuticals, Turkey. However, Phar-merit staff had complete access to all data and had final controlover the content, review, and submission of the manuscript.

R E F E R E N C E S

[1] Kessler RC, Berglund P, Demler O, et al. Lifetime prevalence and age-of-onset distributions of DSM-IV disorders in the National ComorbiditySurvey Replication. Arch Gen Psychiatry 2005;62:593–602.

[2] Kessler RC, Angermeyer M, Anthony JC, et al. Lifetime prevalence andage-of-onset distributions of mental disorders in the World HealthOrganization’s World Mental Health Survey Initiative. World Psychiatry2007;6:168–76.

[3] Alonso J, Angermeyer MC, Bernert S, et al. Prevalence of mentaldisorders in Europe: results from the European Study of theEpidemiology of Mental Disorders (ESEMeD) project. Acta PsychiatrScand Suppl 2004;420:21–7.

[4] Ministry of Health. Turkey Burden of Disease Study. Ankara: Ministry ofHealth, 2006.

[5] Burcusa SL, Iacono WG. Risk for recurrence in depression. Clin PsycholRev 2007;27:959–85.

[6] Khan A, Khan SR, Leventhal RM, Brown WA. Symptom reduction andsuicide risk in patients treated with placebo in antidepressant clinicaltrials: a replication analysis of the Food and Drug AdministrationDatabase. Int J Neuropsychopharmacol 2001;4:113–8.

[7] Murray CJ, Lopez AD. Evidence-based health policy–lessons from theGlobal Burden of Disease Study. Science 1996;274:740–3.

[8] Wang H, Kindig DA, Mullahy J. Variation in Chinese population healthrelated quality of life: results from a EuroQol study in Beijing, China.Qual Life Res 2005;14:119–32.

[9] Anderson I, Pilling S, Barnes A, et al. Depression: the treatment andmanagement of depression in adults. Leicester, UK: NationalCollaborating Centre for Mental Health Commissioned by the NationalInstitute for Health and Care Excellence, 2009.

[10] Sobocki P, Ekman M, Agren H, et al. Health-related quality of lifemeasured with EQ-5D in patients treated for depression in primarycare. Value Health 2007;10:153–60.

[11] Stouthard ME, Essink-Bot ML, Bonsel G, et al. Disability Weights forDiseases in the Netherlands. Rotterdam: Department of Public Health, 1997.

[12] Bennett KJ, Torrance GW, Boyle MH, Guscott R. Cost-utility analysis indepression: the McSad utility measure for depression health states.Psychiatr Serv 2000;51:1171–6.

[13] Revicki DA, Wood M. Patient-assigned health state utilities fordepression-related outcomes: differences by depression severity andantidepressant medications. J Affect Disord 1998;48:25–36.

[14] Sapin C, Fantino B, Nowicki ML, Kind P. Usefulness of EQ-5D inassessing health status in primary care patients with major depressivedisorder. Health Qual Life Outcomes 2004;2:20.

[15] Ministry of Health Turkey. National Burden of Disease and CostEffectiveness Project. Ankara, Turkey: RSHMB School of Public HealthDirectorate, 2004.

[16] Sobocki P, Ekman M, Agren H, et al. The mission is remission: healtheconomic consequences of achieving full remission withantidepressant treatment for depression. Int J Clin Pract 2006;60:791–8.

[17] Trivedi MH, Rush AJ, Wisniewski SR, et al. Evaluation of outcomes withcitalopram for depression using measurement-based care in STAR*D:implications for clinical practice. Am J Psychiatry 2006;163:28–40.

[18] Trivedi MH. Treating depression to full remission. J Clin Psychiatry2009;70:e01.

[19] Ansseau M, Demyttenaere K, Heyrman J, et al. Objective: remission ofdepression in primary care. The Oreon Study. Eur Neuropsychopharmacol2009;19:169–76.

[20] Ministry of Health Turkey. Short product summary aripiprazole.Available from: http://www.iegm.gov.tr/Folders/KubKT/Ruhsatlý%20Ürünler-2%20Þube%20Müdürlüðü/ABILIFY%205%20mg%20KUB%20autism_006c814.pdf. [Accessed May 22, 2010].

[21] Taneja C, Papakostas GI, Jing Y, et al. Cost-effectiveness of adjunctivetherapy with atypical antipsychotics for acute treatment of majordepressive disorder. Ann Pharmacother 2012;46:642–9.

[22] Berman RM, Marcus RN, Swanink R, et al. The efficacy and safety ofaripiprazole as adjunctive therapy in major depressive disorder: a

multicenter, randomized, double-blind, placebo-controlled study. J ClinPsychiatry 2007;68:843–53.

[23] Berman RM, Fava M, Thase ME, et al. Aripiprazole augmentation inmajor depressive disorder: a double-blind, placebo-controlled study inpatients with inadequate response to antidepressants. CNS Spectr2009;14:197–206.

[24] Marcus RN, McQuade RD, Carson WH, et al. The efficacy and safety ofaripiprazole as adjunctive therapy in major depressive disorder: asecond multicenter, randomized, double-blind, placebo-controlledstudy. J Clin Psychopharmacol 2008;28:156–65.

[25] Bauer M, Pretorius HW, Constant EL, et al. Extended-release quetiapineas adjunct to an antidepressant in patients with major depressivedisorder: results of a randomized, placebo-controlled, double-blindstudy. J Clin Psychiatry 2009;70:540–9.

[26] McIntyre A, Gendron A, McIntyre A. Quetiapine adjunct to selective sero-tonin reuptake inhibitors or venlafaxine in patients with major depression,comorbid anxiety, and residual depressive symptoms: a randomized,placebo-controlled pilot study. Depress Anxiety 2007;24:487–94.

[27] Shelton RC, Williamson DJ, Corya SA, et al. Olanzapine/fluoxetinecombination for treatment-resistant depression: a controlled study ofSSRI and nortriptyline resistance. J Clin Psychiatry 2005;66:1289–97.

[28] Thase ME, Corya SA, Osuntokun O, et al. A randomized, double-blindcomparison of olanzapine/fluoxetine combination, olanzapine, andfluoxetine in treatment-resistant major depressive disorder. J ClinPsychiatry 2007;68:224–36.

[29] Corya SA, Williamson D, Sanger TM, et al. A randomized, double-blindcomparison of olanzapine/fluoxetine combination, olanzapine,fluoxetine, and venlafaxine in treatment-resistant depression. DepressAnxiety 2006;23:364–72.

[30] Nierenberg AA, Papakostas GI, Petersen T, et al. Lithium augmentationof nortriptyline for subjects resistant to multiple antidepressants. J ClinPsychopharmacol 2003;23:92–5.

[31] Joffe RT, Singer W, Levitt AJ, MacDonald C. A placebo-controlledcomparison of lithium and triiodothyronine augmentation of tricyclicantidepressants in unipolar refractory depression. Arch Gen Psychiatry1993;50:387–93.

[32] Lunn D, Thomas A, Best N, Spiegelhalter D. WinBUGS—a Bayesianmodelling framework: concepts, structure, and extensibility. StatComput 2002;10:325–37.

[33] Treur M, Postema R, Loze JY, et al. The comparative efficacy and safetyof adjunctive antipsychotics in major depressive disorder patients thatfailed to respond to conventional antidepressants. In: ISPOR 15thAnnual International Meeting. Atlanta, May 17–19, 2010.

[34] Rush AJ, Trivedi MH, Wisniewski SR, et al. Acute and longer-termoutcomes in depressed outpatients requiring one or several treatmentsteps: a STAR*D report. Am J Psychiatry 2006;163:1905–17.

[35] Solomon DA, Keller MB, Leon AC, et al. Multiple recurrences of majordepressive disorder. Am J Psychiatry 2000;157:229–33.

[36] Bernal M, Haro JM, Bernert S, et al. Risk factors for suicidality in Europe:results from the ESEMED study. J Affect Disord 2007;101:27–34.

[37] Devrimci-Ozguven H, Sayil I. Suicide attempts in Turkey: results of theWHO-EURO Multicentre Study on Suicidal Behaviour. Can J Psychiatry2003;48:324–9.

[38] Bilici M, Bekaroglu M, Hocaoglu C, et al. Incidence of completed andattempted suicide in Trabzon, Turkey. Crisis 2002;23:3–10.

[39] Burstrom K, Johannesson M, Diderichsen F. Swedish population health-related quality of life results using the EQ-5D. Qual Life Res 2001;10:621–35.

[40] Ministry of Health, Directorate of Pharmaceuticals, Official Web Page.Available from: http://www.iegm.gov.tr/Default.aspx?sayfa=fiyat_listesi.[Accessed February 2, 2010].

[41] Karamustafalioglu O, Hemels MEH, Save D, Mene S, Özmen E.Economic evaluation of first-line treatments for depression inTurkey: a cost-effectiveness model. In: ISPOR 10th AnnualInternational Meeting. Washington, DC, May 2005. Value Health2005;(8):392.

[42] Health implementation guideline. Turkish Republic Official Gazette2010;27532(Suppl. 8):98.

[43] Revicki DA, Brown RE, Palmer W, et al. Modelling the cost effectivenessof antidepressant treatment in primary care. Pharmacoeconomics1995;8:524–40.

[44] Revicki DA, Brown RE, Keller MB, et al. Cost-effectiveness of newerantidepressants compared with tricyclic antidepressants in managedcare settings. J Clin Psychiatry 1997;58:47–58.

[45] National Institute for Health and Care Excellence. Guide in the methods oftechnology appraisal (reference NO515). Available from: http://www.nice.org.uk/media/B52/A7/TAMethodsGuideUpdatedJune2008.pdf. [AccessedMay 22, 2010].

[46] Drummond MF, Jefferson TO. Guidelines for authors and peerreviewers of economic submissions to the BMJ. The BMJ EconomicEvaluation Working Party. BMJ 1996;313:275–83.

[47] Sobocki P, Ekman M, Agren H, et al. Model to assess the cost-effectiveness of new treatments for depression. Int J Technol AssessHealth Care 2006;22:469–77.



























































http://www.iegm.gov.tr
































































http://www.iegm.gov.tr/Default.aspx?sayfa=fiyat_listesi

























Cost-Utility Analysis of Depot Atypical Antipsychotics for ChronicSchizophrenia in CroatiaVlado Jukic, MD, PhD1, Miro Jakovljevic, MD, PhD2, Igor Filipcic, MD, PhD2, Miroslav Herceg, MD, PhD1, Ante Silic, MD, PhD3,Tatjana Tomljanovic, MD4, Roman Zilbershtein, MSc5, Rasmus C.D. Jensen, MSc6, Michiel E.H. Hemels, MSc, MBA6,Thomas R. Einarson, PhD7,�

1University Psychiatric Hospital “Vrapce,” Zagreb, Croatia; 2Department of Psychiatry, University Hospital of Zagreb, Zagreb, Croatia; 3University PsychiatricHospital “Sv Ivan,” Zagreb, Croatia; 4Janssen Division of Johnson & Johnson S.E., Zagreb, Croatia; 5Pivina Consulting, Inc., Mississauga, Canada; 6Janssen Cilag,Birkerød, Denmark; 7Leslie Dan Faculty of Pharmacy, University of Toronto, Toronto, ON, Canada

A B S T R A C T

Objectives: As a nation with a developing economy, Croatia is facedwith making choices between pharmaceutical products, includingdepot injectable antipsychotics. We conducted a pharmacoeconomicanalysis to determine the cost-effectiveness of atypical depots inCroatia. Methods: A 1-year decision-analytic framework modeleddrug use. We determined the average direct cost to the CroatianInstitute for Health Insurance of using depot formulations of paliper-idone palmitate long-acting injectable (PP-LAI), risperidone LAI (RIS-LAI), or olanzapine LAI (OLZ-LAI). An expert panel plus literature-derived clinical rates populated the core model, along with costsadjusted to 2012 by using the Croatian consumer price index. Clinicaloutcomes included quality-adjusted life-years, hospitalization rates,emergency room treatment rates, and relapse days. Robustness ofresults was examined with one-way sensitivity analyses on importantinputs; overall, all inputs were varied over 10,000 simulations in aMonte Carlo analysis. Results: Costs (quality-adjusted life-years) perpatient were €5061 (0.817) for PP-LAI, €5168 (0.807) for RIS-LAI, and

€6410 (0.812) for OLZ-LAI. PP-LAI had the fewest relapse days,emergency room visits, and hospitalizations. Results were sensitiveagainst RIS-LAI with respect to drug costs and adherence rates, butwere generally robust overall, dominating OLZ-LAI in 77.3% and RIS-LAI in 56.8% of the simulations. Conclusions: PP-LAI dominated theother drugs because it had the lowest cost and best clinical outcomes.Compared with depots of olanzapine and risperidone and oralolanzapine, PP-LAI was the cost-effective atypical LAI for treatingchronic schizophrenia in Croatia. Using depot paliperidone in place ofeither olanzapine or risperidone would reduce the overall costs ofcaring for these patients.

Keywords: Croatia, long-acting injectable, paliperidone palmitate,pharmacoeconomic analysis, risperidone, schizophrenia.


Introduction

Croatia is a country with a developing economy whose health caresystem has been developing over some time [1,2]. The CroatianInstitute for Health Insurance [3] was established in 1993 to managethe health system that is provided for all citizens. This system isbased on a national health insurance model, with compulsorycontributions for all employed persons and employers, with co-payments for drugs and services. Noninsured services are paid eitherout of pocket or through additional voluntary health insurance.

In 1998, a law was passed that guaranteed treatment and alsosafeguarded the personal rights for persons with schizophreniarequiring involuntary hospitalization [4,5]. Since that time, therehas been an increase in both the availability and use of anti-psychotics [6,7]. Utilization patterns have changed, with a shiftfrom the older, less expensive first-generation drugs to newerand more costly atypical antipsychotics [6–10].

Problems in drug use, especially nonadherence and polyphar-macy, have exacerbated the situation [11,12]. As a result, therehave been increasing numbers of hospitalizations [13]. Harveyet al. [14] noted that hospitalization consumed the largestproportion of total health care costs (430%) in Croatia. Inaddition, the plan to reintegrate persons with schizophrenia intothe community has not been entirely successful [15]. All thesefactors have served to increase costs for the health system;however, its financial resources are insufficient to cover all thedemands. Therefore, cost-effective approaches are needed tomaintain and improve the treatment of persons with chronicschizophrenia.

In 2002, the Croatian Institute for Health Insurance funded aproject to address rational drug use [14]. A main component wasdetermining how savings could be made by incorporating phar-macoeconomic principles into the selection and purchase ofdrugs on the Croatian formulary. They observed that “there was






� Address correspondence to: Thomas R. Einarson, Leslie Dan Faculty of Pharmacy, 144 College Street, Room 674, University of Toronto,Toronto, ON, Canada M5V 3M8.


dx.doi.org/10.1016/j.vhri.2013.06.008

dx.doi.org/10.1016/j.vhri.2013.06.008

dx.doi.org/10.1016/j.vhri.2013.06.008


dx.doi.org/10.1016/j.vhri.2013.06.008

a lack of information about what constituted cost-effective treat-ment appropriate to the Croatian economic situation.” Thisobservation highlights the need for evidence-based informationon the cost-effectiveness of psychopharmaceuticals.

At the same time, the rights and dignity of these people mustbe respected. Two critical aspects are image [16] and quality oflife [17,18]. Nawková et al. [19] assessed articles in the lay press inCroatia describing mental health and found that 40 out of the 75(53%) articles portrayed a negative image. Mentally ill personswere mostly presented as dangerous and involved in aggressivecrimes such as homicide (49%) or physical assaults (31%). Martić-Biocina and Barić [12] also identified dissatisfaction with the roleof the media in that respect. They also found a high level ofstigma toward people with schizophrenia and that it correlatedwith medication nonadherence and hospitalizations. The samewas found with quality-of-life issues [18]. These authors alsoreported that the atypical antipsychotics were superior to tradi-tional drugs with respect to increasing quality of life in personswith chronic schizophrenia. Jukić et al. [20] suggested that theirside-effect profile may be responsible for improved quality of life.

Depot forms of antipsychotic drugs were developed to at leastpartially address issues of nonadherence [21]. In the past decade,long-acting injectable (LAI) formulations of atypical agents havebeen marketed to fill a perceived need. Risperidone LAI (RIS-LAI)was the first such drug [22], and has recently been joinedby olanzapine LAI (OLZ-LAI) [23] and paliperidone LAI (PP-LAI)[24]. In another country undergoing economic change, a

pharmacoeconomic analysis found that PP-LAI was cost-effective when compared with RIS-LAI [25]. It is not currentlyknown whether the outcomes would be similar in this country.We therefore undertook this research to assess the cost-utility ofPP-LAI compared with other LAIs in Croatia from the point ofview of the Croatian Institute for Health Insurance.

Methods

Target Population

We examined the use of atypical LAIs in persons with stablechronic schizophrenia but who had a history of relapses andhospitalizations. They have been referred to as “revolving door”patients who are difficult to treat and have problems withadherence to prescribed medications [26,27]. Consequently, theyimpose a very large burden on health care resources.

Drugs Analyzed

The drug of primary interest was PP-LAI. Comparison drugsincluded the other atypical depots (i.e., RIS-LAI and OLZ-LAI).According to the product summaries of the European MedicinesAgency, PP-LAI (Xeplion) can be dosed monthly [28], OLZ-LAI(Zypadhera) is administered every 2 or 4 weeks [29], and RIS-LAI(Risperdal Consta) requires biweekly injections [30].

Fig. 1 – Model used for the pharmacoeconomic analysis. LAI, long-acting injectable; OLZ-LAI, olanzapine pamoate long-actinginjectable; PP-LAI, paliperidone palmitate long-acting injectable.


Model and base caseWe adapted a previously validated decision tree [25] for use inCroatia, using input from clinical and administrative experts.Figure 1 depicts the model. To begin the analysis, we start with

an average patient having chronic relapsing schizophrenia butwhose disease is currently stabilized. Because of adherenceproblems, patients are maintained on standard doses of depotantipsychotics. They may be either adherent or nonadherent, as

Table 1 – Clinical inputs into the model.

LAIantipsychotic

Clinical input(amount/rate)

Rate Source

Olanzapine Dose toinitializetherapy

300 mg q2 weeks � 3injections

European Medicines Agency [29]

Maintenancedose

432 mg q4 weeks Kane et al. [34]

Dose for acuterelapse

473 mg q4 weeks Lauriello et al. [35]

Adherence 0.803 Ascher-Svanum et al. [36]Stable disease 0.793 Kane et al. [34]ER visits 0.062 Kane et al. [34]Hospitalization

rate0.145 Calculation [1 – rates of (ER visits þ stable disease)]

Paliperidone Dose toinitializetherapy

150 mg week 1, 100 mgweek 2, then 82.8 mgevery 4 wk

European Medicines Agency [28], Hough et al. [37]

Maintenancedose

69.3 mg monthly Average from Gopal et al. [38] and Fleischhacker et al. [39]


84.9 mg monthly Gopal et al. [40], Pandina et al. [41], Hough et al. [42], Nasrallah et al.[43], Pandina et al. [44]

Adherence 0.872 RIS-LAI rate from Olivares et al. [45] adjusted via Mehnert and Diels[46]

Stable disease 0.803 Calculation [1 – rates of (ER visits þ hospitalization)]ER visits 0.059 Hospital rate � ER:hospital ratio from Ascher-Svanum et al. [36]Hospitalization

rate0.138 Hough et al. [37], Gopal et al. [38]

Risperidone Maintenancedose

40.3 mg biweekly Fleischhacker et al. [39], Kissling et al. [47], Lee et al. [48],Lindenmayer et al. [49], Olivares et al. [50]


50 mg biweekly Prorated from PP-LAI dose; similar to doses used by Kane et al. [22],Chue et al. [51], Eerdekens et al. [52] who used 58 mg, but 50 mg isthe maximum allowable dose [30]

Adherence 0.823 Olivares et al. [45]Stable disease 0.763 Calculation [1 – rates of (ER visits þ hospitalization)]ER visits 0.071 Hospital rate � ER:hospital ratio from Ascher-Svanum et al. [36]Hospitalization

rate0.166 Olivares et al. [50], Weiden and Olfson [53]

ER, emergency room; LAI, long-acting injectable; PP, paliperidone palmitate; RIS, risperidone microspheres.

Table 2 – Cost inputs (2012€).

Resource Item Unit Cost (€) Source

Drugs Olanzapine depot mg 3.50 Croatian Official Gazette [58]Risperidone depot mg 3.50 Croatian Official Gazette [58]Paliperidone depot mg 3.76 Croatian Official Gazette [58]Olanzapine tablets mg 0.21 Croatian Official Gazette [58]Risperidone tablets mg 0.20 Croatian Official Gazette [58]Clozapine tablets mg 0.0036 Croatian Official Gazette [58]

Medical Primary care physician 1 visit 16.62 Local medical price listPsychiatrist outpatient visit 1 visit 14.10 Vrapče hospital price listPsychiatric nurse 1 h 7.45 Vrapče hospital price list

Hospital Emergency room 1 visit 230.32 Vrapče hospital price listHospital bed acute care 1 d 96.81 Vrapče hospital price listHospital long-term bed 1 d 33.51 Vrapče hospital price listDay care visit 1 d 26.20 Vrapče hospital price list


per published rates and expert opinion. Some will continue in thestable state, while the rest will relapse. All relapsers would beseen at the emergency room, and the more severe cases would beadmitted to the acute psychiatric unit. Those unable or unwillingto tolerate the initial treatment would be switched. Thosediscontinuing PP-LAI or RIS-LAI would then receive OLZ-LAI,and those switching from OLZ-LAI would receive PP-LAI; dis-continuers of oral OLZ would receive RIS-LAI. Patients who failedtwo different drugs would be given clozapine, in accordance withNational Institute for Health and Clinical Excellence guidelines[31] and local practice using doses reported in the literature[32,33].

Clinical Inputs

Table 1 lists the clinical inputs used in this mode, by drug, aswell as the sources of information [22,28–30,34–53]. The dosesof drugs used were derived from randomized clinical trialsand long-term studies published in the literature. Where datawere presented for nonadherent patients, we extracted ratesfrom the published articles. In other cases, we used eitherthe rates for placebo in trials (e.g., PP-LAI; OLZ-LAI assumedequal) or we calculated rates by using the ratio of adherent:nonadherent patients from Weiden and Olfson [53] (e.g., RIS-LAI).

With each drug, adjunct therapy was added, in accordancewith typical clinical trials. Gopal et al. [38] indicated that PP-LAIwas augmented with oral risperidone in a dose of 6.8 mg/d for30.5 days. In a similar trial, Möller et al. [54] reported that 22%of the patients receiving RIS-LAI required oral supplementationwith 3.2 mg/d for 43 days. Ascher-Svanum et al. [55] notedthat OLZ-LAI required 10.8 mg/d for 31 days in 21% of thepatients.

Some economic analyses have used standard doses such asdefined daily doses (DDDs) [56]; however, DDDs reflect theaverage dose when used for the most common indication. Itshould be remembered that we are dealing with revolving doorpatients who comprise only a subset of patients who representthe extreme of the spectrum. Thus, we feel that DDDs wouldunderestimate the doses used in actual practice when managingsuch problematic patients.

In calculating adherence rates, we used experience from largeobservational studies. The rate for RIS-LAI was taken from a largepatient registry (n ¼ 1648) of patients with longstanding schizo-phrenia [45]. Because there was insufficient long-term experiencewith PP-LAI, that rate was adjusted by using results from thestudy by Mehnert and Diels [46]. They compared adherencebetween RIS-LAI when administered monthly and twice weekly,finding a minimum of 5.1% increase in adherence with monthlyinjections. That factor was applied to PP-LAI, which is adminis-tered monthly and is a metabolite of RIS-LAI, thereby having thesame adverse-event profile. For OLZ-LAI, we used the rate from alarge cohort (n ¼ 1906). Even though that rate was with oral drugs(which normally have lower adherence rates than do depots), weused that value of 80.2% because it was higher than the 72.7%rate in 931 patients found with the depot form by Ascher-Svanum et al. [57], and was quite similar to the rates of otherdepot atypicals.

Cost Inputs

We considered all direct costs from the viewpoint of the NationalHealth Service of Croatia, as presented in Table 2 [58] (localmedical price list and Vrapče hospital price list). We did notinclude indirect costs such as time lost from work, becausevery few of these people participate fully in the workforce.A multicountry study in Europe reported that less than 10% of

Table

3–Clinicaland

eco

nom

icoutp

uts

from

thedecision

treem

odelanaly

sisfo

rCro

atia.

Dru

gTotal

cost

per

patient(€)*

Day

sin

remission

Emerge

ncy

room

visits

Hos

pitalizations

Total

QALY

sper

patient

Increm

entalc

ostper

patient(€)†

Increm

entalQALY

sper

patient†

Econ

omic

conclusion

PP-LAI

5061

329.3

0.12

90.25

20.81

7Dominan

t‡

RIS-LAI

5168

322.8

0.14

80.30

50.80

7�1

070.01

0Dominated

OLZ

-LAI

6410

325.5

0.14

30.28

00.81

2�1

349

0.00

5Dominated

LAI,long-ac

tinginjectab

le;PP

,paliperidonepalmitate;

RIS,risp

eridonemicro

spheres

;QALY

s,quality-ad

justed

life-yea

rs.

�Allco

stsarein

2012

euro

s.†Bec

ause

ofdominan

ce,co

mparisonsareag

ainst

PP-LAI.

‡Has

lowes

tco

stan

dhighes

tQALY

san

dth

erefore

isth

edru

gofch

oice.


3996 persons with schizophrenia were employed full-time andanother 12.3% worked part-time [59]. We did not apply discount-ing because the analytic time horizon was 1 year. Priceswere taken from current lists or from the literature, and theninflated to 2011 euros by using the consumer price index forCroatia [60].

Utilities

Utilities for the analysis were obtained from the literature;values obtained were simply averaged [61–65]. Stable diseasehad a utility of 0.890; an exacerbation requiring outpatienttreatment at the hospital emergency room had a utility of0.659 for emergency room exacerbation and 0.490 forhospitalization.

Analysis and Outputs

No official guidelines currently exist for pharmacoeconomicanalyses in Croatia. We therefore used a standard approach thathad been used in previous analyses in Europe [25]. The decisiontree produced expected outcomes for the average patient treatedwith average doses of each drug. These outcomes included thecost per patient treated, measured in 2012 euros, as well asnumbers of hospitalizations, emergency room visits, days withstable disease, and quality-adjusted life-years (QALYs) for eachdrug. The economic outcome of prime interest was the incre-mental cost per QALY.

We explored the effect of variations in input values on out-puts by applying one-way sensitivity analyses on all importantinputs such as costs and clinical rates. Break-even analysisidentified the points where outcomes changed qualitatively. We

Fig. 2 – Cloud diagram from the probabilistic sensitivity analysis comparing incremental costs (Y axis) and QALYs (X axis)between (A) PP-LAI and RIS-LAI and (B) PP-LAI and OLZ-LAI. OLZ-LAI, olanzapine pamoate long-acting injectable; PP-LAI,paliperidone palmitate long-acting injectable; QALY, quality-adjusted life-year; RIS-LAI, risperidone long-acting injectable.


also conducted a set of pairwise probabilistic analyses by using10,000 Monte Carlo simulations on all inputs and standarddistributions (i.e., beta for rates and gamma for costs) [66].Proportions of incremental cost-effectiveness ratios falling intoeach of the four major quadrants (cost vs. QALYs) were calculatedand compared.

Results

Results of the base-case analysis appear in Table 3. PP-LAI hadthe lowest overall cost to treat one patient for 1 year (€5061),followed by RIS-LAI (€5168), with OLZ-LAI costing the most(€6410). Clinical outcomes were also better in all cases for PP-LAI; it had the most days in remission and the fewest hospital-izations and emergency room visits. Also, it was associatedwith the highest QALY score, but differences for this outcomewere not large. Because its cost was lowest and QALYs (andother beneficial clinical outcomes) highest, it dominated theother drugs. That is, it should be considered the preferredchoice.

In one-way sensitivity analyses, PP-LAI was not sensitive tochanges in cost relative to OLZ-LAI. Its cost would need toincrease 64% or that of OLZ-LAI decrease by 54% to lose itsdominance. However, it would not dominate RIS-LAI with a 4%increase in the cost of PP-LAI or a 4.4% decrease in the cost of RIS-LAI. If the unit cost of PP-LAI were equal to that of RIS-LAI, theexpected cost/patient would decline to €4756. Results weresensitive to reasonable changes in adherence rate (�10% forOLZ-LAI and �18% for RIS-LAI). Hospitalization rates were notsensitive.

Figure 2A,B depicts results from the probabilistic sensitivityanalyses. PP-LAI dominated OLZ-LAI in about 77.3% of thesimulations and RIS-LAI in 56.8% of the simulations. It wascost-effective (i.e., incremental cost-effectiveness ratios o€30,000) compared with RIS-LAI in another 37% of the trials(overall 93%). However, PP-LAI was dominated in 1.3% of the20,000 iterations, in total.

Discussion

A search of the literature could find no examples of pharmacoe-conomic analyses that examined the pharmacotherapy of schiz-ophrenia in Croatia. Therefore, we believe that this is the firstsuch analysis. Because decision makers and health care providersare being faced with increasing demands from patients and theiradvocates without a corresponding increase in revenues, theymust take advantage of these quantitative approaches to aid inselecting what to fund. Relying solely on acquisition prices ofdrugs, services, or other products can be misleading because allfactors impacting the choice are not being considered. Because ofenhanced efficacy, a drug with a higher price may be the bestchoice if it prevents the consumption of other resources, such ashospitalization.

In this analysis, PP-LAI dominated the other available atypicalLAIs. Results could change against RIS-LAI with changes in costor against OLZ-LAI with changes in adherence. The overallprobabilistic sensitivity analyses, however, did indicate that PP-LAI would be the drug of choice in the majority of cases.

In addition to the clinical and economic advantages, there isan advantage for PP-LAI with respect to convenience. This drugmay be administered monthly, while its nearest competitor, RIS-LAI, must be given every 2 weeks. Monthly dosing would seem tobe the preferable situation for both the patient and the busypractitioner.

This analysis has some limitations, which should be noted.Rates of adherence and hospitalization were taken from theliterature and were assumed to apply as well in this country.We considered only persons with chronic schizophrenia in astable state. Those who are hospitalized or experiencing an acuteexacerbation of symptoms would require more aggressive treat-ment; therefore, costs and outcomes might vary.

We did not consider the treatment of adverse events otherthan postinjection syndrome in this analysis for a number ofreasons. First, these events are quite common in all antipsychoticdrugs. Many of these problems can be managed by reducing thedose or changing to another drug. Also, these patients arerequired to visit their physician or other practitioner (e.g.,psychiatric nurse or psychologist) on a regular basis, and so theywould not likely incur extra visits because of adverse events. Inaddition, many require treatment with drugs (e.g., anticholiner-gics) that are very inexpensive and add little to the overall cost ofcare, especially on a comparative basis. Finally, reports fromofficial agencies have concluded that adverse events associatedwith these drugs have little appreciable impact overall [67,68].

Conclusions

In this analysis, we found that PP-LAI was the dominant choicefor treating chronic relapsing schizophrenia in Croatia. Its higheracquisition cost was more than offset by reductions in otherhealth care areas, such as decreased hospitalizations, visits toemergency room, and visits to other health care practitioners.Results were sensitive to minor changes in adherence rates.

Source of financial support: This study was funded by JanssenA/S, Beerse, Belgium. Dr. Tomljanovic, Mr. Rasmussen, and Mr.Hemels are all employees of Janssen. Dr. Einarson receivedfunding for this research. Mr. Zilbershtein is an employee ofPivina Conculting, Inc., who received funding for this research.

R E F E R E N C E S

[1] European Observatory on Health Care Systems. Health Care Systems inTransition: Croatia (2nd ed.). Copenhagen: WHO Regional Office forEurope, 1999. Available from: http://www.euro.who.int/__data/assets/pdf_file/0018/75150/E68394.pdf. [Accessed March 16, 2012].

[2] Mastilica M, Kušec S. Croatian healthcare system in transition, from theperspective of users. BMJ 2005;331:223–7.

[3] Croatian Institute of Health Insurance. Available from: http://www.hzzo-net.hr. [Accessed May 3, 2012].

[4] Jukić V, Goreta M, Kozumplik O, et al. Implementation of first CroatianLaw on Protection of Persons with Mental Disorders. Coll Antropol2005;29:543–9.

[5] Kozumplik O, Jukić V, Goreta M. Involuntary hospitalizations ofpatients with mental disorders in Vrapce Psychiatric Hospital: fiveyears of implementation of the first Croatian law on protection ofpersons with mental disorders. Croat Med J 2003;44:601–5.

[6] Jukić V, Herceg M, Savić A. Availability of psychiatric medications toCroatian healthcare users and the influence of availability of atypicalantipsychotics on psychiatric hospital morbidity. Psychiatr Danub2011;23:320–4.

[7] Jukić V, Savić A, Herceg M. Availability of new psychiatric medications,especially antipsychotics, in context of patient rights anddestigmatization of psychiatric patients. Psychiatr Danub2011;23:316–9.

[8] Jukić V, Herceg M, Brecić P, et al. Dynamic in prescribing antipsychoticdrugs during five year period (2001–2005) in the Psychiatric HospitalVrapce, Zagreb, Croatia. Coll Antropol 2008;32(Suppl 1):211–3.

[9] Štimac D, Čulig J. Outpatient utilization of psychopharmaceuticals inthe city of Zagreb 2001–2006. Psychiatr Danub 2009;21:56–64.

[10] Štimac D, Vukusić I, Čulig J, et al. Outpatient utilization ofpsychopharmaceuticals: comparison between Croatia andScandinavian countries (2001–2003). Coll Antropol 2009;33:237–43.

[11] Loga-Zec S, Loga S. Polypharmacy in the treatment of schizophrenicpatients in three university centers in the Federation of Bosnia andHerzegovina (F/BH). Psychiatr Danub 2011;23:60–3.


http://refhub.elsevier.com/S2212-1099(13)00071-X/othref0005


http://www.euro.who.int/&underscore;&underscore;data/assets/pdf&underscore;file/0018/75150/E68394.pdf

http://www.euro.who.int/&underscore;&underscore;data/assets/pdf&underscore;file/0018/75150/E68394.pdf

http://refhub.elsevier.com/S2212-1099(13)00071-X/sbref1


http://www.hzzo-net.hr

http://www.hzzo-net.hr



























[12] Martić-Biocina S, Barić V. Factors important for compliance amongCroatian patients suffering from schizophrenia—how to improvepsychiatric services in Croatia? Psychiatr Danub 2005;17:36–41.

[13] Bilić P, Ivanis A, Vidović D, Jukić V. Changing the structure of thehospitalized patients at the Psychiatric Clinic Vrapce. Srp Arh Celok Lek2011;139(Suppl 1):33–5.

[14] Harvey K, Kalanj K, Stevanović R. Croatian pharmaceutical sectorreform project: rational drug use. Croat Med J 2004;45:611–9.

[15] Jukić V. (Re)Integration of mental patients—mixed media messages.Psychiatr Danub 2008;20:433–6.

[16] Bilić B, Georgaca E. Representations of “mental illness” in Serbiannewspapers: a critical discourse analysis. Qual Res Psychol2007;4:167–86.

[17] Pesek MB. Therapy and quality of life of patients with psychosis.Psychiatr Danub 2009;21(Suppl. 1):146–8.

[18] Jašović-Gašić M, Lačković M, Dunjić-Kostić B, et al. Critical review ofstudies on quality of life in psychiatric patients published in Serbianmedical journals from 2000 to 2009. Psychiatr Danub 2010;22:488–94.

[19] Nawková L, Nawka A, Adámková T, et al. The picture of mental health/illness in the printed media in three central European countries. JHealth Commun 2012;17:22–40.

[20] Jukić V, Barić V, Culav-Sumić J, et al. The impact of novel antipsychoticdrugs on quality of life among people suffering from schizophrenia.Coll Antropol 2003;27(Suppl. 1):119–24.

[21] Nasrallah H. The case for long-acting antipsychotic agents in the post-CATIE era. Acta Psychiatr Scand 2007;115:60–267.

[22] Kane J, Eerdekens M, Lindenmayer J-P, et al. Long-acting injectablerisperidone: efficacy and safety of the first long-acting atypicalantipsychotic. Am J Psychiatry 2003;60:1125–32.

[23] European Medicines Agency. Assessment report for Zypadhera. Doc.Ref EMEA/608654/2008. Procedure No. EMEA/H/C/000890. 2008:1–63.Available from: http://www.ema.europa.eu/docs/en_GB/document_library/EPAR_-_Public_assessment_report/human/000890/WC500054428.pdf. [Accessed March 3, 2012].

[24] European Medicines Agency. Xeplions opinion 17-12-2010. Availablefrom: http://www.ema.europa.eu/ema/index.jsp?curl=pages/medicines/human/medicines/002105/smops/Positive/human_smop_000162.jsp&murl=menus/medicines/medicines.[Accessed March 3, 2012].

[25] Einarson TR, Geitona M, Chaidemenos A, et al. Pharmacoeconomicanalysis of paliperidone palmitate for treating schizophrenia in Greece.Ann Gen Psychiatry 2012;11:18.

[26] Haywood TW, Kravitz HM, Grossman LS, et al. Predicting the “revolvingdoor” phenomenon among patients with schizophrenic, schizoaffective,and affective disorders. Am J Psychiatry 1995;152:856–61.

[27] Weiden P, Glazer W. Assessment and treatment selection for “revolvingdoor” inpatients with schizophrenia. Psychiatr Q 1997;68:377–92.

[28] European Medicines Agency. Xeplions summary of productcharacteristics. Available from: http://www.ema.europa.eu/docs/en_GB/document_library/EPAR_-_Product_Information/human/002105/WC500103317.pdf. [Accessed March 3, 2012].

[29] European Medicines Agency. Zypadheras summary of productcharacteristics. Available from: http://www.ema.europa.eu/docs/en_GB/document_library/EPAR_-_Product_Information/human/000890/WC500054429.pdf. [Accessed March 3, 2012].

[30] European Medicines Agency. Risperdal Constas summary of productcharacteristics. Available from: http://www.ema.europa.eu/docs/en_GB/document_library/Referrals_document/Risperdal_Consta_30/WC500008170.pdf. [Accessed March 3, 2012].

[31] National Institute for Health and Care Excellence. Schizophrenia: coreinterventions in the treatment and management of schizophrenia inadults in primary and secondary care. NICE Clinical Guideline 82(update). Available from: www.nice.org.uk. [Accessed April 4, 2012].

[32] Simonsen E, Friis S, Opjordsmoen S, et al. Early identification of non-remission in first-episode psychosis in a two-year outcome study. ActaPsychiatr Scand 2010;122:375–83.

[33] Wahlbeck K, Cheine M, Essali A, Adams C. Evidence of clozapine’seffectiveness in schizophrenia: a systematic review and meta-analysisof randomized trials. Am J Psychiatry 1999;156:990–9.

[34] Kane J, Detke H, Naber D, et al. Olanzapine long-acting injection: a 24-week, randomized, double-blind trial of maintenance treatment inpatients with schizophrenia. Am J Psychiatry 2010;167:181–9.

[35] Lauriello J, Lambert T, Andersen S, et al. An 8-week, double-blind,randomized, placebo controlled study of olanzapine long-actinginjection in acutely ill patients with schizophrenia. J Clin Psychiatry2008;69:790–9.

[36] Ascher-Svanum H, Faries D, Zhu B, et al. Medication adherence andlong-term functional outcomes in the treatment of schizophrenia inusual care. J Clin Psychiatry 2006;67:453–60.

[37] Hough D, Gopal S, Vijapurkar U, et al. Paliperidone palmitatemaintenance treatment in delaying the time-to-relapse in patientswith schizophrenia: a randomized, double-blind, placebo-controlledstudy. Schiz Res 2010;116:107–17.

[38] Gopal S, Vijapurkar U, Lim P, et al. A 52-week open-label study of thesafety and tolerability of paliperidone palmitate in patients withschizophrenia. J Psychopharmacol 2010;25:685–97.

[39] Fleischhacker W, Gopal S, Lane R, et al. A randomized trial ofpaliperidone palmitate and risperidone long-acting injectable inschizophrenia. Int J Neuropsychopharmacol 2011: [Epub ahead ofprint].

[40] Gopal S, Hough D, Xu H, et al. Efficacy and safety of paliperidonepalmitate in adult patients with acutely symptomatic schizophrenia: arandomized, double-blind, placebo-controlled, dose-response study.Int Clin Psychopharmacol 2010;25:247–56.

[41] Pandina GJ, Lane R, Gopal S, et al. A double-blind study of paliperidonepalmitate and risperidone long-acting injectable in adults withschizophrenia. Prog Neuropsychopharmacol Biol Psychiatry2011;35:218–26.

[42] Hough D, Lindenmayer J-P, Gopal S, et al. Safety and tolerability ofdeltoid and gluteal injections of paliperidone palmitate inschizophrenia. Prog Neuropsychopharmacol Biol Psychiatry2009;33:1022–31.

[43] Nasrallah H, Gopal S, Gassmann-Mayer C, et al. A controlled, evidence-based trial of paliperidone palmitate, a long-acting injectableantipsychotic, in schizophrenia. Neuropsychopharmacology2010;35:2072–82.

[44] Pandina GJ, Lindenmayer J-P, Lull JM, et al. A randomized, placebo-controlled study to assess the efficacy and safety of 3 doses ofpaliperidone palmitate in adults with acutely exacerbatedschizophrenia. J Clin Psychiatry 2010;30:235–44.

[45] Olivares J, Peuskens J, Pecanek J, et al. Clinical and resource-useoutcomes of risperidone long-acting injection in recent and long-termdiagnosed schizophrenia patients: results from a multinationalelectronic registry. Curr Med Res Opin 2009;25:2197–206.

[46] Mehnert A, Diels J. Impact of administration interval on treatmentretention with long-acting antipsychotics in schizophrenia. Presentedat the Tenth Workshop on Costs and Assessment in Psychiatry-MentalHealth Policy and Economics, March 25–27, 2011, Venice, Italy.

[47] Kissling W, Heres S, Lloyd K, et al. Direct transition to long-actingrisperidone—analysis of long-term efficacy. J Psychopharmacol2005;19:15–21.

[48] Lee M, Ko Y, Lee S, et al. Long-term treatment with long-actingrisperidone in Korean patients with schizophrenia. HumPsychopharmacol 2006;21:399–407.

[49] Lindenmayer J-P, Khan A, Eerdekens M, et al. Long-term safety andtolerability of long-acting injectable risperidone in patients withschizophrenia or schizoaffective disorder. Eur Neuropsychopharmacol2007;17:138–44.

[50] Olivares JM, Rodrigues-Morales A, Diels J, et al. Long-term outcomes inpatients with schizophrenia treated with risperidone long-actinginjection or oral antipsychotics in Spain: results from the electronicSchizophrenia Treatment Adherence Registry (e-STAR). Eur Psychiatry2009;24:287–96.

[51] Chue P, Eerdekens M, Augustyns I, et al. Comparative efficacy andsafety of long-acting risperidone and risperidone oral tablets. EurNeuropsychopharmacol 2005;15:111–7.

[52] Eerdekens M, Van Hove I, Remmerie B, Mannaert E. Pharmacokineticsand tolerability of long-acting risperidone in schizophrenia. SchizophrRes 2004;70:91–100.

[53] Weiden PJ, Olfson M. Cost of relapse in schizophrenia. Schiz Bull1995;21:419–29.

[54] Möller HJ, Llorca PM, Sacchetti E, et al. Efficacy and safety of directtransition to risperidone long-acting injectable in patients treated withvarious antipsychotic therapies. Int Clin Psychopharmacol2005;20:121–30.

[55] Ascher-Svanum H, Peng X, Montgomery W, et al. Assessing theinfrequent oral supplementation of olanzapine long-acting injection inthe treatment of schizophrenia. Eur Psychiatry 2011;26:313–9.

[56] ATC/DDD Index 2011. Available from: http://www.whocc.no/atc_ddd_index/. [Accessed May 3, 2012].

[57] Ascher-Svanum H, Montgomery WS, McDonnell DP, et al. Treatment-completion rates with olanzapine long-acting injection versusrisperidone long-acting injection in a 12-month, open-label treatmentof schizophrenia: indirect, exploratory comparisons. Int J Gen Med2012;5:391–8.

[58] Croatian Institute of Health Insurance: List of Drugs. Official GazetteNo. 69, 2010; No. 54, 2011 and No. 48 2012.

[59] Papageorgiou G, Cañas F, Zink M, Rossi A. Country differences inpatient characteristics and treatment in schizophrenia: data from aphysician-based survey in Europe. Eur Psychiatry 2011;26(Suppl.1):17–28.

[60] Consumer price index for Croatia. Available from: http://www.indexmundi.com/facts/croatia/consumer-price-index. [Accessed May 3,2012].

































http://www.ema.europa.eu/docs/en&underscore;GB/document&underscore;library/EPAR&underscore;-&underscore;Public&underscore;assessment&underscore;report/human/000890/WC500054428.pdf




http://www.ema.europa.eu/ema/index.jsp?curl=pages/medicines/human/medicines/002105/smops/Positive/human&underscore;smop&underscore;000162.jsp&murl=menus/medicines/medicines.jsp&mid=WC0b01ac058001d127













http://www.ema.europa.eu/docs/en&underscore;GB/document&underscore;library/EPAR&underscore;-&underscore;Product&underscore;Information/human/002105/WC500103317.pdf








http://www.ema.europa.eu/docs/en&underscore;GB/document&underscore;library/Referrals&underscore;document/Risperdal&underscore;Consta&underscore;30/WC500008170.pdf






www.nice.org.uk






















































































http://www.whocc.no/atc&underscore;ddd&underscore;index/

http://www.whocc.no/atc&underscore;ddd&underscore;index/












http://www.indexmundi.com/facts/croatia/consumer-price-index

http://www.indexmundi.com/facts/croatia/consumer-price-index

[61] Briggs A, Wild D, Lees M, et al. Impact of schizophrenia andschizophrenia treatment-related adverse events on quality of life:direct utility elicitation. Health Qual Life Outcomes 2008;6:105.

[62] Cummins C, Stevens A, Kisely S. The Use of Olanzapine as a First andSecond Choice Treatment in Schizophrenia. Birmingham, UK:Department of Public Health & Epidemiology, University ofBirmingham, 1998. A West Midlands Development and EvaluationCommittee Report.

[63] Lenert L, Sturley A, Rapaport M, et al. Public preferences for healthstates with schizophrenia and a mapping function to estimate utilitiesfrom positive and negative symptom scale scores. Schiz Res2004;71:155–65.

[64] Oh P, Lanctôt K, Mittmann N, et al. Cost-utility of risperidone comparedwith standard conventional antipsychotics in chronic schizophrenia. JMed Econ 2001;4:137–56.

[65] Revicki D, Shakespeare A, Kind P. Preferences for schizophrenia-relatedhealth states: a comparison of patients, caregivers and psychiatrists.Int Clin Psychopharmacol 1996;11:101–8.

[66] Johnson ML, Crown W, Martin BC, et al. Good research practices forcomparative effectiveness research: analytic methods to improvecausal inference from nonrandomized studies of treatment effectsusing secondary data sources: the ISPOR Good Research Practices forRetrospective Database Analysis Task Force report—part III. ValueHealth 2009;12:1062–73.

[67] All Wales Medicines Strategy Group. Final appraisal report: olanzapinedepot (ZypAdheras), Lilly UK. Advice No: 1510 – October 2010.Available from: http://www.wales.nhs.uk/sites3/Documents/371/olanzapine%20depot%20(ZypAdhera)%20schizophrenia.pdf. [AccessedJuly 7, 2012].

[68] Farahati F, Boucher M, Moulton K, et al. Atypical AntipsychoticMonotherapy for Schizophrenia: Clinical Review and EconomicEvaluation of First Year of Treatment (Technology report number 91).Ottawa, Canada: Canadian Agency for Drugs and Technologies inHealth, 2007.





























http://www.euro.who.int/__data/assets/ pdf_file/0018/75150/E68394.pdf

http://www.euro.who.int/__data/assets/ pdf_file/0018/75150/E68394.pdf








Cost-Utility Analysis of Pharmaceutical Care Intervention Versus UsualCare in Management of Nigerian Patients with Type 2 DiabetesMaxwell O. Adibe, BPharm, MPharm, PhD1,2,*, Cletus N. Aguwa, PharmD1,2, Chinwe V. Ukwe, BPharm, MPharm, PhD1,2

1Department of Clinical Pharmacy and Pharmacy Management, University of Nigeria, Nsukka, Enugu, Nigeria; 2Pharmacotherapeutic Group, Department ofClinical Pharmacy and Pharmacy Management, University of Nigeria, Nsukka, Enugu, Nigeria

A B S T R A C T

Objective: To assess the cost-effectiveness of pharmaceutical care(PC) intervention versus usual care (UC) in the management of type 2diabetes. Methods: This study was a randomized, controlled studywith a 12-month patient follow-up in two Nigerian tertiary hospitals.One hundred and ten patients were randomly assigned to each of the“intervention” (PC) and the “control” (UC) groups. Patients in the UCgroup received the usual/conventional care offered by the hospitals.Patients in the PC group received UC and PC in the form of structuralself-care education and training for 12 months. The economic evalu-ation was based on patients’ perspective. Costs of management ofindividual complications were calculated from activities involved intheir management by using activity-based costing. The impact of theinterventions on quality of life was estimated by using the HUI23-S4EN.40Q (Mark index 3) questionnaire. The primary outcomes wereincremental cost-utility ratio and net monetary benefit. An intention-to-treat approach was used. Two-sample comparisons were made byusing Student’s t tests for normally distributed variables data at

baseline, 6 months, and 12 months. Comparisons of proportions weredone by using the chi-square test. Results: The PC intervention led toincremental cost and effect of Nigerian naira (NGN) 10,623 ($69) and0.12 quality-adjusted life-year (QALY) gained, respectively, with anassociated incremental cost-utility ratio of NGN 88,525 ($571) perQALY gained. In the cost-effectiveness acceptability curve, the prob-ability that PC was more cost-effective than UC was 95% at the NGN250,000 ($1613) per QALY gained threshold and 52% at the NGN 88,600($572) per QALY gained threshold. Conclusions: The PC interventionwas very cost-effective among patients with type 2 diabetes at theNGN 88,525 ($571.13) per QALY gained threshold, although consider-able uncertainty surrounds these estimates.

Keywords: cost-effectiveness analysis, cost-utility analysis, Nigeria,patients with type 2 diabetes, pharmaceutical care, usual care.


Introduction

Analytic techniques used for economic evaluation in health care,for example, cost-benefit analysis, cost-effectiveness analysis,and cost-consequences analysis, are designed to compare alter-native courses of action in terms of costs and outcomes. Thechoice of the technique depends on the decision the healtheconomists intend to influence. Quality-adjusted life-years(QALYs) measure health as a combination of the duration of lifeand the health-related quality of life [1]. The primary outcome ofa cost-utility analysis is the cost per QALY, or incremental cost-utility ratio (ICUR), which is calculated as the difference in theexpected cost of two interventions divided by the difference inthe expected QALYs produced by the two interventions. Theresults of a cost-utility analysis are compared with a thresholdincremental cost-effectiveness ratio (ICER); interventions with anICER below this threshold are funded, whereas those with anICER above the threshold tend not to be. Economic evaluations

using QALYs as the principal measure of outcome, often termedcost-utility studies, have become increasingly popular in theliterature and have also been adopted by a number of healthtechnology assessment agencies as the methodology of choice[1].

Cost-utility analysis was developed to help decision makerscompare the value of alternative interventions that have verydifferent health benefits, and it facilitates these comparisonswithout recourse to placing monetary values on different healthstates. Cost-utility analysis specifies what value is attached tospecific health states, and thus increasingly facilitates the trans-parency of resource allocation processes [2].

Cost-utility analysis was developed to address the problem ofconventional cost-effectiveness analysis, which did not allowdecision makers to compare the value of interventions for differ-ent health problems. The utilities can now be obtained fromstandardized and validated health status instruments, makingthe evidence required to inform cost-utility analysis relatively





E-mail: [email protected]; [email protected].

� Address correspondence to: Maxwell O. Adibe, Department of Clinical Pharmacy and Pharmacy Management, University of Nigeria,Nsukka, Enugu, Nigeria.


dx.doi.org/10.1016/j.vhri.2013.06.009

dx.doi.org/10.1016/j.vhri.2013.06.009

dx.doi.org/10.1016/j.vhri.2013.06.009



dx.doi.org/10.1016/j.vhri.2013.06.009

straightforward and cheap to acquire—certainly when comparedwith the cost of acquiring evidence on clinical effectiveness, andindeed the cost of many of the treatments being reviewed [3].

Diabetes mellitus (DM) is associated with considerable morbid-ity andmortality [4]. It is also a major risk factor for cardiovasculardisease, stroke, and kidney failure [5]. In Africa, DM probably hasthe highest morbidity and mortality rates of all chronic non-infective diseases [6].

DM was once regarded as a disease of the affluent, but it isnow vastly visible as a growing health problem in developingeconomics because almost 80% of diabetes deaths occur in low-and middle-income countries [7,8]. The national standardizedprevalence rate of DM in Nigeria is 2.2%, while the crudeprevalence rate is 7.4% in those aged 45 years and above wholive in urban areas [9]. Global estimates of the prevalence ofdiabetes showed that the prevalence of diabetes in Nigeria in2010 was 4.7% (vs. 3.9% for world population) and that it would be5.5% (vs. 4.3% for world population) in 2030 [10].

With the increasing demand for better management of type 2diabetes, attention has focused on the potential benefits of pharma-ceutical care (PC) to improve patients’ health outcomes. Many PCprograms have been established in various countries to enhanceclinical outcomes and the health-related quality of life. Theseprograms were implemented by pharmacists, with the cooperationof physicians and other health care professionals. PC and theexpanded role of pharmacists are associated with many positivediabetes-related outcomes, including improved clinical measures[11], improved patient and provider satisfaction [12,13], andimproved cost of management [12,14]. The pharmacists can, there-fore, in collaboration with physicians and other health care profes-sionals, contribute to an improvement in the quality of life ofpatients with diabetes by informing and educating patients, answer-ing their questions, and, at the same time, monitoring the outcomesof their treatment [15]. In view of the above issues, the objective ofthis study was to assess the cost-effectiveness of the PC interventionin the management of type 2 diabetes versus usual care (UC).

Methods

Study Design

This study was a randomized, controlled, and longitudinal pro-spective study with a 12-month patient follow-up. The studyfollowed the Consolidated Health Economic Evaluation ReportingStandards guideline for reporting economic evaluation of inter-ventions [16]. The study protocol was approved by the ResearchEthical Committees of the University of Nigeria Teaching Hospi-tal, Ituku Ozalla, and Nnamdi Azikiwe University TeachingHospital, Nnewi, in which this study was conducted. Thesehospitals are tertiary hospitals that serve as referral centers tomost of the hospitals in the southeastern part of Nigeria.

Inclusion Criteria

Patients with type 2 DM who fulfilled the entrance criteria wereidentified and included in the study. Inclusion criteria includedpatients with type 2 diabetes who were on oral hypoglycemictherapy and provided written informed consent in addition towillingness to abide by the rules of the study and being certifiedfit to take part by the consulting physician.

Exclusion criteria were patients who were diagnosed with type 1diabetes (to avoid complexity in the scope of the study), patients whowere younger than 18 years (they are legally regarded as dependentsand consequently they cannot take decisions of their own), patientswho were pregnant (they are generally not allowed to participate in astudy of this nature by the institutions used for the study), and

patients who expressed willingness to withdraw from the study(participation is voluntary). The sample size determination showedthat a sample size of at least 104 patients was required in each of thecontrol and intervention groups [17]. Based on these data, to ensuresufficient statistical power and to account for “dropouts” during thestudy, a target sample size of 220 patients was recruited (110 patientsfrom each of the hospitals). The folders of the 110 selected patients ineach hospital were assigned numbers 1 to 110, which represented anindividual patient. Patients were randomly assigned to one of twogroups (intervention group or control group) on the basis of thenumber assigned to their folders by using online “random sequencegenerator” [18] with sequence boundaries of 1 to 110 (boundariesinclusive) set in a two-column format: the first column was a prioridesignated to the intervention group PC (55 patients) and the secondcolumn to the control group UC (55 patients).

Patients in the UC group received the usual/conventional careoffered by the hospitals, which included hospital visits on appoint-ment or on a sick day, consultations with physicians, prescription ofdrugs and routine laboratory tests, review of diagnosis and medi-cations, refilling of prescriptions by patients, and referral. This UCwas offered with education/training of the patients in an uncoordi-nated manner and without structured educational materials.Patients in the PC group received UC and PC for 12 months onmonthly schedule. This additional PC included a stepwise approach:setting priorities for patient care, assessing patients’ specific educa-tional needs and identification of drug-related problems, develop-ment of a comprehensive and achievable PC plan in collaborationwith the patient and the physician, implementation of this plan, andmonitoring and review of the plan from time to time [19]. Thenurses collaborated with the pharmacists in terms of organizing thepatients and patients’ folders, taking point-of-care testing, counsel-ing the patients, and reinforcing the information given to thepatients during training sections. The physicians provided thevisitation/appointment schedule for the patients, and prescriptionof laboratory tests. They were also involved in the implementationof consensus strategies in managing drug-related problems in areasof changing, substitution, and withdrawal of medications. All themembers of the health care team were trained before the imple-mentation of the intervention.

The medical and educational contents of the training materialswere evaluated by the physicians and nurses in diabetes clinicsbefore the researchers conducted the training for the patients. Thephysicians and nurses were asked to rate the materials as beingexcellent, very good, good, fair, poor, and useless.

The monthly educational/training program for the patientsconsisted of four sections of 90 to 120 minutes. The programcovered the following areas: diabetes overview and its complica-tions, self-monitoring blood glucose techniques and interpreta-tion of diabetes-related tests, medications and their side effects,lifestyle modification, counseling, and effective interaction withhealth providers. PC provided ground for the patients to monitorand react to changes in their blood glucose levels, allowing themto integrate their diabetes into the lifestyle they preferred.

Data Collection

Data were collected on utilization of health care resources for 12months for control and intervention groups at baseline, 6 months,and 12 months. Information was obtained on the frequency ofself-monitoring, number and average duration of visits to ahospital, daily doses of drugs taken regularly, and the variable of“other health care resource use,” including primary care (generalpractitioner and nurse consultations), hospital care (visits to anaccident and emergency department, outpatient care, day hospitalcare, and inpatient care), auxiliary health care (services of apodiatrist, optician, or dietitian), and private health care. Thesedata were collected by means of patients’ PC diaries notes


supplemented when necessary by information from patients’medical records.

Patient-specific data on the incremental resources requiredfor intervention and control groups and the resources associatedwith the treatment of complications were taken. Treatmentresources included doses of drugs used for treating diabetes,antihypertensive drugs, other drugs, blood-glucose tests, self-monitoring resources such as test strips, lancets, and glucome-ters, and visits to general practitioners, practice nurses, andclinics. Resources associated with complications included thenumber, duration, and specialty of admissions to hospital; out-patient consultations; medical procedures such as photocoagu-lation and cataract extraction; and day case episodes. The unitcost of these resource volumes (drugs and other consumables,laboratory tests, and specialty care per visit) was derived by usingthe National Health Insurance Scheme price list [20] and Interna-tional Drug Price Indicator Guide 2010 edition [21], and the cost ofall inpatient bed-days and outpatient visits was derived fromWorld Health Organization-Choosing Intervention that is Cost-Effective (WHO-CHOICE) [22] unit-cost estimates. Costs of themanagement of individual complications were calculated fromactivities involved in their management by using the ingredientapproach or activity-based costing as opined by the experts; allcosts were adjusted to 2011 cost [23]. The effect of either a higheror lower adjustment rate was examined in the sensitivity anal-ysis. All costs are reported in year 2011 values of Nigerian naira(NGN 155 ¼ $1).

Each item for resource use was categorized into the “cost of theintervention,” the “cost of drugs,” and the “cost of other healthcare resource use” (including primary care, hospital care, andauxiliary health care). The costs were calculated by multiplyingthe volume of resource use in each category by the associated unitcost in 2010 prices (Table 1). Average costs were estimated in eacharm of the study for the 12 months of follow-up Table 2.

The impact of the interventions on quality of life wasestimated by using the HUI23S4EN.40Q (developed by HUInc -Mark index 2&3) questionnaire at baseline, 6 months, and 12months in accordance with the HUI procedures manual (HUI23-S4EN.40Q, HUI23-40Q.MNL) [24]. We adopted the QALY [2]because this measure captures both increases in life expectancyand improved quality of life that results from the prevention ofcomplications, providing a composite outcome measure of fataland nonfatal events that permits comparison between manyhealth interventions.

Because the economic evaluation perspective was that of thehealth care purchaser, only direct health service costs wereincluded. These included treatment costs, visits to a nurse or ageneral practitioner based on “standard practice” assumptions,and costs of treating diabetes complications. Not included in thisanalysis were nonmedical costs such as out-of-pocket expensesincurred when visiting clinics, cost of informal care provided byfamily members, and production losses resulting from workabsences, long-term disability, or premature death.

Statistical Analysis

Statistical analyses were performed by using the SPSS package,version 14 (SPSS, Inc., Chicago, IL). An intention-to-treat approachwas used. Data were summarized as means � SD, mean differ-ences with 95% confidence intervals. Two-sample comparisonswere made by using Student’s t tests for normally distributedvariables or Mann-Whitney U tests for nonnormally distributeddata (0, 6, and 12 months). Comparisons of proportions werecarried out by using chi-square, Fisher’s exact, or McNemar’stests. An a priori significance level of P less than 0.05 was usedthroughout. Based on the overall health-related quality-of-lifescore for the patients at baseline, 6 months, and 12 months,

QALYs were determined. Areas under the curves were deter-mined by using WinNonlin standard edition version 2.1 [3,22].

Sensitivity analysisTo address uncertainty around the ICUR, univariate sensitivityanalysis was conducted, where one cost variable was varied at atime (upper and lower limits) while keeping all other variablesconstant at their mean base-case cost. Then, two alternative-caseoutcomes of ICUR were generated on the basis of upper and lowerboundaries of �20% of the mean base-case cost.

To assess how a simultaneous change in several variables(QALYs, total intervention cost, cost of antidiabetes medications,cost of antihypertensives, cost of antidiabetes antihypertensivesmedications, total cost of drugs, hospital care cost, auxiliaryhealth care cost) affects the cost-utility ratio, a Monte-Carlosimulation (a type of multivariate sensitivity analysis) wasperformed. This technique runs a large number of simulations(here 1000) by repeatedly drawing samples from probabilitydistributions of input variables. Thus, it provides a probabilitydistribution of the output variable; that is, QALYs, incrementalcosts, incremental effectiveness, and ICURs. Beta and gammadistributions were assumed for utility (QALYs) and unit cost,respectively [25–27].

Given that the interpretation of negative ICURs is ambiguous,the ICURs were transformed into net monetary benefits (NMBs).The decision rule used was to adopt the intervention in questionif the NMB is greater than zero. Given that the appropriate valueof λ is unknown, λ was varied from NGN 0 to NGN 450,000. A cost-effectiveness acceptability curve was generated on the basis ofthe distribution of NMB for each λ. A cost-effectiveness accept-ability curve allows a decision maker to consider whether anintervention (PC) is cost-effective in relation to the maximumamount a decision maker is willing to pay for a QALY. A discountrate of 3% and 6% was used in sensitivity analysis [28].

At each ceiling value for the willingness to pay for a QALY, thecost-effectiveness curve shows the probability that the treatmentis cost-effective. All calculations were done in Microsoft Excel2007 (Microsoft Corporation, Redmond, WA).

Results

Economic Outcomes

The medical and educational content of the training course wasrated positively by the 17 physicians and 29 nurses: the majority38 (82.6%) rated the content as “excellent” and the remaining 8rated the content as “very good” or “good”; only 3 (6.5%) of themsuggested little modification or changes.

The number of patients who completed the study and whosedata were analyzed at 6 months and 12 months in UC and PCarms were 98 (89.09%) versus 102 (92.73%) and 93 (84.55%) versus99 (90.0%), respectively.

The general cost of care/laboratory cost per patient for UCversus PC at 12 months was NGN 16,519 � 7,905 ($107 � $51)versus NGN 17,369 � 6,673 ($112 � $43), P ¼ 0.4208. PC-specificcost per patient was NGN 7,345 � 2,651 ($47 � $17), while thecosts of antidiabetes medications for UC and PC arms were NGN9,703 � 4,632 ($63 � $30) and NGN 7,808 � 4,183 ($50 � $27), P ¼0.0033, respectively. The cost of antihypertensives for UC wasNGN 6,625 � 4,691 ($43 � $30) while that of PC was NGN 5,155 �

2,619 ($33 � $17), P ¼ 0.0228. The cost of antidiabetes medicationsplus antihypertensives for UC was NGN 16,328 � 5,086 ($105 �

$33) as against NGN 12,963 � 7,549 ($84 � $49) for PC, P ¼ 0.0004.The cost of other medications was NGN 3,243 � 2,637 ($21 � $17)and NGN 4,945 � 1,687 ($32 � $11), P o 0.0001, for UC and PC,respectively.


Table 1 – Categories of resources used and their cost sources.

Cost centers Unit cost in NGN (cost used in study) Source

InterventionMeter (Acu-check Active) 3,300 Market priceTest strips/50 strips 2,200 Market priceLancet/200 600 Market priceBP apparatus (Aneroid sphygmomanometer and stethoscope) 2,900 Market priceTraining/educational materials 800 Market price

Laboratory testsHb A1c 1,700 UNTHFasting blood glucose 250 [20]Liver function test 700 [20]HDL 400 [20]LDL 700 [20]Triglyceride 600 [20]Total cholesterol 300 [20]Others 34,000 [20]

DrugsOral antidiabetes drugs per tablet� 15–60� [20,21]Human insulin per vial 3,200 [20,21]Insulin soluble per vial 950 [20]Insulin zinc per vial

Suspension (insulin zinc suspension)850 [20]

Antihypertensives (tablet, injection, injection powder, syrup)� 5–1,400� [20,21]Others (tablet, injection, injection powder, syrup)� 5–280� [20,21]

Hospital care (per episode)Emergency care 8,000–20,000 (12,000)† Experts’ opinion (UNTH/NAUTH)Outpatient care (tertiary hospital) 2,149.50 [22]Inpatient (per day) (tertiary hospital) 4,404 [22]

Primary careGeneral practitioner consultation 700 [20]Nurse consultation 700 [20]

Auxiliary health care (per session)Dietician 700 [20]Optician 700 [20]Podiatrist 700 [20]Others 700 [20]

Note. NGN 155 ¼ $1.BP, blood pressure; Hb A1c, hemoglobin A1c; HDL, high-density lipoprotein; LDL, low-density lipoprotein; NAUTH, Nnamdi Azikiwe University Teaching Hospital; NGN, Nigerian naira; NHIS,National Health Insurance Scheme; UNTH, University of Nigeria Teaching Hospital.� When the individual drug was not in the NHIS price list, the International Drug Price Indicator Guide 2010 edition was used; the total cost of drug category was presented because many drugswere encountered. The ranges of their prices are represented.

† The price in parentheses was used.

VALUE

INH

EALTH

REGIO

NAL

ISSUES

2(2

013)189–198

192

The total drug cost was NGN 19,571 � 7,514 ($126 � $49) for UCas against NGN 17,908 � 8,549 ($116 � $55) for PC, P ¼ 0.1549. Thehospital care costs for UC and PC were NGN 10,302 � 5,657 ($67 �

$37) and NGN 9,766 � 4,234 ($63 ± $27), P ¼ 0.4565, respectively,while their auxiliary health care costs were NGN 4,060 � 1,675 andNGN 8,687� 2,365 ($56 ± $15), Po 0.0001, respectively. The total costper patient per year was NGN 50,452 � 35747 ($326 ± $231) for UCand NGN 61,075 � 43763 ($394 ± $282), P ¼ 0.1009, for PC (Table 3).

Cost-effectiveness or cost-utility was NGN 78524.51 ($507) perQALY for UC and NGN 80,098.36 ($517) per QALY for PC, while theincremental cost and incremental QALY were NGN 10,623 ($69) and0.12, respectively. Thus, the ICUR was NGN 88,525 ($571) per QALY.

Sensitivity Analysis

The cost-effectiveness plane that was obtained from a MonteCarlo simulation with 1000 iterations showed that 93.8% of the

simulations were within the northeast quadrant, where the PCintervention resulted in gain in QALY and cost, whereas 5.6% ofthe simulations were in the southeast quadrant, where the PCintervention resulted in gain in QALY and reduced cost. Only 0.5%of the simulations were within the northwest quadrant, wherethe addition of PC resulted in loss in QALY and increased cost.The 1000 iterations produced an incremental QALY that rangedfrom −0.022 to 0.293 and an incremental cost that ranged fromNGN −8,276.40 to NGN 28,294.27 (Fig. 1).

The mean NMB within a willingness to pay of NGN 0 to NGN450,000 was greater in the PC intervention whatever the willing-ness to pay was. This result also revealed that 90% of PCcredibility interval was far above the mean of UC though theinterval overlapped with about 5% of the UC (Fig. 2).

The PC intervention led to incremental cost and incrementalQALY/effect of NGN 10,623 and 0.12 QALY gained, respectively,with an associated ICUR of NGN 88525 per QALY gained. The

Table 2 – Baseline characteristics of the patients in PC and UC arms.

Demographic data UC (n ¼ 110) PC (n ¼ 110) P

Mean age � SD (y) 52.8 � 8.2 52.4 � 7.6 0.708Grouped age: 453 y, n (%) 81 (73.64) 75 (68.18) 0.373Sex: male, n (%) 49 (44.55) 44 (40) 0.495Level of education, n (%) 0.406Primary school 3 (2.72) 6 (5.45)Secondary school 71 (64.55) 63 (57.27)University 36 (32.73) 41 (37.27)

Marital status, n (%) 0.409Currently married 37 (33.64) 46 (41.82)Widowed 71 (64.54) 63 (57.27)Single 2 (1.82) 1 (0.91)

Occupation, n (%) 0.611Self-employed 37 (33.64) 34 (30.91)Employee 35 (31.82) 42 (38.18)Retired 38 (34.54) 34 (30.91)

Smoking status: smoker, n (%) 34 (30.91) 21 (19.09) 0.043�

Duration, mean � SD 4.5 � 2.2 4.8 � 2.8 0.378Duration: ≥5 y, n (%) 62 (56.36) 71 (64.55) 0.215Family history of diabetes, n (%) 71 (64.55) 62 (56.36) 0.214Physical activity/exercise, n (%) 18 (16.36) 23 (20.91) 0.387ComorbiditiesHypertension 60 (54.55) 73 (66.36) 0.073Congestive heart failure 11 (10.00) 15 (13.64) 0.404Ischemic heart disease 7 (6.36) 8 (7.27) 0.789Arthritis 37 (33.64) 43 (39.09) 0.400

≥2 comorbidities, n (%) 72 (65.45) 81 (73.64) 0.187Overnight hospitalization, n (%) 9 (8.18) 7 (6.36) 0.604Emergency room, n (%) 1 (0.91) 2 (1.82) 0.561Use of insulin, n (%) 17 (15.45) 13 (11.82) 0.432Antidiabetic medications, n (%) 103 (93.64) 107 (97.27) 0.195Other medications, n (%)Daily aspirin 43 (39.09) 57 (51.82) 0.058Diuretics 71 (64.55) 84 (76.36) 0.055Antihypertensives 98 (89.91) 78 (70.91) 0.0007�

Lipid-lowering 23 (20.91) 14 (12.73) 0.105Complications, n (%)Myocardial infarction 2 (1.82) 4 (3.64) 0.408Stroke 9 (8.18) 6 (5.45) 0.422Foot ulcer 2 (1.82) 3 (2.73) 0.651Blindness 1 (0.91) 1 (0.91) 1.000Renal failure 3 (2.73) 8 (7.27) 0.122

NGN, Nigerian naira; PC, pharmaceutical care; UC, usual care.�P o 0.05.


QALY value of PC was varied by �10% (0.68625 and 0.83875) in thesensitivity analysis, which made the ICUR moved from NGN88,525 per QALY gained to upper and lower limits of NGN252,173 per QALY gained and NGN 54,777 per QALY gained,respectively. When a 3% and 6% adjustment rate of costs wasapplied, the lower and upper extremes of ICURs were NGN 85,936per QALY gained and NGN 88,799 per QALY gained from the basevalue of NGN 88,525 per QALY gained.

Varying the base costs by �20% (lower limit to upper limit) andusing these extreme values in simulations, the total interventioncosts center produced ICURs that ranged from NGN 47,426.41 toNGN 126,544.5 per QALY gained. The antidiabetes cost centerproduced ICURs that ranged from NGN 74,817.97 to NGN104,206.6 per QALY gained. The ICURs moved from NGN 79,210.06to NGN 98,038.14 per QALY gained when the base cost of theantihypertensives cost center was varied. The antidiabetes medi-cations plus antihypertensives produced ICURs that ranged fromNGN 70,020.39 to NGN 109,161.3 per QALY gained, and total drugcost center produced ICURs that ranged from NGN 58,950.02 toNGN 120,469.9 per QALY gained. The ICURs of hospital care costand auxiliary health care centers ranged from NGN 72,582.82 to106,161.6 per QALY gained and NGN 74,454.77 to 105,414.0 perQALY gained, respectively, when base costs of the cost centerswere varied. The ICUR was most sensitive to variation in QALY and“total intervention-specific cost center” variable followed by that oftotal drug cost center (Fig. 3). In the cost-effectiveness acceptabilitycurve, the probability that PC was cost-effective versus UC was 95%at the threshold of NGN 250,000 per QALY gained and 52% at thethreshold of NGN 88,600 per QALY gained (Fig. 4).

Discussion

Quality-Adjusted Life-Years

QALYs associated with PC were significantly higher than thoseassociated with UC after 12 months. This indicates that extend-ing this study beyond 1 year could offer more benefits to patientswith diabetes in terms of QALYs gained. Some studies haddemonstrated that extension of PC beyond 1 year could offerextra benefits to patients with diabetes [29,30].

At the end of this period, the PC intervention resulted in anincremental gain in QALYs and cost compared with the UC. Thiseconomic evaluation demonstrates that PC is the most cost-effective strategy for managing patients with type 2 diabetes ifthe patients are willing to pay at least NGN 88,600 per QALYgained. The addition of the PC intervention to UC, as noted in thisstudy, should be considered a highly cost-effective managementoption for patients with type 2 diabetes because treatmentscosting no more than £20,000 (NGN 4,761,905) to £30,000 (NGN7,142,857) per QALY gained are generally considered to be cost-effective [1,31]. This PC intervention also generated greater NMBswhen compared with UC; therefore, the addition of PC to UCmight be considered an appropriate management option forpatients with diabetes who have comorbidities where the prob-ability or likelihood of drug-related problem is higher.

Cost-Effectiveness Plane

The cost-effectiveness plane showed that most of the simula-tions were within the northeast quadrant, where the addition ofPC to UC resulted in gain in QALY and cost, which indicated thatalthough the PC intervention generated more QALYs than did UC,it was more costly. The 1000 iterations showed that the 95%confidence interval of incremental QALYs and incremental costwas wide. This wide range shows that there are uncertaintiessurrounding both QALYs and cost. The magnitude of QALYsgained, specific intervention cost, and cost of all drugs werefound to have affected the ICUR most. This provides avenues forurgent intervention to reduce the cost of drugs used for themanagement of diabetes and its comorbidities and an urgentinstitution of intervention that will improve the quality of life ofpatients with diabetes in Nigerian tertiary hospitals.

Net Monetary Benefit

The NMB approach provides a useful mechanism for identifyingwhich arm of the study is most cost-effective. The NMB of additionalPC over UC alone for a willingness to pay of NGN 0 to NGN 450,000,the additional PC alternative, was associated with the greater meanNMB whatever the willingness to pay was. It is interesting to notethat the lowest trough (NMB) of PC was far higher than the mean ofUC NMB. Addition of the PC intervention was found to be superior toUC alone in all willingness to pay, even as the willingness to pay

Table 3 – Costs and QALY per patient per year at the end of the 12-mo follow-up period for UC versus PC (NGN155 ¼ $1).

Cost per patient per year UC PC P Mean costdifference

95% CI perpatient

General intervention and laboratorycost

16,519 � 7,905 17,369 � 6,673 0.4208 850 −1,228.1 to 2,928.1

Specific intervention cost for PC NA 7,345 � 2,651 NA NA NATotal cost of intervention 16,519 � 7,905 24,714 � 11,655 o0.0001� 8,195 5,341.9–11,048Antidiabetes medications 9,703 � 4,632 7,808 � 4,183 0.0033� −1,895 −3,150.1 to −639.92Antihypertensives 6,625 � 4,691 5,155 � 2,619 0.0228� −1,470 −2,733.7 to −206.34Antidiabetes medications plus

antihypertensives16,328 � 5,086 12,963 � 7,549 0.0004� −3,365 −5,209.3 to −1,520.7

Other medications 3,243 � 2637 4,945 � 1,687 o0.0001� 1,702 1,074.7–2,328.3Total drug cost 19,571 � 7514 17,908 � 8,549 0.1549 1,663 −3,960.2 to 634.16Hospital care cost 10,302 � 5,657 9,766 � 4,234 0.4565 −536 −1,952.9 to 880.88Auxiliary health care cost 4,060 � 1,675 8,687 � 2,365 o0.0001� 8,627 8,040.2–9,213.80Total cost per patient 50,452 � 35,747 61,075 � 43,763 0.1009 10,623 −2,088.5 to 23,335QALY per patient per year 0.6425 � 0.13 0.7625 � 0.15 o0.0001� 0.1200 (0.07–0.1601)

CI, confidence interval; NA, not applicable; PC, pharmaceutical care; QALY, quality-adjusted life-year; UC, usual care.� P ≤ 0.05; Negative cost differences indicate cost savings associated with the PC intervention.


increases. This result shows that if the willingness to pay is in therange of NGN 0 to NGN 450,000, there is a net monetary gain orsaving of NGN 56,148 in 1 year. This amount is more than 3 monthssalaries of a Nigerian low-income earner based on the current NGN18,000 minimum wage [32]. Therefore, there is need for introductionand exploitation of the PC intervention in Nigerian health facilitiesbecause this is very cost-effective with enormous NMB.

Sensitivity Analyses

This study found out that a little variation in QALY gained in PCto the tune of �10% resulted in a tremendous increase and mildlowering of the base ICUR, respectively. This result showed that

the economic burden placed on patients with diabetes by 10%health deficit was enormous; therefore, interventions such as PCthat would be aimed at improving the quality of life of patientsand resolution/reduction of drug-related problems that wouldultimately reduce the cost of drugs would certainly reduce thecost per QALY associated with diabetes.

These results indicate that the additional PC intervention has acost per QALY gained that is lower than that of UC. In the UnitedKingdom, interventions appear to have a high chance of acceptanceby the National Institute for Clinical Excellence if their cost-effectiveness is more favorable than approximately £30,000 perQALY [1]. Several other studies had classified cost-effectiveness.WHO-CHOICE classified interventions on the basis of the level of

150000

200000

250000

300000

350000

400000

0 50000 100000 150000 200000 250000 300000 350000 400000 450000

Net

Mon

etar

y Be

nefit

s (N

GN

)

Willingness to pay (NGN)

PC

UC

Fig. 2 – Net monetary benefit of PC and UC at different levels of willingness to pay for a QALY. NGN, Nigerian naira; PC,pharmaceutical care; QALY, quality-adjusted life-year; UC, usual care.

-15000.000

-10000.000

-5000.000

0.000

5000.000

10000.000

15000.000

20000.000

25000.000

30000.000

35000.000

-0.050 0.000 0.050 0.100 0.150 0.200 0.250 0.300 0.350

Incr

emen

tal c

ost

(NG

N)

Incremental QALYs

Fig. 1 – Cost-effectiveness plane. NGN, Nigerian naira; QALYs, quality-adjusted life-years.


cost-effectiveness by convention as described in the literature [33–35]. An intervention is cost saving when the intervention generates abetter health outcome and costs less than the comparison inter-vention. The intervention is cost neutral if the ICER is 0. Theintervention is very cost-effective when the ICER is more than 0 or$25,000 or less per QALY or life-year gained (LYG) while theintervention is cost-effective when the ICER is between more than$25,000 to $50,000 per QALY or LYG. The intervention is marginallycost-effective when the ICER is between more than $50,000 and$100,000 per QALY or LYG, whereas an intervention is said to be notcost-effective when the ICER is more than $100,000 per QALY or LYG.

WHO-CHOICE published in 2005 the cost-effectiveness thresh-old for different regions of the world. WHO-CHOICE suggested acost-effectiveness threshold based on gross domestic product(GDP) per capita. An intervention that produces cost per QALYgained of less than GDP per capita of the country is said to be verycost-effective while an intervention with cost per QALY gained ofbetween one to three times the GDP per capita of the country iscost-effective. An intervention with cost per QALY gained ofgreater than three times the GDP per capita of the country isnot cost-effective. For AFRO D where Nigeria belongs, the cost-effectiveness threshold ranges from $1,695 to $5,086 [7]. With aconversion factor of NGN 155 ¼ $1, the threshold ranges from

NGN 262,725 to NGN 788,330. World Bank in 2010 published a GDPper capita, considering purchasing power parity-current interna-tional $; for Nigeria, it is $2,381 (NGN 369,055:00) [34]. Theassociated ICUR from this study was NGN 88,525 per QALY($571.13/QALY) gained in the PC arm, which was far lower thanthe GDP per capita of Nigeria in 2010 [36].

Based on the above facts, the PC intervention with an ICUR ofNGN 88,525 per QALY gained is very cost-effective although thismay still not be affordable for low-income earners in relation to theNGN 18,000 minimum wage approved in Nigeria in 2011 becauseNGN 88,525 is about 5 months’ salary of this group of Nigerians [32].

In probabilistic sensitivity analysis, based on the cost-effectiveness acceptability curve, the PC dominated UC at thethreshold of NGN 88,600 per QALY gained and the probability ofPC being more cost-effective approached 95% at the threshold ofNGN 250,000 per QALY gained. Nevertheless, if a patient is willingto pay NGN 400,000 per QALY gained, the probability that PC isthe most cost-effective option for managing patients with dia-betes increases to 97%. In contrast, the probability that UC is themost cost-effective option at the threshold of NGN 400,000 perQALY gained approaches zero.

Studies of this kind must address inherent potential threatsto internal validity [37,38]. The major limitations of this study

54,777

85,936

47,426

74,81879,210

70,020

58,950

72,583 74,455

252,173

88,799

126,545

104,20798,038

109,161

120,470

106,162 105,414

0

50,000

100,000

150,000

200,000

250,000

300,000

Cost

Per

QA

LY (N

GN

)

Variables

Fig. 3 – Univariate sensitivity analysis of cost and utility variables on incremental cost-effectiveness ratio. NGN, Nigerian naira;QALY, quality-adjusted life-year. A ¼ QALY (�10%); B ¼ Adjustment rate (3% and 6%); C ¼ Total intervention-specific cost (�20%);D ¼ Cost of antidiabetes medications (�20%); E ¼ Cost of antihypertensives (�20%); F ¼ Cost of antidiabetes + antihypertensivesmedications (�20%); G ¼ Total cost of drugs (�20%); H ¼ Hospital care cost (�20%); I ¼ Auxiliary health care cost (�20%).


were missing data, selection bias, short period of study, attrition,and consideration of only direct cost. Data on humanisticoutcome measures were self-reported; however, self-reporteddata about diabetes status have been established to be bothvalid and reliable [39]. We recommend that future researchstudies of this kind address these limitations. This pharmaceut-ical intervention could be adopted for patients suffering fromother chronic diseases such as HIV, hypertension, asthma,psychosis, epilepsies, and cerebrovascular and cardiovasculardiseases.

Conclusions

The PC intervention was very cost-effective among patients withtype 2 diabetes at the NGN 88,525 ($571.13) per QALY gainedthreshold, although considerable uncertainty surrounds theseestimates. This study also revealed that cost incurred and QALYsgained by patients in the PC group were higher than those of theircounterparts in the UC group. This indicates that the extra costpaid for extra QALYs gained is worth it because it saves futureexpenditures and improves the quality of life of patients.

The results of this study illustrate a convincing economic ration-ale for improving standards of care for patients with type 2 diabetesthrough the PC intervention. This study provides further evidencethat the cost-effectiveness of interventions to reduce the burden ofdiabetes-related complications compares favorably with that of otheraccepted uses of health care resources. The results should be ofinterest and used by other economists and health service researchers,and in particular should be considered by decision makers whenconsidering the allocation of resources to diabetes care.

Cost-utility analysis thus increasingly facilitates the trans-parency of resource allocation processes. The usefulness of cost-utility analysis to decision makers explains the rapid expansionin the utilization of cost-utility analysis over the last decade.

Acknowledgment

We acknowledge Health Utility Incorporated for granting andawarding us HUI23S4En.40Q and HUI23.40Q.MNL.

Source of financial support: Funding for this project wasprovided from Science and Technology Education Post Basic(STEP-B) through the University of Nigeria. The views expressedin this article are those of the authors, and no official endorse-ment by STEP-B is intended or should be inferred.

R E F E R E N C E S

[1] Rawlins M, Culyer A. National Institute for Clinical Excellence and itsvalue judgments. BMJ 2004;329:224–7.

[2] Torrance G. Measurement of health state utilities for economicappraisal: a review. J Health Econ 1986;5:1–30.

[3] Drummond M, O’Brien B, Stoddart G, Torrance G. Methods for theEconomic Evaluation of Health Care Programmes. (2nd ed.). Oxford:Oxford University Press, 1997.

[4] Akanji AO, Adetunji A. The pattern of presentation of foot lesions inNigerian diabetic patients. West Afr J Med 1990;9:1–4.

[5] American Diabetes Association. National Diabetes Fact Sheet. Availablefrom: http://www.diabetes.org/main/info/facts/factsna tl.jsp. [AccessedDecember 15, 2002].

[6] McLarty DG, Swai ABM, Kitange HM, et al. Prevalence of diabetes andimpaired glucose tolerance in rural Tanzania. Lancet 1989;1:871–5.

[7] International Diabetes Federation. International Diabetes Federation:diabetes atlas. Available from: http://da3.diabetesatlas.org/newsc269.html. [Accessed November 15, 2011].

[8] Odili VU, Ugboka LU, Oparah AC. Quality of life people with diabetes inBenin City as measured with WHOQoL-BREF. Internet J Law HealthcareEthics 2010;6(2). http//dx.doi.org/10.5580/18a.

[9] Nyenwe E, Odia O, Ihekwala A, et al. Type 2 diabetes in adult Nigerians:a study of its prevalent and risk factors in Port Harcourt, Nigeria.Diabetes Res Clin Pract 2003;62:177–85.

[10] Shaw J, Sicree R, Zimmet P. Global estimates of the prevalence ofdiabetes. Diabetes Res Clin Pract 2010;87:4–14.

[11] Jaber L, Halapy H, Fenret M, et al. Evaluation of a pharmaceutical caremodel on diabetes management. Ann Pharmacother 1996;30:238–43.

[12] Sadur C, Moline N, Costa M, et al. Diabetes management in a healthmaintenance organization: efficacy of care management using clustervisits. Diabetes Care 1999;22:2011–7.

[13] Majumdar SR, Guirguis LM, Toth EL, et al. Controlled trial of amultifaceted intervention for improving quality of care for ruralpatients with type 2 diabetes. Diabetes Care 2003;26:3061–6.

[14] Coast-Senior E, Kroner B, Kelley C, Trili L. Management of patients withtype 2 diabetes by pharmacists in primary care clinics. Ann Pharma-cother 1998;32:636–41.

[15] Hawkins D, Bradberry JC, Cziraky MJ, et al. National PharmacyCardiovascular Council treatment guidelines for the management oftype 2 diabetes mellitus: toward better patient outcomes and new rolesfor pharmacists. Pharmacotherapy 2002;22:436–44.

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

0 50000 100000 150000 200000 250000 300000 350000 400000 450000

Prob

abili

ty c

ost-

effec

�ve

ness

Treshold Willingness to pay for a QALY (NGN)

Pharmaceutical Care

Usual care

Fig. 4 – Cost-effectiveness acceptability curves. NGN, Nigerian naira; QALY, quality-adjusted life-year.












http://www.diabetes.org/main/info/facts/factsna




http://da3.diabetesatlas.org/newsc269.html

http://da3.diabetesatlas.org/newsc269.html
























[16] Husereau D, Drummond M, Petrou S, et al. ISPOR Health EconomicEvaluation Publication Guidelines-CHEERS Good Reporting PracticesTask Force. Consolidated Health Economic Evaluation ReportingStandards (CHEERS)—explanation and elaboration: a report of theISPOR Health Economic Evaluation Publication Guidelines GoodReporting Practices Task Force. Value Health 2013;16:231–50.

[17] Oyejide OO. Health Reseach Methods. Ibadan: Leniks Publishers, 1992.[18] Mads-Haahr. Random Sequence Generator (1998–2011). Available from:

http://www.random.org/sequences/. [Assessed February 15, 2012].[19] Hepler C, Strand L. Opportunities and responsibilities in pharma-

ceutical care. Am J Hosp Pharm 1990;47:533–43.[20] National Health Insurance Scheme. NHIS Healthcare Providers Service

Price List. Abuja, Nigeria: Ministry of Health, 2005.[21] Management Science for Health (MSH) and World Health Organization.

International Drug Price Indicator Guide. In: Frye JE, ed. Cambridge:Management Science for Health, 2010.

[22] WHO-CHOICE. Cost-effectiveness thresholds (2005 International $); byregion. Available from: http://www.who.int/choice/costs/CER_levels/en/.[Accessed February 15, 2012].

[23] Drummond M. Introducing economic and quality of life measurementsinto clinical studies. Ann Med 2001;33:344–9.

[24] Furlong W, Feeny D, Torrance G. Algorithm for Determining HUI Mark 2(HUI2)/Mark 3 (HUI 3) Health Status Classification Levels, Health States,Single-Attribute Level Utility Scores and Overall Health Related Qualityof Life Utility Scores from HUI23.40Q Questionnaires. Dundas, Ontario,Canada: Health Utilities Inc., 2000.

[25] Briggs A, Claxton K, Sculpher M. Decision Modelling for HealthEconomic Evaluation. Oxford: Oxford University Press, 2006.

[26] Briggs B, Sculpher M, Buxton M. Uncertainty in the economicevaluation of health care technologies: the role of sensitivity analysis.Health Econ 1994;3:95–104.

[27] Stinnet A, Mullahy J. Net health benefits: a new framework for theanalysis of uncertainty in cost-effectiveness analysis. Med DecisMaking 1998;18:68–80.

[28] World Health Organization. Making Choices in Health: WHO Guide toCost Effectiveness Analysis. Geneva: World Health Organization, 2003.

[29] Cranor C, Bunting B, Christensen D. The Asheville Project: long-termclinical and economic outcomes of a community pharmacy diabetescare program. J Am Pharm Assoc 2003;43:173–84.

[30] Neto P, Marusic S, Júnior DP, et al. Effect of a 36-month pharmaceuticalcare program on coronary heart disease risk in elderly diabetic andhypertensive patients. J Pharm Pharm Sci 2011;14:249–63.

[31] McCabe C, Claxton K, Culyer A. The NICE cost-effectiveness threshold:what it is and what that means. Pharmacoeconomics 2008;26:733–44.

[32] An act to amend the National Minimum Wages Act Cap. N61 Laws ofthe Federation of Nigeria, 2004 to provide for a revised nationalminimum waged and; for related matters. Available from: http://www.aksjlegalresource.com/resource/Laws_of_the_Federation%5CNATIONAL%20MINIMUM%20WAGE%20_AMENDMENT_%20ACT%202011.pdf. [Accessed May 27, 2013].

[33] Klonoff D, Schwartz D. An economic analysis of interventions fordiabetes. Diabetes Care 2000;23:390–404.

[34] Laupacis A, Deeny D, Detsky A, Tugwell P. How attractive does a newtechnology have to be to warrant adoption and utilization? Tentativeguidelines for using clinical and economic evaluations. Can Med AssocJ 1992;146:473–81.

[35] Grosse S. Assessing cost-effectiveness in health care: history of the$50,000 per QALY threshold. Value Health 2008;8:165–78.

[36] The World Bank. Data by country 2010: Nigeria. Available from: http://data.worldbank.org/country/nigeria. [Accessed February 15, 2012].

[37] Campbell D, Stanley J. Experimental and Quasi-Experimental Designsfor Research. Chicago, IL: Rand McNally College Publishing Company,1963.

[38] Cook T, Campbell D. Quasi-experimentation: Design and AnalysisIssues for Field Studies. Chicago, IL: Rand McNally College PublishingCompany, 1979.

[39] West J, Goldberg K. Diabetes self-care knowledge among outpatients ata Veterans Affairs medical center. Am J Health-Syst Ph 2002;59:849–52.










http://www.random.org/sequences/









http://www.who.int/choice/costs/CER&underscore;levels/en/




























http://www.aksjlegalresource.com/resource/Laws&underscore;of&underscore;the&underscore;Federation%5CNATIONAL%20MINIMUM%20WAGE%20&underscore;AMENDMENT&underscore;%20ACT%202011.pdf












http://data.worldbank.org/country/nigeria

http://data.worldbank.org/country/nigeria











Economic Burden of Cardiovascular Diseases in the Russian FederationAnna Kontsevaya, PhD*, Anna Kalinina, PhD, Rafael Oganov, PhD

Department of Primary Prevention in Healthcare, National Research Center for Preventive Medicine, Moscow, Russia

A B S T R A C T

Objectives: In the Russian Federation, cardiovascular disease (CVD) isthe primary cause of death and premature death; however, to date,there have been no systematic cost-of-illness studies to assess theeconomic impact of CVD. Methods: The economic burden of CVD wasestimated from statistic data on morbidity, mortality, and health careresource use. Health care costs were estimated on the basis ofexpenditure on primary, outpatient, emergency, and inpatient care,as well as medications. Non–health care costs included economiclosses due to morbidity and premature death in the working age.Results: CVD was estimated to cost Russia RUR 836.1 billion (€24,517.8million) in 2006 and RUR 1076 billion (€24,400.4 million) in 2009. Of thetotal costs of CVD, 14.5% in 2006 and 21.3% in 2009 were due to healthcare, with 85.5% and 78.7%, respectively, due to non–health care costs.Conclusions: CVD is a leading public health problem. We first

assessed the economic burden of CVD in Russia. Our results can beused for planning investments in prevention programs and measuresfor improving care for patients with CVD. Regular monitoring of theeconomic burden of CVD in the future at the federal, regional, andmunicipal levels will allow assessment of the dynamics of economicburden, as well as the effectiveness of investments in the economy inprimary and secondary prevention. Because data are relativelyunavailable, there are important limitations to this study, whichhighlight the need for more accurate CVD-specific information.Keywords: cardiovascular disease, coronary heart disease, cerebro-vascular diseases, cost-of-illness study, economic burden, Russia.


Introduction

In the Russian Federation, cardiovascular disease (CVD) is theprimary cause of death and premature death [1]. There has been,however, no systematic cost-of-illness study to assess its eco-nomic impact. The World Bank calculated health care expendi-tures in two regions of Russia and extrapolated these data to theentire country [2]. These regions, however, were not representa-tive of the entire country because the calculations included onlyhealth care costs. Another attempt to calculate the economicburden of CVDs in Russia was made by the World HealthOrganization. It calculated the economic burden for 2005 andpredicted a prognosis of burden for 2015 and 2030 of the mostprevalent noncommunicable diseases from a macroeconomicperspective based on death rate [3].

The objectives of this study were to estimate the economiccosts of CVD in Russia, including health care costs and produc-tivity loss, and to estimate the proportion of total CVD costattributable to coronary heart disease (CHD) and cerebrovasculardiseases as was estimated in the study of Leal et al. [4] in theEuropean Union (EU).

Methods

Methodological Background

Cost-of-illness analyses involved the identification, measure-ment, and valuing of resources related to CVD in Russia in theperiod 2006 to 2009. The calculation included health care costsand costs outside the health care sector (productivity lossesassociated with premature death or morbidity and disabilitypensions). All expenditures were measured for the period 2006to 2009 in the prices of the appropriate year. Additional file 1includes the sources of information used for calculations.

The national currency rubles was converted to euros (€) byusing a weighted exchange rate for the period 2006 to 2009.Epidemiological and health care utilization data were acquiredfrom the Ministry of Health of the Russian Federation and themarket research company COMCON from the published litera-ture. Analysis was based on the International Statistical Classifica-tion of Diseases, 10th Revision (ICD-10) categories: CVD (ICD-10category I00–I99), hypertensive diseases (ICD-10 category I10–I15), CHD (ICD-10 category I20–I25), and cerebrovascular disease






� Address correspondence to: Anna Kontsevaya, Department of Primary Prevention in Healthcare, National Research Center forPreventive Medicine, Petroverigski Lane, 10, 101990, Moscow, Russia.


dx.doi.org/10.1016/j.vhri.2013.06.010

dx.doi.org/10.1016/j.vhri.2013.06.010

dx.doi.org/10.1016/j.vhri.2013.06.010


dx.doi.org/10.1016/j.vhri.2013.06.010

(ICD-10 category I60–I69). All sources of information used for thecalculations are listed in the Additional information file.

Health Care Expenditure

The following categories of CVD health care services were includedin the calculations: primary and outpatient care, accident andemergency (A&E) care, hospital inpatient care, cardiosurgery andpercutaneous coronary interventions (PCIs), and medications. Car-diosurgery and PCI expenditures were calculated separately fromhospital inpatient care because they are financed through differentsources in Russia. Hospital inpatient care is paid by the healthinsurance system, while cardiosurgery and PCIs are paid by directpayment from the federal budget in the framework of the federalprogram on high technology and costly medical care or by thepatient. Other types of activities related to the CVD were notincluded because of the difficulties in locating information. Datafor A&E care and hospital inpatient care were received from theMinistry of Health of the Russian Federation; data on primary andoutpatient care were obtained from several information sources;and data on medications were received from annual pharmacoepi-demiology surveys through the COMCON company and otherliterature data.

Health Service Utilization

Primary and outpatient carePrimary care activities consisted of CVD-related visits to generalpractitioners (GPs), as well as GP visits to patients’ homes. There is aspecific statistical form for all medical organizations that includesdata on the number of patients who visited this organization foroutpatient care during the year according to the ICD-10 categories.This statistical form is centrally received in the Ministry of Healthand processed as a single form for the entire country. We received asingle form for the entire country for 2006 to 2009. The number ofvisits for each outpatient was calculated on the basis of thesestatistical forms and data from previous studies examining themean number of visits of patients with CVD during the year.

Hospital inpatient careInpatient care was estimated on the basis of the number of CVD-related days in the hospital. There is also a specific statisticalform for all inpatient medical organizations that includes data onthe number of hospitalizations and number of hospital days inthis organization during the year according to the ICD-10 catego-ries. This statistical form is also centrally received in the Ministryof Health and processed as a single form for the entire country.We received the statistical single form for inpatient care in Russiafor 2006 to 2009 and selected data on ICD-10 categories of interest.

A&E careA&E care consisted of all CVD-related hospital emergency visits.A 2009 special statistical form for all inpatient medical organ-izations included information regarding the number of hospitalemergency visits according to the ICD-10 categories. We selecteddata on ICD-10 categories in 2009 and extrapolated these valuesfor 2006 to 2008.

Cardiosurgery and PCIThe main cardiosurgery institution in Russia, the Bakoulev Centerfor Cardiovascular Surgery, centrally collects information from allmedical organizations involved in such interventions in Russia andannually publishes a statistical yearbook. We selected data regard-ing the number of PCI, coronary artery bypass grafting, and someother cardiosurgeries performed in Russia in 2006 to 2009.

Health care unit costsUnit costs of an inpatient day, outpatient visit, and emergency visitwere obtained from the Ministry of Health. The official Web siteannually publishes information on the mean costs of inpatient days,outpatient visits, and emergency visits in different specialties andtotal expenditures in the framework of the program of the govern-mental guarantees of the medical care.

The costs of cardiosurgery and PCI paid directly by the federalbudget are published annually on the Web site of the Ministry ofHealth in the description of the Federal Program on HighTechnology and Costly Medical Care.

Expenditure on medicationThere are no national sources of information regarding nationalexpenditures on medications in Russia. We used data from severalpharmacoepidemiology surveys made in Russia in 2006 to 2009 andextrapolated these data for the entire country. The main source wasthe databases of annual surveys conducted by the COMCONcompany; other studies were used to identify patients with CVDregularly taking medication for long periods as well as some otherdata. Costs of medications were calculated on the basis of meanprices during the studding years, including value added tax (VAT).

Non–Health Service Costs

Non–health service costs included productivity losses associatedwith premature death and morbidity and disability pensions.Because little information was found on informal care costs andout-of-pocket expenses across the country, these costs were notincluded in the calculations.

Estimation of productivity costs due to premature death duringworking ageProductivity costs due to premature death during working ageincluded the gross domestic product (GDP) per employed personrelated to CVD attributable to mortality.

The productivity loss from CVD-mortality was estimated bycalculating the following:

1. number of CVD-related deaths during working age (retirementage is 60 years for men and 55 years for women);

2. number of remaining work years at the time of death (toestimate the likely GDP that an individual who died wouldhave otherwise produced);

3. annual GDP per employed person; and4. economic activity and unemployment rates.

Future GDP was not indexed in the main analysis, as the usualdiscount rate of 3% to 3.5% is not reasonable for Russia. Theinflation rate was 9% in 2006, 11.9% in 2007, 13.3% in 2008, and8.8% in 2009. The effects of indexation on productivity costs usingrates of 10% and 15% were studied through sensitivity analysis.

Estimation of productivity costs due to cardiovascular morbidityMorbidity costs were defined as those associated with CVD-attributable absence from work, estimated by multiplying thenumber of certified days off work due to CVD by GDP produced inone working day.

The number of CVD-related working days lost was obtainedfrom a special statistical form for all medical organizations thatincluded data on the number of disability days during the yearaccording to the ICD-10 categories. This statistical form is alsocentrally received at the Ministry of Health and processed as asingle form for the entire country. We received a single form forworking days lost in Russia for 2006 to 2009 and selected data onICD-10 categories related to CVD.


Sensitivity analysisWe examined the effects of 20% changes in health care costs.Because medication costs were estimated by using subjectiveassumptions, the effect of 50% changes in these categories wastested. We also assessed the effects of indexation on productivitycosts by using rates of 10% and 15%.

Results

Table 1 shows the average unit costs used for calculations andthose aggregated from different sources.

Health Care Costs

CVD accounted for more than 69.6 million hospital bed days inRussia in 2006 and 68.0 million in 2009. The number of hospital-izations did not change, but the mean duration of hospital staydecreased each year (from 15.0 to 14.1 days). The number ofhospital bed days for CHD was 27.1 million in 2006 and 26.3 millionin 2009, and for cerebrovascular diseases it was 17.6 and 17.2million, respectively. CVD represented 479.2 hospital bed days per1000 persons in 2009 (Table 2), 185.2 days for CHD, and 121.2 daysfor cerebrovascular diseases. This parameter is higher than thetotal for the EU [4] but comparable to some European countries.

The number of GP and outpatient visits was 1603.2 per 1000persons in 2009, with CHD representing 300.9 visits for 1000persons and cerebrovascular diseases representing 286.6 (Table 2).

CVD cost Russian health care systems approximately RUR 121billion or €3558.5 million in 2006 and (Table 3) RUR 229.5 billion or€5204.4 million in 2009. The major component of CVD-relatedhealth care expenditure was inpatient care, which accounted forRUR 52.7 billion (€1545.1 million) in 2006 and RUR 109 million(€2470.8 million) in 2009, representing 43.4% and 47.5% of totalhealth care costs, respectively. CVD-related medication expendi-ture was also a large cost, representing 27% in 2006 (RUR 32.6billion) of total costs and 21% in 2007 (RUR 47.6 billion) of healthcare costs. The third largest component of CVD-related health careexpenditure was outpatient care, which represented 20.2% of thetotal health care costs in 2006 and 21.7% of the total health carecosts in 2009. The other two cost components (emergency care andcardiosurgery) accounted for 9.5% in 2006 and 10.0% in 2009 ofcosts, with A&E representing the smallest component. The struc-ture of the CVD costs in Russia was similar to that of the EU [4],where inpatient care and pharmaceutical expenditure were majorcomponents of CVD-related health care expenditure.

Non–Health Care Costs

Mortality losses were relatively high in Russia because of high CVDdeath rates in the general population and in the working agepopulation. The number of CVD deaths per 1000 persons was severaltimes higher than in Europe [4]. There is a prominent difference in

the number of deaths in working-age men and women. The deathrate of working-age men (younger than 60 years) was 3.2 in Russia in2009, while for working-age women it was 0.7. This explains thesignificant difference in working year losses between men andwomen (22.1 and 3.4 for 1000 persons in 2009). In the EU, workingyear losses in men were also higher than in women [4], but to alesser degree than in Russia. CHD accounted for approximately halfof CVD mortality losses, while values for cerebrovascular diseaseswere much lower, particularly regarding working-year losses.

CVD accounted for 2.1 million working years lost in 2006 and1.7 million in 2009 owing to deaths during working ages. GDPlosses due to mortality in working ages were estimated to costapproximately RUR 630 billion (€18,483.3 million) in 2006 and RUR117.1 billion (€16,509.6) in 2009, respectively, after adjusting forworking status (Table 3).

There were 70.1 million working days lost in 2006 and 67.3million in 2009 because of CVD morbidity. This represented a costof RUR 83.8 billion (€2458.4 billion) in 2006 and RUR 117.4 billion(€2662.5 billion) in 2009 after adjusting for working status(Table 3). The costs of disability pensions were rather small,€17.7 million in 2006 and €23.9 million in 2009.

Total Costs of CVD

Overall, CVD was estimated to cost the Russian economy RUR836.1 billion (€24,517.8 million) in 2006 and RUR 1076 billion(€24,400.4 million) in 2009 (Table 3). Of the total costs of CVD,14.5% in 2006 and 21.3% in 2009 were due to health care, whereasnon–health care costs were 85.5% and 78.7%, respectively. Thestructure of costs is completely different from the EU data, wherethe health care cost was the most significant [4].

Share of GDP (CVD)

The total costs of CVD were equal to 3.1% of the GDP of theRussian Federation in 2006 and 2.8% in 2009, respectively(Table 4).

Costs of CHD and Cerebrovascular Diseases

The total costs of CHD were RUR 339.7 billion (€9962.2 million) in2006 (Table 4) and RUR 406.6 billion (€9220.4 million) in 2009. Healthcare costs were RUR 58.6 billion (€1719.0 million) and RUR 103.9billion (€2355.2 million) in 2009. The share of health care costsrelative to total costs increased from 17.3% in 2006 to 25.5% in 2009.

The total costs of cerebrovascular diseases were muchlower than those of CHD. Total care costs were RUR 139.2 billion(€4083.3 million) in 2006 and RUR 184.5 billion (€4183.9 million) in2009. Health care cerebrovascular costs were RUR 18.8 billion(€551.9 million) in 2006 (Table 4) and RUR 38.6 billion (€874.1million) in 2009. The share of health care costs compared withtotal costs increased from 13.5% in 2006 to 20.9% in 2009.

Table 1 – Average unit costs in Russia in 2006–2009.

2006 2007 2008 2009

Ruble Euro Ruble Euro Ruble Euro Ruble Euro

GDP per employed personAnnual 389,021 11,408.2 467,463 13,668.5 584,338 16,053.2 562,500 12,755.1Daily 1,195 35.0 1,445 42.3 1,837 50.5 1,745 39.6

Health care unit costGP and outpatient visit 115 3.4 149 4.4 181 5.0 219 5.0A&E visit 686 20.1 888 26.0 1,110 30.5 1,387 31.5Inpatient day 754 22.1 947 27.7 1,232 33.8 1,602 36.3

A&E, accident and emergency; GDP, gross domestic product; GP, general practitioner.


Sensitivity Analysis

Varying total health care costs upwards and downwards by 20%produced a variation of 1% in the baseline of total CVD-relatedcosts in the period 2006 to 2009. Our results did not varysignificantly when the assumptions used to derive medicationcost estimates were varied by 50%, resulting in changes of 1% intotal costs.

Future earning losses at a 10% discount rate were associatedwith a reduction of 8% in costs, while a 15% discount rate wasassociated with a reduction of 11% in costs.

Discussion

This is the first study to estimate the full burden of CVD,including health care and non–health care costs in Russia. Weestimated that the total burden of CVD varied between €24.4and €32.3 billion in the period 2006 to 2009. In a recent Europeanstudy, the estimated cost of CVD was €169 billion [4]. Becausethe population of EU is several times higher than that of Russia,the amounts were comparable. We found, however, that the

structures of the total costs were completely different in Russiaand the EU. A European study found that health care expendi-ture accounted for 61% of costs, but in our study the share ofhealth care expenditures was only 14.5% of total costs in 2006and 21.3% in 2009. If our study included the cost of informalcare, health care costs would be even lower. Non–health carecosts (mainly due to premature death during working age)account for 80% of total CVD costs in Russia.

Direct costs associated with CVD in the EU included costs forhospital admissions (57%), costs for drug therapy (27%), and costsfor ambulatory care and A&E (16%) [4]. In general, the structure ofdirect costs in our study was similar.

In our study, the share of health care costs for CVD relative tototal costs was 13.5% to 20.9%. In the study of 27 EU countries, itwas shown that total costs from stroke were 68.5% due to directcosts and 31.5% due to indirect costs [5]. In the United States, theeconomic burden of stroke includes 67% of direct health carecosts and 33% of losses in the economy due to prematuremortality and disability days [6]. Thus, in Russia, the structureof the economic burden of CVD is different from that due to thepredominance of indirect costs.

Table 2 – Resource units per 1000 population in the Russian Federation in 2009.

CVD CHD Cerebrovascular diseases

Mortality lossesDeaths

All 8.0 4.1 2.6Men 7.8 4.2 2.2Women 8.2 4.0 3.0

Working years lostAll 11.1 4.5 1.9Men 22.1 9.1 3.5Women 3.4 0.9 0.8

Morbidity lossesWorking days lost 474.2 80.1 73.3

Health care unitGP and outpatient visits 1603.2 300.9 286.6Hospital emergency visits 47.9 5.1 10.6Inpatient days 479.2 185.2 121.2

CHD, coronary heart disease; CVD, cardiovascular disease; GP, general practitioner.

Table 3 – Cost of CVD in Russia in 2006–2009.

2006 2007 2008 2009

RURmillion

€

millionRUR

million€

millionRUR

million€

millionRUR

million€

million

Inpatient care 52,687.9 1,545.1 65,791.5 1,923.7 85,283.7 2,343.0 108,960.9 2,470.8Primary and outpatient care 24,469.5 717.6 32,385.6 946.9 40,262.8 1,106.1 49,833.0 1,130.0A&E care 4,032.4 118.3 5,614.5 164.2 7,410.2 203.6 9,425.8 213.7Cardiosurgery and PCI 7,517.5 220.5 9,451.6 276.4 11,701.9 321.5 13,714.5 311.0Medications 32,637.0 957.1 35,249.0 1,030.7 40,124.4 1,102.3 47,578.7 1,078.9Total health care costs 121,344.3 3,558.5 148,492.3 4,341.9 184,783.0 5,076.5 229,513.0 5,204.4GDP losses due to mortality in

working age630,280.4 18,483.3 687,087.1 20,090.3 860,872.9 23,650.4 728,075.2 16,509.6

GDP losses due to morbidity 83,830.7 2,458.4 104,067.2 3,042.9 127,864.4 3,512.8 117,416.1 2,662.5Disability pensions 602.2 17.7 666.7 19.5 833.8 22.9 1,054.1 23.9Total non–health care costs 714,713.3 20,959.3 791,821.0 23,152.7 989,571.0 27,186.0 846,545.4 19,196.0Total costs 836,057.6 24,517.8 940,313.3 27,494.5 1,174,354.0 32,262.5 1,076,058.3 24,400.4

A&E, accident and emergency; CVD, cardiovascular disease; GDP, gross domestic product; PCI, percutaneous coronary intervention; RUR,rubles.


Several studies demonstrated that health care costs forvarious CVDs differ significantly depending on the country, thelength of hospitalization, and other factors [7–9].

CHD was the most costly CVD (37.8% of total CVD costs).According to estimates of economic burden in the United States,CHD was also the most expensive disorder [10]. In a European study,CHD accounted for 27% of total costs [4]. A UK study showed thatCHD and stroke costs were similar (29% and 27% of total CVD costs,respectively) [11]. In some studies, it was shown that myocardialinfarction typically accounts for the largest share of CHD costsowing to long hospitalization and costly interventions [12,13].

Studies examining the economic burden of diseases enablecomparisons between the burdens of different diseases, allowingdecision makers to prioritize limited research funds to areas withthe highest burden [14]. Furthermore, if such studies are per-formed at regular intervals, the impact of health policy decisionscan be measured. In Russia, these studies can be used to monitorgovernmental programs involved in chronic disease prevention.

The results of calculations performed in this study can beused to plan investments in prevention programs and improvecare to patients with CVD. Regular monitoring of the economicburden of CVD in the future at the federal, regional, andmunicipal levels will allow the assessment of the dynamics ofeconomic burden, as well as the effectiveness of investments inthe economy in primary and secondary prevention.

The next step of such studies is to estimate the economic costof risk factors of chronic disease. This will provide support forprioritizing resources for prevention and public health [15].

To be in a better position to inform policy decisions aimed atreducing the burden of disease, improved information regardingepidemiology and accurate information regarding resource useand unit costs is imperative. In Russia, there is lack of recentreliable epidemiology data as well as gaps in the official statisticson resource use.

Our results are likely underestimated. Some categories ofcosts, such as costs of informal care, were not included becauseof data limitations.

Despite these acknowledged and important data limitations,this is the first study to quantify the burden of CVD in Russia.

Acknowledgments

We are grateful to Michail Khudjakov for useful contributions tothis project. The comments from three anonymous reviewers arealso acknowledged.Source of financial support: The authors haveno other financial relationships to disclose.

Source of financial support: The authors have no otherrelationships to disclose.

R E F E R E N C E S

[1] Population data for 2006–2009 from Russian Federal Service of StateStatistics (Rosstat). Available from: http://www.gks.ru/wps/wcm/connect/rosstat/rosstatsite/main/ [Accessed October 16, 2010].

[2] World Bank. Dying too young: addressing premature mortality and illhealth due to non communicable diseases and injuries in the RussianFederation. 2005: 145. Available from: http://siteresources.worldbank.org/INTECA/Resources/Dying_too_Young_Summary_UPDATED_Oct_19.pdf[Accessed October 16, 2010].

[3] Abegunde DO, Mathers CD, Adam T, et al. The burden and costs ofchronic diseases in low-income and middle-income countries. Lancet2007;370:1929–38.

[4] Leal J, Luengo-Fernández R, Gray A, et al. Economic burden ofcardiovascular diseases in the enlarged European Union. Eur Heart J2006;27:1610–9.

[5] European Heart Network. European Cardiovascular Disease Statistics2008. Brussels: European Heart Network, 2008. Available from: http://

Table

4–Cost

ofCVD,CHD,and

cere

bro

vasc

ula

rdisease

sin

Russ

iain

2006–2009.

2006

2007

2008

2009

CVD

CHD

Cere

CVD

CHD

Cere

CVD

CHD

Cere

CVD

CHD

Cere

Totalco

sts

RUR million

836,05

7.6

339,71

2.3

139,23

9.6

940,31

3.3

377,18

2.3

158,04

4.4

1,17

4,35

4.0

471,83

5.9

196,85

1.7

1,07

6,05

8.3

406,61

8.8

184,51

1.9

€million

24,517

.89,96

2.2

4,08

3.3

27,494

.511

,028

.74,62

1.2

32,262

.512

,962

.55,40

8.0

24,400

.49,22

0.4

4,18

3.9

Shareof

GDP

(%)

3.1

1.26

0.52

2.8

1.14

0.48

2.8

1.14

0.48

2.8

1.04

0.47

Hea

lth

care

costs

RUR million

121,34

4.3

58,619

.118

,821

.214

8,49

2.3

69,141

.823

,837

.618

4,78

3.0

84,148

.430

,731

.222

9,51

3.0

103,86

2.9

38,550

.0

€million

3,55

8.5

1,71

9.0

551.9

4,34

1.9

2,02

1.7

697.0

5,07

6.5

2,31

1.8

844.3

5,20

4.4

2,35

5.2

874.1

Shareof

total

costs

(%)

14.5

17.3

13.5

15.8

18.3

15.1

15.7

17.8

15.6

21.3

25.5

20.9

Cere,

cerebrova

sculardisea

se;CHD,co

ronaryhea

rtdisea

se;CVD,ca

rdiova

sculardisea

se;GDP,

gross

domes

ticpro

duct;RUR,ru

bles

.



http://www.gks.ru/wps/wcm/connect/rosstat/rosstatsite/main/

http://www.gks.ru/wps/wcm/connect/rosstat/rosstatsite/main/



http://siteresources.worldbank.org/INTECA/Resources/Dying_too_Young_Summary_UPDATED_Oct_19.pdf










www.herzstiftung.ch/uploads/media European_cardiovascular_disease_statistics_2008.pdf [Accessed October 16, 2010].

[6] Go AS, Mozaffarian D, Roger VL, et al. Heart disease and stroke statistics—2008 update: a report from the American Heart Association Statistics Com-mittee and Stroke Statistics Subcommittee. Circulation 2008;117:e25–146.

[7] Kauf TL, Velazquez EJ, Crosslin DR, et al. The cost of acute myocardialinfarction in the new millennium: evidence from a multinationalregistry. Am Heart J 2006;151:206–12.

[8] Tiemann O. Variations in hospitalisation costs for acute myocardialinfarction—a comparison across Europe. Health Econ 2008;17(1, Suppl.):S33–45.

[9] Epstein D, Mason A, Manca A. The hospital costs of care for stroke innine European countries. Health Econ 2008;17(Suppl.):S21–31.

[10] Druss BG, Marcus SC, Olfson M, Pincus HA. The most expensivemedical conditions in America. Health Aff 2002;21:105–11.

[11] Luengo-Fernández R, Leal J, Gray A, et al. Cost of cardiovasculardiseases in the United Kingdom. Heart 2006;92:1384–9.

[12] Reinhold T, Lindig C, Willich SN, Brüggenjürgen B. The costs ofmyocardial infarction—a longitudinal analysis using data from a largeGerman health insurance company. J Public Health 2011;19:579–86.

[13] Ioannides-Demos LL, Makarounas-Kirchmann K, Ashton E, et al. Costof myocardial infarction to the Australian community: a prospectivemulticenter survey. Clin Drug Investig 2010;30:533–43.

[14] Gross CP, Anderson GF, Powe NR. The relation between funding by theNational Institutes of Health and the burden of disease. N Engl J Med1999;340:1881–7.

[15] Scarborough P, Bhatnagar P, Wickramasinghe KK, et al. The economicburden of ill health due to diet, physical inactivity, smoking, alcoholand obesity in the UK: an update to 2006–07 NHS costs. J Public Health(Oxf) 2011;33:527–35.


































Cost for Treatment of Chronic Lymphocytic Leukemia in SpecializedInstitutions of UkraineOlena Mandrik, PhD1,2,*, Isaac Corro Ramos, PhD3, Olga Zalis’ka, PhD1, Andriy Gaisenko, PhD4, Johan L. Severens, PhD2,3

1Danylo Halytsky Lviv National Medical University, Lviv, Ukraine; 2Institute of Health Policy & Management, Erasmus University Rotterdam, Rotterdam, TheNetherlands; 3Institute for Medical Technology Assessment, Erasmus University Rotterdam, Rotterdam, The Netherlands; 4National Cancer Institute, Kyiv,Ukraine

A B S T R A C T

Objective: The aim of this study was to identify, from a health careperspective, the cost of treatment for chronic lymphocytic leukemia inspecialized hospitals in Ukraine. Methods: Cost analysis was per-formed by using retrospective data between 2006 and 2010 frompatient-file databases of two specialized hospitals (145 patients).Uncertainty was assessed by using bootstrapping and multivariatesensitivity analyses. Linear regression analysis was used to analyzewhether patients’ characteristics are related to health care costs. Inaddition, one-way analysis of variance (Welch test) and paired-sample t test were conducted to compare mean costs of treatmentbetween the two hospitals and mean expenses for drugs and in-hospital stay. Results: The average annual cost for a patient’s drugtreatment is 2047 EUR. The cost of hospitalization was significantlylower (t ¼ 5.026; significance two-tailed ¼ 0.000) and equal to 541 EURper person, resulting in total expenditures of 2589 EUR. Mean totalcosts in the bootstrap analysis were equal to 2584 EUR (median 2576

EUR, 97.5th percentile 3223 EUR; 2.5th percentile 1987 EUR). Theregression analysis did not reveal a relation between patients’ char-acteristics and health care costs, although hospital choice was aninfluential parameter (β ¼ −0.260; significance ¼ 0.002). Significantdifference in mean costs of two analyzed hospitals was also con-firmed by one-way analysis of variance (Welch statistics 19.222, P ¼0.000). Conclusions: Drug treatment comprises the largest portion oftotal costs, but differences between hospitals exist. Because manypatients in Ukraine pay out of pocket for in-hospital drugs, these costsare a high economic burden for patients with chronic lymphocyticleukemia.Keywords: chronic lymphocytic leukemia, cost of treatment, hematologicmalignancies.


Introduction

Globally, there are approximately 7.4 million cancer deaths peryear, which is approximately 13% of deaths from all causes.Because the population of many countries around the world isaging, it can be expected that cancer incidence will increase [1].Among oncologic diseases, chronic hematologic malignancies arecomparatively rare. In Ukraine in 2010 the officially registeredtotal morbidity rate for patients with diagnosed leukemia was 7.8per 100,000 people, of which 39.3% did not live a year afterdiagnosis [2]. Chronic lymphocytic leukemia (CLL) is the mostfrequent form of leukemia in Western countries, and it accountsfor approximately 30% to 40% of all leukemias [3,4]. It is charac-terized by the clonal proliferation and accumulation of neoplasticB lymphocytes in the blood, bone marrow, lymph nodes, andspleen. Although the median age of patients at diagnosis ishigher than retirement age and so has no significant impact onstate productivity loss [5], the economic impact of CLL is signifi-cant due to long duration and high expenses related to treatment,combined with low cure rates. Nevertheless, early diagnosis andeffective treatment of hematologic malignancies shift the

indicators of a patient’s life expectancy to positive values [6].For example, in the United States for the time period 1999 to 2005,the 5-year survival rate for leukemia was 82% (79% for CLL),although in the time period 1975 to 1977, this indicator was closeto Ukrainian data—35% [7,8].

In-hospital medical care for patients with CLL is generallyprovided in 35 hematologic departments, based in district hospi-tals (16), state city hospitals (12), oncologic dispensaries (4), andspecialized institutes of the National Academy of Medical Scienceof Ukraine (3) [2,9]. To the latter group belong two hematologicinstitutes and the National Cancer Institute, which is a leadingstate institution additionally responsible for methodological andscientific development in this clinical area. Treatment schemesfor patients with CLL are based on a clinical protocol thatproposes a number of treatment options for patients with CLLand was first developed and approved under an order of theMinistry of Health of Ukraine in 2010 [10]. State pharmaceuticalprovision for adult oncologic patients is granted through thenational treatment program “Oncology” for the years 2011 to2016, although governmental financing is insufficient and drugtreatment is usually paid out of pocket by patients [11].






� Address correspondence to: Olena Mandrik, Danylo Halytsky Lviv National Medical University, 69 Pekarska Street, Lviv, Ukraine 79010.


dx.doi.org/10.1016/j.vhri.2013.06.006

dx.doi.org/10.1016/j.vhri.2013.06.006

dx.doi.org/10.1016/j.vhri.2013.06.006


dx.doi.org/10.1016/j.vhri.2013.06.006

Although CLL has a significant impact on patients’ quality oflife [12,13], studies exploring economic costs and burden ofhematologic malignancies are relatively rare in English-language publications worldwide [14,15]. Possible reasons for thislack of information appear to include the low incidence rate andaged study population (over 60 years old), which make broad,well-designed economic analyses a challenge for most research-ers [5,16]. These few cost reviews identified cost drivers for CLL aschemotherapy costs, intravenous immunoglobulin costs, trans-plantation costs, and costs associated with the differential stain-ing cytotoxicity assay, with the main cost drivers related to thetreatment chosen [14–16].

The health care system itself, including organization ofmedical care for oncologic patients, is going through a stage oftransformation. Changes include the implementation of a uni-versal reimbursement system to begin in 2016, standardization ofmedical help with enhanced control on follow-up of clinicalprotocols, and more strict division between primary, secondary,and tertiary levels of medical help. Considering that the majorrecipients of the central state budget are specialized institutions(tertiary level of help), the primary aim of this research was toidentify the cost of treatment for CLL in specialized hospitals inUkraine from a health care perspective and to understandwhether patient characteristics are related to these costs.

Methods

The study was conducted from a health care perspective,accounting for direct medical costs to illustrate which costs willbe paid by the Ministry of Health after the health care systemtransformation.

Analysis included data from databases of two specializedhospitals—National Cancer Institute and State Institute of Hem-atology—that receive state financing through the national treat-ment program “Oncology.” These hospital databases were madein the programs Access and Word for the purpose of data storageand included all the information available in hard copies ofhospital cards; the data were typed into the hospital databasesretrospectively by qualified personnel (hospital assistants). After-ward, data from the two hospital databases were transferred intoExcel, and SPSS databases were created for the purpose of dataanalysis.

The study population included all newly diagnosed andrelapsed patients with CLL (145 in total) who were hospitalizedduring the period from 2006 to 2010 and whose data wererecorded in the electronic database. The information derivedfrom the hospital cards (excluding patients’ identification infor-mation) contained the following data: sex of the patient, ageduring diagnosis and treatment, number of years a patient liveswith the disease, year of treatment, therapies prescribed andduration of treatment, the number of hospitalizations per year,and the duration of hospitalization. Stage of the patient’s diseaseand health state on Eastern Cooperative Oncology Group (ECOG)performance status were excluded from the factor list becausedata on these parameters were frequently missing.

Costs related only to CLL diagnosis for the last observationalyear were calculated. These costs included drug expenses and in-hospital costs. The cost of diagnostics, medical procedures, hotelservices, and medical personnel is included in the integral in-hospital cost, based on data of the economic department of theNational Cancer Institute. These costs reflect the approximatecosts for oncologic patients in a specialized hospital and areequal to 16.3 EUR per patient-day [17]. Out-of-hospital health carecosts were not calculated because according to the clinicalprotocol [10], the treatment of patients with CLL should beconducted only in hospital. The average length of hospital stay

and drug costs were assessed by retrospective analysis of patientfile data.

To assess drug usage, daily defined doses and total amountreceived during the year were recorded. To calculate drug costs,we used a stepwise approach to determine an average price,depending on the availability of information: tariff in govern-mental purchases (2010); price, registered in the Ministry ofHealth; and distributors’ price.

Multivariate sensitivity analysis was conducted. Price devia-tions for the sensitivity test of all drugs were calculated by usingthe minimum and maximum prices from the available sources(hospital purchases, state registered prices, and distributors’prices). There are no defined general tariffs for hospital stay inUkraine, which are relatively low in comparison to medical costsin the European countries and may vary from 3.4 to 19.2 EUR [17–19]. All statistical analyses were performed in IBM SPSS Statistics20 (SPSS, Inc., Chicago, IL), and bootstrapping (1000 replications)was performed in Microsoft Excel 2010. To analyze whetherchoice of the hospital and age and sex of a patient have animpact on total health care costs, logarithmic data transforma-tion was performed on nonnormally distributed costs and alinear regression analysis was applied. Because of frequentlymissing data for the parameter “stage of the disease,” as a proxyfor disease progression we included “number of years a patient isliving with the disease” in the linear regression analysis. Basedon Cook’s distance (0.028571), which measures the effect ofdeleting a given observation and so allows to define data pointswith large residuals, we excluded six outliers to improve theresiduals plot and model validity. One-way analysis of variance(Welch test) was conducted to compare mean costs of treatmentin the two hospitals involved (asymptotically F distributed).Paired-sample t test was used to compare difference in meanexpenses for drugs and in-hospital stay.

Results

Overall, data of 113 patients from the first hospital (StateHematology Institute) and of 32 patients from the second hospital(National Cancer Institute) were analyzed. Patients were aged 40to 85 years (mean age 62.9 years, mean age during diagnosis 60.3years, SD 9.8 years). From the sample, 27.6% (40) of the patientswere newly diagnosed. Because of limited sample size, thedistribution of patients by sex was not equal in different agegroups, with the total proportion of men being equal to 60.7%(88 men).

Values of the cost items (drugs) and cost deviations for thesensitivity test are presented in Table 1. As can be seen in Table 1,the highest cost per milligram was for fludarabine, vincristine,and rituximab. Rituximab and fludarabine (if Fludara was pre-scribed) had the highest price per average daily dose, equal to312.03 EUR for fludarabine and 237.93 EUR for rituximab. Becausedrug expenditures depend not only on cost per item but also ontotal volume used, we present cost-items utilization and charac-teristics of population using it in Table 2. Data are presented foritems that were used by more than 3% of patients. Cyclophos-phamide, fludarabine, and vincristine were prescribed to most ofthe patients. Characteristics of the study population using spe-cific cost items showed a significant difference in the percentageof men prescribed cyclophosphamide, mitoxantrone, and chlor-ambucil in comparison to other drugs. No significant difference inthe patients’ age was observed, although on average the age ofpatients receiving alemtuzumab was lower and of those receivingchlorambucil was higher. From Table 2 it also may be observedthat the injectable form of fludarabine is prescribed significantlymore than the oral form.


The average annual cost for a patient’s drug treatment is 2047EUR. The average cost of in-hospital stay is 542 EUR per person,resulting in total expenditures of 2589 EUR. Results indicate thatexpenses for drugs significantly exceed hospitalization costs (t ¼5.026; significance two-tailed ¼ 0.000).

Mean total cost in the bootstrap analysis was 2584 EUR(median 2576, 97.5th percentile 3223 EUR; 2.5th percentile 1987EUR). Sex of the patient, number of years a patient lives with thedisease, and age at the time of hospitalization had no significantimpact on health care costs per patient. Hospital choice (β ¼−0.260; significance ¼ 0.002), however, was a strong determinantof health care costs. One-way analysis of variance also showed asignificant difference in mean costs between the two hospitalsinvolved (Welch statistics 19.222, P ¼ 0.000).

The results of the multivariate sensitivity analysis showedthat in the best-case (lowest cost) scenario, the average annualspending on drug treatment of a patient with CLL is 1659 EUR,and in the worst-case scenario, it is 2332 EUR. The deviation ofdrug costs does not exceed 12% on the negative side and 19% onthe positive side. The annual cost of hospitalization ranges from251 to 597 EUR per person and depends on the type of hospital atwhich a patient is treated.

Discussion

A literature review was conducted in the database PubMed toexplore whether our results were consistent with results fromstudies in other countries and to understand whether factors thatimpact the cost of other cancer conditions are similar to thoseaffecting CLL. The search was limited to a 10-year period of English-language articles studying multiple cancer conditions. The literaturereview showed that the major factors influencing the cost of cancerconditions are related to patients’ characteristics, such as stage atdiagnosis and stage at treatment, degree of comorbidity, age and

gender of a patient, and tumor site. Lal et al. [20], Longo et al. [21,22],and Yabroff et al. [23] recorded increased costs due to higher stageof the disease during treatment. Akushevich et al. [24], in aretrospective analysis on oncologic patients in the United States,determined that the highest costs exist in the period of treatmentimmediately after diagnosis. Yabroff et al. [23] also recorded thatboth the first stage and the last stage of the disease at the time oftreatment are associated with higher costs. The results of thestudies by Lal et al. [20] and Kuse et al. [25] demonstrated aconnection between the degree of comorbidity and treatment costs.The impact of patient’s age on the cost of cancer was significant in anumber of studies, but differed in scale and type of impact [20–22,24,25]. Some research described an impact of tumor site on totalcosts of the diseases [21,22,24]. Yabroff et al. [23] showed that costfor the treatment of localized diseases is lower, a conclusionsupported by Lai et al. [20] who noted that the highest costs werefor hematological malignancies among other types of cancer.

A limited number of economic analyses that describe factorsinfluencing the cost of CLL treatment specifically were found.These studies showed a positive correlation between age andcost of drug treatment [6,26]. Danese et al. [26] also concludedthat male gender is associated with higher cost for CLL drugtreatment.

Similar to Yabroff et al. [23], CLL phase-specific health carecosts for the US-Medicare population were found to have a U-shaped pattern over lifetime [27]. A study in the United States byLafeuille et al. [27] also reported significantly higher health carecosts for a CLL population compared to matched controls, mainlybecause of the higher costs for physicians, caregivers, andinpatient care. Besides inpatient hospital stay, pharmaceuticalswere the main cost drivers of CLL in a study in Germany [28],where the cost of treatment per case is about twice as high as thecost per case for highly prevalent diseases, such as chronicobstructive pulmonary disease or diabetes. This study alsorevealed that the average annual cost for patients with CLL from

Table 1 – Price values of the drugs for the cost and sensitivity analyses.

Drugs* Base-case price(EUR per mg)

Range used in sensitivityanalysis (EUR per mg)

Source†

Alemtuzumab (inj.) 0.4500 – Tariff in governmental purchases 2010Bleomycin (inj.) 1.5100 1.5100–1.6300 Distributors’ priceChlorambucil 0.0080 0.0080–0.2290 Tariff in governmental purchases 2010Cyclophosphamide (Adriablastine) 1.0600 0.6700–1.4500 Price, registered in the MOHCyclophosphamide (other generics) 0.0077 0.0010–0.0120 Tariff in governmental purchases 2010Dexamethasone (inj.) 0.0200 0.0040–0.0840 Price, registered in the MOHDexamethasone (po) 0.0002 – Price, registered in the MOHDoxorubicin (inj.) 0.1300 0.1300–0.4400 Tariff in governmental purchases 2010Etoposide phosphate (po) 0.1000 0.0980–0.1000 Price, registered in the MOHFludarabine (Fludara inj.) 3.7400 3.2200–3.7400 Tariff in governmental purchases 2010Fludarabine (Netran inj.) 0.7700 – Tariff in governmental purchases 2010Fludarabine (Netran po) 0.1500 – Tariff in governmental purchases 2010Fludarabine (other generics) 2.7800 0.7600–3.3700 Price, registered in the MOHMethylprednisolone (inj.) 0.1100 0.0140–0.1100 Price, registered in the MOHMethylprednisolone (po) 0.0260 0.0220–0.0300 Price, registered in the MOHMitoxantrone (inj.) 0.3800 0.3800–3.8300 Tariff in governmental purchases 2010Prednisolone (inj.) 0.0160 0.0130–0.0170 Price, registered in the MOHPrednisolone (po) 0.0072 0.0072–0.0077 Price, registered in the MOHRituximab (inj.) 2.0300 1.3500–2.9600 Tariff in governmental purchases 2010Vincristine (inj.) 3.6700 2.8600–4.0600 Tariff in governmental purchases 2010Vinblastine (inj.) 0.6300 0.4000–0.6300 Price, registered in the MOH

inj., injection; MOH, Ministry of Health; po, per os (by mouth).� Trade name is indicated if the product was prescribed specifically by it.† To value the use of the drugs from a health care perspective, we used a stepwise approach to determine an average price, depending on theavailability of information: tariff in governmental purchases (2010); price registered in the MOH; distributors’ price.


the sickness fund perspective decreased with increasing age until60 to 65 years, and thereafter increased.

Results of these US and German studies [27,28] differed fromthe results of our analysis conducted in Ukraine, where the costof medical care is relatively low and the major expenses are drugrelated. Similar to our results, a literature review conducted byStephens et al. [16] concluded that the cost of drug therapy is themain driver for CLL treatment costs, significantly exceedinghospitalization costs.

Previous research has suggested that the major factorsinfluencing the cost of cancer conditions are stage at diagnosisand stage at treatment, degree of comorbidity, age and gender,tumor site, and type of therapy received. Our study on aUkrainian sample from two specialized institutions, however,showed that only hospital choice had a significant impact on thecost of drug treatment. Possible explanations may be risk-patientselection or difference in treatment practice within the hospitals.High usage of injectable forms of drugs such as fludarabine anddexamethasone also was observed in this study. Because nohealth technology assessment agency currently exists, there areno recommendations comparing injectable and oral forms devel-oped in Ukraine. Nevertheless, the National Institute for Healthand Clinical Excellence [29] recommends giving preference to theoral form of fludarabine because of its higher efficiency. Ration-ality of the use of the injectable form of fludarabine in CLLtreatment practice may be a potential topic for further researchin Ukraine.

Ukraine is a country with a post-Semashko model of thehealth care system, and currently there is no state reimburse-ment system. Limited reimbursement for in-hospital treatment,however, is provided under governmental programs for suchdiseases as AIDS, tuberculosis, diabetes, cardiovascular diseases,

and cancer, among other diseases. These in-hospital statepurchases cover from 7% to 40% of oncologic patients’ needsdepending on the region and hospital type [30,31]. Majorexpenses on drugs are covered by patients’ out-of-pocket pay-ments. Thus, the high treatment cost of chronic conditions suchas CLL may be a significant economic burden, especially forpatients with low income.

The average annual cost of drug treatment for patients withCLL is 2047 EUR, with the majority of costs paid out of pocket.From December 1, 2011, the minimum annual subsistence levelin Ukraine is equal to 1155 EUR for people of working age and 920EUR for those who are retired. These figures are lower thanannual costs of drug treatment for patients with CLL in Ukraine,according to current clinical practice in specialized hospitals.This may impose a significant economic impact of the disease onvulnerable populations (e.g., elderly poor), taking into accountonly limited governmental subsidy. With high costs for thetreatment of hematologic malignancies [15,26], and insufficientreimbursement level for drug treatment in Ukraine [9], the treat-ment of CLL in specialized hospitals may be financially difficultfor economically unprotected patients because of hightherapy costs.

Implications

Our analysis indicates that there is likely to be a significantdifference in the practice of treating CLL within different hospi-tals of Ukraine, resulting in a significant deviation in drugexpenditures. Therefore, it is not clear whether treatment stand-ards are being followed within the hospitals and whether theschemes used are evidence based and rational. These issuesshould be explored further in future studies.

Table 2 – Drug utilization related to patients’ population characteristics.

Drugs* Patientsusing thedrug (%)

Average age (SD) ofpatients using thisitem (y)

n (%)Meanvolume perpatient (mg)

Men among oneswho are using thisdrug

Prescriptions duringthe first year oftreatment

Alemtuzumab (inj.) 10.3 57.90 � 9.25 9 (60.00) 0 (0.00) 87.14Chlorambucil (po) 7.6 74.45 � 4.61 5 (45.50) 6 (54.5) 4.76Cyclophosphamide

(Adriablastine inj.)4.8 64.29 � 5.85 6 (85.70) 4 (57.10) 7.24

Cyclophosphamide(all brand names)

66.2 62.95 � 9.13 59 (61.50) 29 (30.20) 2900.34

Dexamethasone(inj.)

10.0 60.07 � 8.71 19 (65.50) 7 (24.10) 22.70

Dexamethasone (po) 5.4 59.20 � 6.50 3 (60.00) 1 (20.00) 4.41Fludarabine (Fludara

inj.)44.4 60.56 � 8.82 36 (57.10) 13 (20.60) 377.87

Fludarabine (po) 3.4 59.80 � 8.35 3 (60.00) 0 (0.00) 30.48Fludarabine inj. (all

brand names)47.6 60.65 � 8.82 40 (58.00) 13 (18.80) 411.79

Methylprednisolone(po)

4.8 56.29 � 10.42 4 (57.1) 1 (14.3) 22.59

Mitoxantrone (inj.) 5.5 58.12 � 6.31 7 (87.50) 0 (0.00) 4.35Prednisolone (inj.) 5.5 60.63 � 10.64 5 (62.50) 2 (25.00) 224.63Prednisolone (po) 26.2 66.63 � 8.07 26 (68.40) 14 (36.80) 179.37Rituximab (inj.) 12.4 59.28 � 6.72 13 (72.20) 5 (27.80) 258.62Vincristine (inj.) 31.0 64.51 �9.57 34 (75.60) 17 (37.80) 1.66

inj., injection; po, per os (by mouth).� Trade name is indicated if a product was prescribed not by generic but by a trade name with a high frequency (for cyclophosphamide andfludarabine).


Limitations

Retrospective analysis allowed us to make an estimation oftreatment cost for CLL in specialized medical institutions ofUkraine and to explore its correlation with patient character-istics. Our research, however, suffered from several limitations.

First, our research does not allow us to assess the economicburden of CLL in Ukraine. It is expected that costs of treatment inthe current study may be higher than in the regional oncologydispensaries because of larger state subsidiaries and patients’expenditures on drugs.

Moreover, data on stage of the patient’s disease and healthstate on Eastern Cooperative Oncology Group (ECOG) criterionwere missing and thus excluded from the factor list. It is possible,however, that these factors may have an impact on CLL costs.

We conducted a linear regression analysis on logarithmic-transformed costs data while excluding six outliers on the basisof Cook’s distance. Disregarding these six observations may haveincreased the significance of the statistical analysis and thestrength of the relation between the independent variable (hos-pital choice) and the dependent variable (health care costs).

Conclusions

Drug treatment comprises the largest portion of total costs, whichpresumably may be a high economic burden for a patient with CLLwho is the major payer of treatment expenses in Ukraine. Costs ofdrug treatment significantly depend on the type of hospital selected.

Acknowledgment

We thank Carter Mandrik, PhD, for help in manuscript editing.Olena Mandrik is employed by MSD Ukraine and affiliated withErasmus University Rotterdam, Institute of Health Policy &Management by means of a PhD-hospitality agreement. Theviews expressed in this article are those of the authors andshould not be attributed to the authors’ employers.

Source of financial support: The authors have no other financialrelationships to disclose.

R E F E R E N C E S

[1] Dranitsaris G, Truter I, Lubbe MS, et al. Advances in cancertherapeutics and patient access to new drugs. Pharmacoeconomics2011;29:213–24.

[2] National Cancer Institute. National Cancer Register. 2011;12. Availablefrom: http://users.iptelecom.net.ua/�ucr/eng/index_e.htm. [AccessedDecember 3, 2011].

[3] Cheson BD, Bennett JM, Grever M, et al. National Cancer Institutesponsored working group guidelines for chronic lymphocytic leukemia:revised guidelines for diagnosis and treatment. Blood 1996;87:4990–7.

[4] Kalil N, Cheson BD. Management of chronic lymphocytic leukaemia.Drugs Aging 2000;16:9–27.

[5] Redaelli A, Botteman MF, Stephens JM, et al. Economic burden of acutemyeloid leukemia: a literature review. Cancer Treat Rev 2004;30:237–47.

[6] Rozman C, Montserrat E. Chronic lymphocytic leukemia. N Engl J Med1995;333:1052–7.

[7] American Cancer Society. Cancer facts and figures. 2010. Availablefrom: http://www.cancer.org/acs/groups/content/@nho/documents/document/acspc-024113.pdf. [Accessed February 7, 2012].

[8] National Cancer Institute. Surveillance, Epidemiology, and End ResultsProgram, 1975–2005, Division of Cancer Control and PopulationSciences. 2008. Available from: http://seer.cancer.gov/csr/1975_2005/.[Accessed February 7, 2012].

[9] [Achievements and challenges of the domestic haematology].Ukraiinskiy medyuchnuy chasopyus. 2011. Available from: http://www.umj.com.ua/wp-content/uploads/2011/06/Hematolodji.pdf.[Accessed March 9, 2013].

[10] Order of the Ministry of Health of Ukraine #647 from 30.07.2010. Aboutapproval of the clinical protocols on medical help provision forpatients on specialty “Hematology.”.

[11] Order of the Ministry of Health of Ukraine #769 from 13.09.2010.About approval of the concept of pharmaceutical sector developmentin health care branch in Ukraine for the years 2011–2020.

[12] Levin TT, Li Y, Riskind J, Rai K. Depression, anxiety and quality of life ina chronic lymphocytic leukemia cohort. Gen Hosp Psychiatry2007;29:251–6.

[13] Shanafelt TD, Bowen D, Venkat C, et al. Quality of life in chroniclymphocytic leukemia: an international survey of 1482 patients. Br JHaematol 2007;139:255–64.

[14] Goor KM, Schaafsma MR, Huijgens PC, van Agthoven M. Economicassessment on the management of chronic lymphocytic leukaemia.Expert Opin Pharmacother 2005;6:1179–89.

[15] Kasteng F, Sobocki P, Svedman C, Lundkvist J. Economic evaluations ofleukemia: a review of the literature. Int J Technol Assess Health Care2007;23:43–53.

[16] Stephens JM, Gramegna P, Laskin B, et al. Chronic lymphocyticleukemia: economic burden and quality of life: literature review. Am JTher 2005;12:460–6.

[17] Mandrik O, Zalis’ka O. [Assessment of pharmacoeconomic aspectsof chronic lymphocytic leukemia treatment in Ukraine].Upravlinnya, economica, ta zabezpechenya yakosti v farmacii2012;3:62–7.

[18] Gorlyachenko O, Shulgay AG, Gorlyachenko A, et al. [Cost of medicalhelp]. Novosti medicinyu I farmacii 2012;5(403): Available from:http://www.mif-ua.com/archive/article_print/27173. [Accessed March9, 2013].

[19] Reforming of the secondary medical help in Ukraine: basicproblematic and solution options]. Combined report on the projectEuropeAid/123236/C/SER/UA. EPOS Health Consultants/NI-CO/ECORY. Available from: Shttp://www.eu-shc.com.ua/UserFiles/File/SR_V04_ua.pdf. [Accessed March 9, 2013].

[20] Lal A, Bhurgri Y, Rizvi N, et al. Factors influencing in-hospital lengthof stay and mortality in cancer patients suffering from febrileneutropenia. Asian Pac J Cancer Prev 2008;9:303–8.

[21] Longo CJ, Deber R, Fitch M, et al. An examination of cancer patients’monthly ‘out-of-pocket’ costs in Ontario, Canada. Eur J Cancer Care(Engl) 2007;16:500–7.

[22] Longo CJ, Fitch M, Deber RB, Williams AP. Financial and family burdenassociated with cancer treatment in Ontario, Canada. SupportCare Cancer 2006;14:1077–85.

[23] Yabroff KR, Lamont EB, Mariotto A, et al. Cost of care forelderly cancer patients in the United States. J Natl Cancer Inst2007;99:14–23.

[24] Akushevich I, Kravchenko J, Akushevich L, et al. Medical cost trajectoriesand onsets of cancer and noncancer diseases in US elderly population.Comput Math Methods Med 2011. http://dx.doi.org/10.1155/2011/857892.

[25] Kuse R, Colberg H, Marbé W, et al. Which factors render cost-covering lump-sum charging difficult for the treatment of patientswith acute leukemias? Onkologie 2001;24:292–4.

[26] Danese M, Gleeson M, Reyes C, et al. Cost of chronic lymphocyticleukemia (CLL) in Medicare patients. J Clin Oncol 2008;26(Suppl): abstr17531.

[27] Lafeuille MH, Vekeman F, Wang ST, et al. Lifetime costs to Medicare ofproviding care to patients with chronic lymphocytic leukemia. LeukLymphoma 2012;53:1146–54.

[28] Blankart CR, Koch T, Linder R, et al. Cost of illness and economicburden of chronic lymphocytic leukemia. Orphanet J Rare Dis 2013;8:32.

[29] Guidance on the Use of Fludarabine for B-Cell Chronic LymphocyticLeukaemia. NICE technology appraisal guidance TA29. London, UK:National Institute for Health and Care Excellence, 2001. Available from:http://publications.nice.org.uk/guidance-on-the-use-of-fludarabine-for-b-cell-chronic-lymphocytic-leukaemia-ta29/guidance. [Accessed March 9, 2013].

[30] [On efficiency of costs use by the Ministry of Health]. State financeinspection of Ukraine. Available from: http://www.dkrs.gov.ua/kru/uk/publish/article/75100;jsessionid=5305963B5762BAF2C3E27BEAC3E2A6F6.[Accessed March 9, 2013].

[31] Adamchuk I. [The program exists, where is effectiveness?]. Statefinance inspection of Ukraine. Available from: http://www.dkrs.gov.ua/kru/uk/publish/printable_article/51727;jsessionid=BB368B7E53A135A1BF1A7994BC73505E. [Accessed March 9, 2013].






http://users.iptelecom.net.ua/~ucr/eng/index_e.htm

http://users.iptelecom.net.ua/~ucr/eng/index_e.htm











http://www.cancer.org/acs/groups/content/@nho/documents/document/acspc-024113.pdf

http://www.cancer.org/acs/groups/content/@nho/documents/document/acspc-024113.pdf



http://seer.cancer.gov/csr/1975_2005/


http://www.umj.com.ua/wp-content/uploads/2011/06/Hematolodji.pdf

http://www.umj.com.ua/wp-content/uploads/2011/06/Hematolodji.pdf




























http://www.mif-ua.com/archive/article_print/27173




Shttp://www.eu-shc.com.ua/UserFiles/File/SR_V04_ua.pdf

Shttp://www.eu-shc.com.ua/UserFiles/File/SR_V04_ua.pdf













10.1155/2011/857892















http://publications.nice.org.uk/guidance-on-the-use-of-fludarabine-for-b-cell-chronic-lymphocytic-leukaemia-ta29/guidance




http://www.dkrs.gov.ua/kru/uk/publish/article/75100;jsessionid=5305963B5762BAF2C3E27BEAC3E2A6F6

http://www.dkrs.gov.ua/kru/uk/publish/article/75100;jsessionid=5305963B5762BAF2C3E27BEAC3E2A6F6


http://www.dkrs.gov.ua/kru/uk/publish/printable_article/51727;jsessionid=BB368B7E53A135A1BF1A7994BC73505E





Costs of Medically Attended Acute Gastrointestinal Infections: The PolishProspective Healthcare Utilization SurveyMarcin Czech, PhD, MD, MBA1,*, Magdalena Rosinska, PhD, MD2, Justyna Rogalska, MSc2, Ewa Staszewska, MSc2,Pawel Stefanoff, PhD, MD, MSc2

1Faculty of Pharmacy, Department of Pharmacoeconomics, Medical University of Warsaw, Warsaw, Poland; 2Department of Epidemiology, National Institute ofPublic Health–National Institute of Hygiene, Warsaw, Poland

A B S T R A C T

Objectives: The burden of acute gastrointestinal infections (AGIs)on the society has not been well studied in Central Europeancountries, which prevents the implementation of effective, targetedpublic health interventions. Methods: We investigated patients of11 randomly selected general practices and 8 hospital units. Eachpatient meeting the international AGI case definition criteria wasinterviewed on costs incurred related to the use of health careresources. Follow-up interview with consenting patients was con-ducted 2 to 4 weeks after the general practitioner (GP) visit ordischarge from hospital, collecting information on self-medicationcosts and indirect costs. Costs were recalculated to US dollars byusing the purchasing power parity exchange rate for Poland.Results: Weighting the inpatient costs by age-specific probabilityof hospital referral by GPs, the societal cost of a medically attendedAGI case was estimated to be US $168. The main cost drivers of

direct medical costs were cost of hospital bed days (US $28), cost ofoutpatient pharmacotherapy (US $20), and cost of GP consultation(US $10). Patients covered only the cost of outpatient pharmaco-therapy. Considering the AGI population GP consultation rate, theage-adjusted societal cost of medically attended AGI episodes wasestimated at US $2222 million, of which 53% was attributable toindirect costs. Conclusions: Even though AGIs generate a low costfor individuals, they place a high burden on the society, attributedmostly to indirect costs. Higher resources could be allocated to theprevention and control of AGIs.Keywords: direct medical costs, direct nonmedical costs, gastrointestinalinfections, indirect costs, Poland.


Introduction

Acute gastrointestinal infections (AGIs) can be caused by viruses,bacteria, and parasites, as well as through physical or chemicalintoxications. Symptoms and high incidence of AGIs may put asubstantial burden on patients and the health care system—frommedical, social, and economic perspectives. An estimation of the trueburden of AGI symptoms on the society, however, is difficult [1]. Indeveloped countries, AGIs are common but usually do not causesevere disease. Sufferers frequently downplay its significance, anddoctors often do not trace the causes of individual cases. Themajority of AGIs in Poland are not treated at all or are treated withrehydration and/or over-the-counter drugs [2,3]. For these reasons,economic consequences of AGIs are not properly assessed and thustheir true burden is underestimated.

Available scarce evidence indicates that the societal cost of theso-called mild gastrointestinal illnesses is considerably higher thanthe costs associated with acute hospitalized cases [4]. To date,high-quality evidence on burden and costs to the society has beencollected mostly for rotavirus-associated AGIs—in relation to bothambulatory care [5,6] and hospital settings [7–9]. Costs for AGIs

caused by other pathogens have not been well documented,although recent evidence indicates that costs for AGIs caused bydifferent etiological factors can be similar [10].

Poland, located in Central Europe, with its 38.4 million habitants isthe 34th most populous country in the world and the 6th largestcountry in the European Union. The gross national product per capitais almost US $16,710 [11]. The overall quality of health care provisionnationwide, as judged by European standards, is regarded as beinghigh, which is reflected in the nation’s average life expectancy,estimated at 71 years for men and 80 years for women [11]. Poland’shealth care system is based on an all-inclusive social insurancesystem. An insured person and members of his or her family areentitled to free health services offered by providers who have signedcontracts with the National Health Fund. The National Health Fund isa state-owned third-party payer. It finances health services andassures reimbursement of medicines. There is a rapidly growingprivate sector consisting of private general practitioners’ (GPs’) andspecialists’ practices and rarely private clinics and hospitals, paid ona fee-for-service or prepaid basis. The drugs reimbursement systemgrants unrestricted access to essential drugs and different level of co-payment for the remaining specialty medicines (depending on the






� Address correspondence to: Marcin Czech, Faculty of Pharmacy, Department of Pharmacoeconomics, Medical University of Warsaw,Zwirki i Wigury 81, 02-091 Warsaw, Poland.


dx.doi.org/10.1016/j.vhri.2013.06.011

dx.doi.org/10.1016/j.vhri.2013.06.011

dx.doi.org/10.1016/j.vhri.2013.06.011


dx.doi.org/10.1016/j.vhri.2013.06.011

Anatomical Therapeutic Chemical [ATC] Classification System group,disease, patient’s status, etc.).

Currently in Poland, no reliable cost data in health care systemare available [12]. These data are not recorded by health careinstitutions, because GP practices are reimbursed for the numberof registered patients and hospital admissions are reimbursed onthe basis of discharge codes of the International Statistical Classifica-tion of Diseases, 10th Revision, not depending on the health careprocedures used. Consequently, there are no health care costdatabases in Poland that are similar to those in other developedcountries, such as the National Health Service database in theUnited Kingdom [13].

The aim of the present article was to summarize directmedical, nonmedical, and indirect costs related to the manage-ment of AGI cases to estimate its societal costs and to enable

better planning and monitoring of AGI prevention programs inthe future.

Methods

We estimated AGI-related resource use linked to GP consultationsand hospitalizations by using a prospective study conductedbetween May 2008 and September 2009. We have describeddetailed methods for study site selection and AGI cases recruit-ment in previous publication [3].

Studied Population

The Polish Prospective Healthcare Utilization Survey was per-formed in a randomly selected population sample served by 11

Table 1 – List of unit costs generated by AGI cases consulting GPs in Poland between May 2008 and September 2009.

Detailed cost categories Number Unit cost per patient (US $)� Total cost (US $)�

Mean 95% CI Median 95% CIPaid bypayer†

Paid bypatient

GP consultation 385/385 10.25 9.64–10.86 8.14 8.05–8.21 3946.54 0.00Laboratory testing 14/385 19.26 13.91–24.62 19.46 12.40–26.52 269.70 0.00Complete blood count 10/385 3.52 3.02–4.03 3.74 2.83–4.65 35.22 0.00Stool sample for microbiological

confirmation9/385 14.17 11.14–17.21 13.51 9.17–17.85 127.57 0.00

Other (blood smear, electrolytes,etc.)

10/385 10.69 7.58–13.79 10.24 6.45–14.03 106.91 0.00

Imaging diagnostics(ultrasonography)

11/385 17.28 14.46–20.10 13.51 6.60–20.43 190.11 0.00

Pharmacotherapy (ATC) 382/385 20.71 19.86–21.55 19.33 18.51–20.15 985.11 6924.65Rehydration fluids (−) 230/385 6.72 6.18–7.25 5.06 4.76–5.37 442.67 1102.28Intestinal anti-infectives (A07A) 229/385 5.77 5.53–6.02 4.75 4.02–5.47 0.00 1322.01Antidiarrheal microorganisms

(A07F)144/385 8.30 8.03–8.57 8.83 – 0.00 1195.4

Other probiotics (−) 89/385 8.97 8.14–9.84 8.03 – 0.00 798.83Other antidiarrheals (A07X) 127/385 7.59 7.48–7.70 7.53 – 0.00 963.81Drugs for functional

gastrointestinal disorders (A03)81/385 5.12 4.36–5.89 4.59 4.37–4.81 41.81 373.22

Analgesics and anti-inflammatoryproducts (N02B, M01A)

82/385 4.02 3.28–4.77 1.74 0–3.58 67.32 262.42

Antibacterials for systemic use (J01) 60/385 8.02 6.68–9.35 7.35 6.85–7.85 159.89 320.94Other pharmaceuticals (A01A,

A02B, A07D, A09A, R05C, R06A,V06D, no ATC code)

68/385 10.63 8.82–12.44 8.47 3.32–13.62 273.42 449.19

Materials used 66/385 0.34 0.17–0.51 0.09 0.04–0.14 22.30 0.00Medical devices (syringes, needles,

ports, fluid transfusion sets)19/385 0.37 0.19–0.55 0.15 0.12–0.28 7.01 0.00

Biological specimen containers 4/385 1.29 0.11–2.46 0.95 0–2.43 5.16 0.00Other (gloves, swabs, disposable

towels, gowns, disinfectants,soaps, etc.)

66/385 0.15 0.10–0.20 0.09 0.04–0.14 10.14 0.00

Emergency department visits 4/115 54.05 – 54.05 – 216.22 0.00Specialist consultation 7/115 17.78 – 17.78 – 124.45 0.00Transport (number of kilometers) 82/115 36.96 17.91–56.01 13.00 9.21–16.79 – –

Cost per kilometer � number ofkilometers

82/115 16.84 8.02–25.66 6.32 4.26–8.39 0.00 1380.85

Home care (family, friends, service) 34/115 127.98 64.37–29.42 62.89 36.11–89.66 4,351.30 54.05Absence from work (number of days

off work)10/115 6.10 3.89–8.31 6.00 3.03–8.97 – –

Cost per day off � number of days 10/115 511.48 330.91–692.04 503.09 262.75–743.44 5,114.77 0.00

AGI, acute gastrointestinal infection; ATC, Anatomical Therapeutic Chemical Classification System; CI, confidence interval; GP, generalpractitioner; WHO, World Health Organization.� US $1 ¼ 1.85 PLN (WHO CHOICE).† State institution (National Health Fund or social insurance).


Table 2 – List of unit costs generated by AGI cases admitted to a hospital in Poland between May 2008 and September 2009.

Detailed cost categories Number Unit cost per patient (US $)* Total cost (US $)*

Mean 95% CI Median 95% CI Paid by payer† Paid by patient

Number of bed days of hospitalization 504/504 4.74 4.41–5.06 4.00 3.86–4.14 – –

Cost per bed day � number of bed days 504/504 832.63 782.84–882.43 688.84 644.99–732.70 419,647.46 0.00Laboratory testing 503/504 49.74 47.36–52.10 47.84 44.46–50.95 25,015.72 0.00

Complete blood count 497/504 4.03 3.86–4.19 4.05 3.96–4.15 2,002.53 0.00Blood smear 384/504 3.27 3.14–3.40 3.24 – 1,254.95 0.00C-reactive protein 310/504 4.51 4.29–4.72 3.62 2.08–5.16 1,396.70 0.00Erythrocyte sedimentation rate 479/504 2.34 2.24–2.44 2.16 – 1,120.80 0.00Electrolytes 494/504 6.57 6.17–6.98 4.59 – 3,248.30 0.00Urea 308/504 3.24 3.10–3.38 3.24 3.21–3.28 999.24 0.00Creatinine 385/504 2.76 2.62–2.89 2.16 1.63–2.70 1,062.27 0.00Glucose 325/504 4.21 3.03–5.39 2.16 2.05–2.27 1,369.04 0.00Gasometry 149/504 5.13 4.47–5.79 3.24 2.49–3.99 764.65 0.00General urine examination 467/504 3.7 3.54–3.86 3.78 3.26–4.30 1,728.10 0.00Stool sample for microbiological confirmation 351/504 13.68 12.89–14.46 9.73 9.26–10.20 4,800.70 0.00Stool sample 78/504 5.83 5.22–6.43 4.86 – 454.35 0.00Blood culture 42/504 13.99 10.83–17.15 9.73 9.33–10.13 587.57 0.00Nasopharyngeal sample for microbiological confirmation 50/504 8.55 7.43–9.67 8.65 5.48–11.82 427.57 0.00Other 288/504 13.19 12.1–14.28 10.27 8.18–12.36 3,798.95 0.00

Imaging diagnostics 251/504 32.29 28.89–35.69 24.32 23.16–25.49 8,104.74 0.00Abdominal ultrasonography 178/504 18,00 16.63–19.37 20,00 19.15–20.85 3,204.39 0.00Chest radiography 96/504 13.19 11.42–14.96 9.73 8.51–10.95 1,266.35 0.00Electrocardiography 109/504 10.06 8.29–11.84 5.41 2.88–2.52 1,096.91 0.00Other 84/504 30.21 24.43–35.98 23.94 18.61–29.26 2,537.09 0.00

Pharmacotherapy 504/504 15.57 13.98–17.15 9.70 8.70–10.71 7,798.49 0.00Rehydration fluids 492/504 5.72 5.25–6.18 4.25 3.69–4.81 2,813.16 0.00Intestinal anti-infectives (A07A) 163/504 3.09 2.58–3.6 2.19 1.92–2.47 503.55 0.00Antidiarrheal microorganisms (A07F) 156/504 1.80 1.45–2.15 1.10 0.87–1.32 280.50 0.00Other antidiarrheals (A07X) 39/504 2.43 1.98–2.87 2.22 1.57–2.87 94.66 0.00Drugs for functional gastrointestinal disorders (A03) 55/504 1.68 1.30–2.06 1.17 0.83–1.51 92.15 0.00Analgesics and anti-inflammatory products (N02B, M01A) 136/504 1.36 1.08–1.65 0.63 0.35–0.90 185.37 0.00Antibacterials for systemic use (J01) 179/504 15.08 13.06–17.09 11.73 10.41–13.05 2,698.90 0.00Other pharmaceuticals (A02A, A02B, A05B, A06A, A07C, A07D, A07E,

A11G, A12B, B05A, H02A, J05A, N03A, N06B, R05C, R06A, V06D,V07A, no ATC code)

338/504 3.34 2.74–3.95 1.71 1.51–1.90 1,130.20 0.00

Specialist consultation 71/504 17.78 15.22–20.33 16.22 15.42–17.02 1,262.16 0.00Materials used 504/504 30.26 25.99–34.53 12.42 10.72–14.13 15,250.69 0.00Medical devices (syringes, needles, fluid transfusion sets) 504/504 3.89 3.59–4.19 2.98 2.70–3.25 1,959.83 0.00Containers and test tubes (urine and other biological specimen

containers)475/504 1.23 1.11–1.36 0.92 0.88–0.97 586.27 0.00

VALUE

INH

EALTH

REGIO

NAL

ISSUES

2(2

013)210–217

212

ambulatories and an independent population served by 8 hospi-tal units located in 6 Polish voivodships (14.4 million inhabitants,38% of Poland’s population). During the study period, localcoordinators recruited patients meeting inclusion criteria. Theinclusion and exclusion criteria used in the present study werecompatible with the AGI definition proposed by the InternationalCollaboration on Enteric Disease Burden of Illness [14] and usedin the recent study of AGI prevalence in the community [15]. TheEthical Committee at the National Institute of Public Health inWarsaw reviewed and accepted the study protocol.

Collection of Data on Costs

For each patient, local coordinators filled a standardized ques-tionnaire on health care resource utilization that was developedspecifically for this study. The questionnaire recording informa-tion on GP visits consisted of 18 items, which were divided intogeneral demographic information, clinical presentation, use ofpharmaceutical preparations, diagnostic tests, materials, andspecialist consultations. For each health care resource used, amonetary value was allocated. The hospital resource utilizationquestionnaire consisted of 26 items, which were divided intogeneral demographic and clinical information as well as patientmanagement at the admission unit and separately at the hospitalward. The local coordinators fill the information on monetarycost incurred on the basis of internal unit prices.

Study coordinators interviewed all consenting patients 2 to 4weeks after GP visit or discharge from hospital. A structuredtelephone interview comprised 16 questions related to thecurrent occupation, effect of the disease on daily activities,absence from work, and use of home care and transport inrelation to AGIs, further GP or specialist consultations, admis-sions to the hospital, and use of prescription and over-the-counter medications and diagnostic tests.

Cost Analysis

We obtained reference values for the calculation of costs fromofficial state publications for the period January to December 2009.We summarized detailed use of health care resources separately forGP consultations and hospital admissions. We collected data ontype, form, and quantity of pharmaceuticals, medical devices,diagnostic procedures, laboratory tests, referrals to specialists andemergency departments, as well as on hospitalizations and groupedthem by using a bottom-up microcosting approach [16]. GP con-sultation cost and hospital cost of bed day was obtained from aquery of studied health care facilities accounting departments byusing accounts breakdown [17], dividing cost items into personnel,building, equipment, administration, food, cleaning, laundry, main-tenance, and other costs. To calculate the cost of GP consultation,we summed monthly costs pertaining to the maintenance of the GPpractice and divided them by the number of physicians working inthe health unit and their working time (cost per minute of GPconsultation). To calculate the cost of hospitalization bed day, wedivided the monthly costs of the hospital ward maintenance by thenumber of patients admitted to the hospital ward. We obtainedcosts of pharmaceuticals (ATC coding), medical devices, and othermedical materials from purchasing units in hospitals. Emergencydepartment costs were taken from an official diagnosis-relatedgroup system. For the imputation of quantitative missing data, weused the mean values from the current study data set or averageprices from participating health unit price lists. Missing values forpharmaceutical costs were obtained from the national pharmaceut-ical formulary [18]. If only international names were provided byrespondents, then we would impute the values on the basis of pricesof the most commonly used generics. Transportation costs werebased on official mileage rates [19]. We calculated indirect costs

Disinfectan

tsan

dso

aps

501/50

41.93

1.62

–2.24

0.70

0.56

–0.85

967.55

0.00

Oth

er(glove

s,sw

abs,

cotton

balls,

pam

pers,

disposa

bletowels,

gowns,

etc.)

504/50

423

.29

19.66–

26.91

7.37

6.44

–8.30

11,737

.04

0.00

Directmed

ical

costs

504/50

494

6.59

893.92

–99

9.25

803.70

733.41

–87

3.98

477,07

9.26

0.00

Transp

ort

(numbe

rofkilometers)

119/14

512

6.13

86.76–

165.51

50.00

31.15–

68.85

––

Cost

per

kilometer

�numbe

rofkilometers

119/14

557

.57

40.10–

75.04

22.59

14.34–

30.84

0.00

6850

.80

Directno

nmed

ical

costs

119/14

557

.57

40.10–

75.04

22.59

14.34–

30.84

0.00

6850

.80

Homeca

re(fam

ily,

friends,

service)

54/145

356.94

257.05

–45

6.83

251.55

179.19

–32

3.90

19,274

.71

0.00

Abs

ence

from

work

(only

patients

aged

18–65

y)–

––

––

––

Dueto

hosp

italization

13/145

202.71

141.36

–26

4.06

179.68

120.54

–23

8.82

2,63

5.24

0.00

After

hosp

italization

3/14

564

2.84

329.75

–95

5.93

838.49

329.31

–11

30.18

1,92

8.52

0.00

Indirect

costs

70/145

360.34

261.30

–45

9.39

251.55

161.06

–34

2.03

23,838

.47

0.00

AGI,ac

ute

gastro

intestinal

infection;ATC,Anatomical

Therap

euticChem

ical

Classifica

tion

System

;CI,co

nfiden

ceinterval;W

HO,W

orldHea

lth

Organ

ization.

*US$1

¼1.85

PLN

(WHO

CHOIC

E).

†Stateinstitution

(National

Hea

lth

Fundorso

cial

insu

rance

).


related to absence from work and time devoted to care by using ahuman capital approach based on average wages lost [16]. We didnot take into account reduced productivity owing to presenteeism orpremature death.

We used three perspectives for cost calculations: household/patient, public payer, and societal. We applied a standard costdivision into direct medical, nonmedical, and indirect costs [16].We estimated separately costs of outpatient and inpatient care,stratified by age group (o4, 5–18, 19–65, and 465 years). For eachcost category, we calculated the mean, median, and total cost. Forthe mean and median, we computed confidence intervals byusing 1000 bootstrap percentile replications.

For the estimation of the overall cost of medically attendedAGI cases, we weighted the costs related to hospitalization byage-specific proportion of patients referred to the hospital. Thesefigures were based on two recent Polish studies of AGI burden[3,15]. All costs collected in the local currency (PLN) wererecalculated into US dollars by using the purchasing power parityexchange rate for Poland [20]. All analyses and data manipula-tions were done in Stata version 10 (StataCorp LP, College Station,TX) [21].

Estimating Population AGI Burden

For the estimation of AGI monetary costs in the society (totalnational cost, TNC), we developed a stochastic model by using@RISK 5.0 (Palisade Corporation, Ithaca, NY), a Monte Carlosimulation add-in to Microsoft Excel:

TNC ¼ ∑age group

pop n IGP n CGP þ pop n IHOSP n CHOSP

where age group is either 0 to 17 and 18 years or more, pop is thepopulation, IGP is the rate of GP visits per person-year, IHOSP is thehospital admission rate per person-year, CGP is the societal costper GP visit, and CHOSP is the societal cost per hospital admission.All parameters are age-group specific.

Our study was designed to collect data on individual costs ofAGI cases at different levels, and it was not feasible to deriveincidence and follow-up data directly. Therefore, to obtainpopulation estimates of AGI-related GP visits and hospitaliza-tions, we related our data to a population-based survey per-formed at the same time [15]. We selected this particular surveyowing to large sample size, similar time span, and geographiccoverage.

Annual rates of AGI-related GP visits and hospital admissionswere represented by beta distributions. The cost estimates per GPvisit and hospital admission were represented by log-logisticdistributions. The distribution parameters were estimated onthe basis of sample cost distributions by using maximum like-lihood estimators. The national annual cost of medicallyattended AGIs was calculated by using Monte Carlo simulationwith the above-mentioned probability distributions and the 2009population estimate for Poland. The model was run for 150,000iterations to stabilize the output distributions. The mean and 95%uncertainty limits were reported.

Results

During the study period, valid questionnaires were collected on385 GP visits, including 113 follow-up interviews, and on 504hospital admissions, including 145 follow-up interviews. Adetailed description of the selection of study sites and studypopulation is presented elsewhere [3].

Direct Medical Cost

Unit costs related to GP consultations of AGI cases are listed inTable 1. Overall, direct medical costs amounted to US $34 per AGIpatient consultation. Prescribed pharmaceuticals, paid in 88% bypatients, were the main cost driver in relation to GP consulta-tions. Particularly, pharmaceuticals of the A07 ATC group (anti-diarrheal and intestinal anti-inflammatory/anti-infective agents)were not reimbursed. However, an important share of anti-inflammatory and analgesic drugs (M01A, N02B) and antibiotics(J01) costs was reimbursed by the state. The next cost driver inprimary care were GP visits’ costs, equal to US $11, entirely paidby the state payer. Laboratory testing and diagnostic procedureswere relatively cheap for an average AGI patient owing to theirrare utilization (US $0.7 and US $0.5, respectively) but substantialin terms of unit cost values (US $19 and US $17, respectively).

Unit costs related to the hospital treatment of AGI cases arelisted in Table 2. Inpatient care generated on average US $947 ofdirect medical costs, including US $833 of bed days (average costper bed day was US $176). The other cost drivers in inpatient carewere costs of laboratory diagnostics, materials, imaging diagnos-tics, and pharmacotherapy. All direct medical costs were coveredby the state-owned third-party payer.

Taking into account the summed costs generated by medicallyattended AGI cases, direct medical costs were highest for patientsaged 5 to 18 years (Table 3). The excess cost was related mostly tothe higher cost of hospitalization bed day in pediatric wards thanin adult wards (US $226 vs. US $117), and substantially lowerprobability of hospitalization among 18- to 64-year-old AGI cases.

Direct Nonmedical Costs

The only nonmedical direct cost considered in our study was thecost of transportation in relation to health care services. Inoutpatient care, the average cost of reaching the service (averagedistance of 20 km) amounted for US $17 (Table 1). In relation toinpatient care, this cost was higher (mean cost per patient was US$58) (Table 2) owing to longer distance to the hospital andcommon necessity of several trips to the hospital undertakenby primary carers. The costs of transportation were higher forpersons younger than 18 years, which was especially marked inrelation to inpatient care-related costs (Table 3). All nonmedicaldirect costs were covered by the patients.

Indirect Costs

The indirect costs assessed in our study related to home care(average cost per hour was US $10 ) and absence from work inrelation to AGI (average cost of 1 day off work was US $84 ). Inrelation to outpatient care, the average cost of home care of 34patients (on average 13 hours of care) was US $128 and theaverage cost of absence of 10 patients from work (on average 6days off work) was US $511 (Table 1). In relation to inpatient care,the average cost of home care of 54 patients was estimated at US$357 and the average cost of absence from work of 13 patientswas US $351 (Table 2). Costs of home care were substantiallyhigher for persons younger than 18 years in outpatient care andwere highest for persons older than 65 years in relation toinpatient care (Table 3). Costs of absence from work were almostexclusively limited to patients aged 18 to 64 years and were by farthe most important cost driver in this age group overall.

Cost Perspectives and Estimated AGI Societal Burden

The average cost of one medically attended AGI case in Poland,taking into account patient management and flow of patients fromprimary to secondary care observed in population-based studies, wasestimated at US $136 from the state-owned third-party payer, US $32


from household, and US $168 from societal perspective (Table 4). Themajority of this cost is paid by the state. The estimated societal costof medically attended AGI episodes amounted to US $2222 million(95% uncertainty interval was US $734 to US $6367 million).

Discussion

The costs of medically attended AGI episodes in Poland wereexamined in primary and secondary care, from state, household,and societal perspectives, in terms of direct and indirect costs. Wedid not assess separately costs attributed to AGI etiological agentsbecause the evidence from the literature showed that their directcosts are in a similar range [10]. When weighting the inpatient costsby age-specific probability of hospital referral by GPs, we estimatedthe societal cost of a medically attended AGI case amounting to US$168. In terms of direct medical costs, the main cost drivers werecost of hospital bed days, cost of GP consultation (both paid by thestate-owned third-party payer), and cost of outpatient pharmaco-therapy, covered in majority by patients. Overall, indirect costsprevailed over direct costs (US $88 vs. US $80), which was mostlypronounced among adults in the productive age, where 76% of thecosts were attributed to indirect costs. Based on the GP consultationrate estimated in a parallel population-based telephone survey [15],the estimated age-adjusted societal cost of medically attended AGIepisodes would be US $1681 million.

International comparisons of economic costs between countriesare difficult from the methodological point of view. The societalcosts of AGI cases depend on the disease prevalence, differences inhealth-seeking behaviors, and health care system organization.Historically, different AGI case definitions were used, differentcurrency exchange rates were applied, and different cost categorieswere taken into account in the cost calculations. Estimates of AGIcosts referring to the general population are sparse. The societalcosts of medically attended AGI cases were estimated to be US$3510 million in the United States [22], which, considering aconservative assumption of 2% annual discounting until June 30,2009, would give US $688 for a medically attended AGI case.Another study estimated the societal costs of all AGI cases inNew Zealand at NZ $89 million [23], which would be equivalent toUS $359 per AGI episode, as of 2009. Straightforward comparisonswith developing countries are difficult because only estimates ofcosts for AGI hospitalization costs were published. The societal costof rotavirus hospitalization ranged from US $36 in Vietnam [24], US$77 in Uzbekistan [25], US $81 in one center in India [10] to US $87in Kyrgyzstan [26]. The study on the hospital cost of bacterialdiarrhea in Thailand showed similar numbers: US $77 per inpatientepisode [8]. The above figures are much lower compared with ourestimate of US $1078 per hospital admission. These differencesreflect the important disparities in the cost of health care servicesbetween the developed and developing world and are concordantin great extent with World Health Organization estimates fordifferent World Health Organization regions [20].

Table 3 – Average societal costs of medically attended AGI cases, by cost category and age group in Poland between May2008 and September 2009.

Cost category (US $)* 0–4 y 5–18 y 19–64 y 65þ y All age groups

OutpatientAverage cost of GP visit 10.94 10.15 9.62 10.42 10.25Laboratory testing 0.89 0.49 0.78 0.00 0.70Imaging diagnostics 0.16 0.69 0.73 0.76 0.49Pharmacotherapy 21.19 19.35 20.71 21.18 20.54Materials 0.06 0.09 0.05 0.01 0.06Emergency department visit 0.00 3.86 3.28 0.00 1.88Specialist consultation 0.67 0.36 0.00 0.00 0.32

Inpatient†

Average cost of hospitalization 38.56 57.20 4.35 22.56 28.11Laboratory testing 1.97 3.68 0.36 1.29 1.68Imaging diagnostics 0.35 0.79 0.14 0.99 0.54Pharmacotherapy 0.50 0.77 0.15 0.63 0.52Materials 1.56 1.36 0.21 0.74 1.02Specialist consultation 0.15 0.21 0.01 0.02 0.08

Total direct medical costs 77.00 99.00 40.39 58.60 66.19Transport related to GP visit 12.89 10.59 12.80 1.51 12.01Transport related to hospitalization‡ 2.27 3.39 0.06 0.49 1.60

Total direct nonmedical costs 15.16 13.98 12.86 2.00 13.61Outpatient

Home care (family, friends, service) related to visit 45.66 51.66 16.52 10.48 38.31Absence from work related to GP visit 0.00 8.98 147.37 0.00 44.48

Inpatient†

Home care related to hospitalization 4.77 5.62 0.81 8.97 4.49Absence from work related to hospitalization 0.00 0.00 2.64 0.00 1.06

Total indirect costs 50.43 66.26 167.34 19.45 88.34Total cost allocated to one studied patient 142.59 179.24 220.59 80.05 168.15Of which direct 92.16 112.98 53.25 60.60 79.81Of which indirect 50.43 66.26 167.34 19.45 88.34

AGI, acute gastrointestinal infection; GP, general practitioner; WHO, World Health Organization.� US $1 ¼ 1.85 PLN (WHO CHOICE).† Inpatient costs weighted by age-specific proportion of GP patients referred to the hospital.‡ Based on the number of kilometers.


In our study, indirect costs constituted 53% of all costsattributed to AGI cases. The predominance of indirect costs overdirect costs generated by AGI cases was seen in all studiesperformed in developed countries. Indirect costs represented64% of total AGI-related costs in the United States [22], 70% inIsrael [6], and 87% in New Zealand [23].

In Poland, AGIs are perceived as low public health priority, whichis probably related to typical mild and self-resolving symptoms.Control measures are limited to the investigation of foodborneoutbreaks, and Salmonella control program in poultry flocks man-dated by the European Commission resolution since 2003 [27].Rotavirus vaccination is also not considered as priority by thenational advisory group for immunization. Currently, the cost ofetiological investigation for selected pathogens (most commonlySalmonella and Shigella) is free of charge only if performed in thecourse of an outbreak investigation. As a result of this situation,most AGI cases reported to surveillance have unknown etiology [15],which prevents implementation of efficient interventions.

Our study indicates that AGIs place a high burden on thesociety. Out-of-pocket spending for an average patient with AGIin Poland seems to be relatively small, amounting only to US $32for a medically attended case. Even coupled with indirect costsassociated with absence from work and home care equal to US$88, the burden to the individual patient may be regarded asrelatively low as equivalent to 7% of mean monthly income of aPolish citizen. Similarly calculated monetary burden on anaffected household was estimated to be 25% in Malaysia [28]and 40% in Taiwan [29]. If we considered the impact of AGIs on

Poland’s economy, the ratio of estimated AGI societal cost to GDPwould be 0.3% and the ratio to the total expenditure per capitawould be equal to 4.3%. Considering the estimate of 81% of thesocietal cost attributable to the public payer, AGIs clearly con-stitute a high burden on the national health care system. Asimilar comparison performed in New Zealand found that theratio of foodborne illness cost to GDP was 0.1% [23].

Our study has a number of limitations. First, restricting thestudy to health care facilities did not allow the estimation ofcosts of AGI episodes not consulting GPs. The follow-up inter-views provided some insight into self-medication practices andindirect costs of home care. These data are however not sufficientto extrapolate to the general population estimates of AGI casesthat were not medically attended. Second, noninclusion of third-level reference hospitals led to the underestimation of our costestimates because it included only the costs generated byuncomplicated AGI cases treated in general hospitals locatednear patients’ residence. Cases with severe complications,although rare, would lead to higher direct as well as indirectcosts of treatment. Third, we did not include private sector healthcare units in our study. In the private practices and privatehospitals, patients are paying for each service. By not includingprivate health care units, we probably underestimated the AGImanagement cost. However, the calculation of the actual costsincurred in the health care units would lead to the approximationof general health care costs, as the private sector units competefor the same patients with the public sector health care and haveto calculate their services accurately.

Table 4 – Average costs of medically attended AGI cases, by cost category and cost perspective, in Poland between May2008 and September 2009.

Cost category (US $)* Patient Payer Societal

OutpatientAverage cost of GP visit 0.00 10.25 10.25Laboratory testing 0.00 0.70 0.70Imaging diagnostics 0.00 0.49 0.49Pharmacotherapy 17.99 2.56 20.54Materials 0.00 0.06 0.06Emergency department visit 0.00 1.88 1.88Specialist consultation 0.00 0.32 0.32

Inpatient†

Average cost of hospitalization 0.00 28.11 28.11Laboratory testing 0.00 1.68 1.68Imaging diagnostics 0.00 0.54 0.54Pharmacotherapy 0.00 0.52 0.52Materials 0.00 1.02 1.02Specialist consultation 0.00 0.08 0.08

Total direct medical costs 17.99 48.21 66.19Transport related to GP visit 12.01 0.00 12.01Transport related to hospitalization† 1.60 0.00 1.60

Total direct nonmedical costs 13.61 0.00 13.61Outpatient

Home care (family, friends, service) related to visit 0.47 37.84 38.31Absence from work related to GP visit 0.00 44.48 44.48

Inpatient†

Home care related to hospitalization 0.00 4.49 4.49Absence from work related to hospitalization 0.00 1.06 1.06

Total indirect costs 0.47 87.87 88.34Total cost allocated to one studied patient 32.07 136.08 168.15Of which direct 31.60 48.21 79.81Of which indirect 0.47 87.87 88.34

AGI, acute gastrointestinal infection; GP, general practitioner; WHO, World Health Organization.� US $1 ¼ 1.85 PLN (WHO CHOICE).† Inpatient costs weighted by age-specific proportion of GP patients referred to the hospital.


Conclusions

These data may imply that much higher resources could beallotted to the prevention and control of acute gastroenteritis,even when considering only their direct medical costs. For bettertargeting of public health interventions, it would be necessary tounderstand the relative importance of different transmissionroutes as well as the relative burden placed by different patho-gens in Poland. An appropriate prioritization of interventions, inturn, would enable improvement in their cost-effectiveness.These results can also impact health care planning, as theprevention of nosocomial transmission of AGI pathogens maylead to substantial savings.

Source of financial support: The Polish Prospective HealthcareUtilization Survey was jointly funded by two European Commis-sion Sixth Framework Programme projects—POLYMOD (SSP22-CT-2004–502084) and the Network of Excellence MED-VET-NET(FOOD-CT-2004-506122)—and cofinanced by the Polish Ministry ofScience and Higher Education.

R E F E R E N C E S

[1] Nathavitharana KA, Booth IW. Pharmacoeconomics of the therapy ofdiarrhoeal disease. Pharmacoeconomics 1992;2:305–23.

[2] Szajewska H, Hoekstra JH, Sandhu B. Management of acutegastroenteritis in Europe and the impact of the new recommendations:a multicenter study. The Working Group on Acute Diarrhoea of theEuropean Society for Paediatric Gastroenterology, Hepatology, andNutrition. J Pediatr Gastroenterol Nutr 2000;30:522–7.

[3] Stefanoff P, Rogalska J, Czech M, et al. Antibacterial prescriptions foracute gastrointestinal infections: uncovering the iceberg. EpidemiolInfect 2012;15:1–9.

[4] Payment P. Epidemiology of endemic gastrointestinal and respiratorydiseases—incidence, fraction attributable to tap water and costs tosociety. Water Sci Technol 1997;35:7–10.

[5] Flores AR, Szilagyi PG, Auinger P, Fisher SG. Estimated burden ofrotavirus- associated diarrhea in ambulatory settings in the UnitedStates. Pediatrics 2010;125:e191–8.

[6] Stein M, Roisin H, Morag B, et al. The burden and cost of ambulatorycases of rotavirus gastroenteritis in Central Israel. Isr Med Assoc J2010;12:168–71.

[7] Osano BO, Wang’ombe JK, Rose W, et al. Cost analysis of care forchildren admitted to Kenyatta national hospital with rotavirusgastroenteritis. Vaccine 2011;29:4019–24.

[8] Riewpaiboon A, Intraprakan K, Phoungkatesunthorn S. Predictingtreatment cost for bacterial diarrhoea at a regional hospital inThailand. J Health Popul Nutr 2008;26:442–50.

[9] Lee BP, Azimi PH, Staat MA, et al. Nonmedical costs associated withrotavirus disease requiring hospitalization. Pediatr Infect Dis J2005;24:984–8.

[10] Mendelsohn AS, Asirvatham JR, Mkaya Mwamburi D, et al. Estimates ofthe economic burden of rotavirus-associated and all-cause diarrhoea inVellore, India. Trop Med Int Health 2008;13:934–42.

[11] World Health Organization. WHO country profile for Poland. Availablefrom: http://www.who.int/countries/pol/en/ [Accessed July 28, 2011].

[12] Czech M. Pharmacoeconomic evaluation and its role in management inhealth care in Poland. Hum Factors Ergon Manuf 2005;1:99–108.

[13] NHS reference costs 2010–11. Available from: http://data.gov.uk/dataset/nhs-reference-costs-2010-11 [Accessed June 1, 2013].

[14] Majowicz SE, Hall G, Scallan E, et al. A common, symptom-based casedefinition for gastroenteritis. Epidemiol Infect 2008;136:886–94.

[15] Baumann-Popczyk A, Sadkowska-Todys M, Rogalska J, Stefanoff P.Incidence of self-reported acute gastrointestinal infections in thecommunity in Poland—a population-based study. Epidemiol Infect2011;20:1–12.

[16] Drummond MF, Sculpher MJ, Torrence GW, et al. Methods for theEconomic Evaluation of Health Care Programmes (3rd ed). Oxford:Oxford University Press, 2005.

[17] Act on Accounting of 29 September 1994. J Laws 1994, No 121, item 591.[18] Medycyna Praktyczna. Pharmaceutical Index with Price List, issue

1/2009. Kraków: Medycyna Praktyczna, 2009.[19] Decree of the Minister of Infrastructure of 25 March 2002 on the

conditions for specifying the amount and method for refunding costsof the use of cars, motorbikes and mopeds not owned by the employerfor official purposes. Offi J Laws Oct 30, 2007, No 201, item 1462.

[20] World Health Organization. Tables of costs and prices used in WHO-CHOICE analysis. Available from: http://www.who.int/choice/costs/en/[Accessed November 13, 2011].

[21] StataCorp. Stata Statistical Software: Release 10. College Station, TX:StataCorp LP, 2007.

[22] Garthright WE, Archer DL, Kvenberg JE. Estimates of incidence andcosts of intestinal infectious diseases. Public Health Rep1988;103:107–16.

[23] Scott WG, Scott HM, Lake RJ, Baker MG. Economic cost to New Zealandof foodborne infectious disease. N Z Med J 2000;113:281–4.

[24] Fischer TK, Anh DD, Antil L, et al. Health care costs of diarrhoealdisease and estimates of the cost-effectiveness of rotavirus vaccinationin Vietnam. J Infect Dis 2005;192:1720–6.

[25] Isakbaeva ET, Musabaev E, Antil L, et al. Rotavirus disease inUzbekistan: cost-effectiveness of a new vaccine. Vaccine2007;25:373–80.

[26] Flem ET, Latipov R, Nurmatov ZS, et al. Costs of diarrheal disease andthe cost-effectiveness of a rotavirus vaccination program inKyrgyzstan. J Infect Dis 2009;200(Suppl.):S195–202.

[27] Regulation no 2160/2003 of the European Parliament and of the Councilof 17 November 2003 on the control of salmonella and other specifiedfood-borne zoonotic agents. Off J Eur Union December 12, 2003, L 325/1.

[28] Chai PF, Lee WS. Out-of-pocket costs associated with rotavirusgastroenteritis requiring hospitalization in Malaysia. Vaccine 2009;27(Suppl. 5):F112–5.

[29] Kow-Tong C, Shiang-Fang F, Ren-Bin T, et al. Hospital-based study ofthe economic burden associated with rotavirus diarrhea in Taiwan.Vaccine 2007;25:4266–72.


































http://www.who.int/countries/pol/en/



http://data.gov.uk/dataset/nhs-reference-costs-2010-11

http://data.gov.uk/dataset/nhs-reference-costs-2010-11


















http://www.who.int/choice/costs/en/




























Radiology Services Costs and Utilization Patterns Estimates inSoutheastern Europe—A Retrospective Analysis from SerbiaMihajlo Jakovljevic, MD, PhD1,�, Ana Rankovic, MD2, Nemanja Rančic, MD3, Mirjana Jovanovic, MD, PhD4, Miloš Ivanovic, PhD5,Olgica Gajovic, MD, PhD6, Zorica Lazic, MD, PhD7

1Faculty of Medical Sciences, Department of Pharmacology and Toxicology, University of Kragujevac, Kragujevac, Serbia; 2Radiology Diagnostic Service,University Clinical Center, Kragujevac, Serbia; 3Medical Faculty, Centre for Clinical Pharmacology, Military Medical Academy, University of Defence, Begrade,Serbia; 4The Psychiatry Clinic, University Clinical Center, Kragujevac, Serbia; 5The Faculty of Science, University of Kragujevac, Kragujevac, Serbia; 6TheInfectious Diseases Clinic, University Clinical Center, Kragujevac, Serbia; 7The Pulmonary Diseases Clinic, University Clinical Center, Kragujevac, Serbia

A B S T R A C T

Objective: Assessment of costs matrix and patterns of prescribing ofradiology diagnostic, radiation therapy, nuclear medicine, and inter-ventional radiology services. Another aim of the study was insightinto drivers of inappropriate resource allocation. Methods: An in-depth, retrospective bottom-up trend analysis of services consump-tion patterns and expenses was conducted from the perspective ofthird-party payer, for 205,576 inpatients of a large tertiarycare university hospital in Serbia (1,293 beds) from 2007 to 2010.Results: A total of 20,117 patients in 2007, 17,436 in 2008, 19,996 in2009, and 17,579 in 2010 were radiologically examined, who consumedservices valued at €2,713,573.99 in 2007, €4,529,387.36 in 2008,€5,388,585.15 in –2009, and €5,556,341.35 in 2010. Conclusions: Themacroeconomic crisis worldwide and consecutive health policy meas-ures caused a drop in health care services diversity offered insome areas in the period 2008 to 2009. In spite of this, in total it

increased during the time span observed. The total cost of servicesincreased because of a rise in overall consumption and populationmorbidity. An average radiologically examined patient got one frontalchest graph, each 7th patient got an abdomen ultrasound examina-tion, each 19th patient got a computed tomography endocraniumcheck, and each 25th patient got a head nuclear magnetic resonance.Findings confirm irrational prescribing of diagnostic procedures andnecessities of cutting costs. The consumption patterns noticed shouldprovide an important momentum for policymakers to interveneand ensure higher adherence to guidelines by clinicians.Keywords: costs, interventional radiology, nuclear medicine, radiationtherapy, radiology diagnostics, utilization patterns.


Introduction

Our time witnesses unseen contemporary advances in medicaltechnology and development of modern equipment in allbranches of medicine. A great number of new diagnostic andtherapeutic methods have come up in radiology as well. They arevery powerful, but their purchase price (which often amounts to afew million euros, for, for example, computed tomography [CT]and nuclear magnetic resonance [NMR]) limits broader usage andreplacement of existing appliances. Apart from diagnostic appli-ances, there has been a technological revolution in radiologicalmethods of intervention radiology and radiotherapy, the develop-ment and application of which have prospered in the last 10 years.Parallel to the invention of such appliances and their procure-ment, a problem has been noted because their services are veryexpensive and keep a constant burden on health funds [1–8].These appliances represent mass consumers of health carebudgets worldwide. This is particularly the case when considering

secondary and tertiary care hospitals. Health economic estimatesof radiation-mediated diagnostic and treatment procedures areseldom reported in the literature. Of those that are available, mostdeal with imaging diagnostics or radiotherapy in oncology on aseparate basis. This would be the first study to compare all theseexaminations and interventions in a large-scale trial.

In Serbia, as a typical upper-middle income southeasternEuropean country, expensive high-tech services were centralizedto several tertiary facilities, the third largest of them being theClinical Center in Kragujevac. For this reason, we chose thisparticular university clinic with approximately 1,300 beds, morethan 50,000 hospital admissions, and 400,000 outpatient exami-nations per year. Another problem lies in the fact that theseexpensive high-tech services are nonrationally prescribed, whichalso contributes to excessive consumption and spending fromthe modest health budget. By analyzing the 3-year-long trend inthe consumption of services (the volume, i.e., the number of, thefrequency, or expenses), we are of the opinion that the key






* Address correspondence to: Mihajlo Jakovljevic, Faculty of Medical Sciences, Department of Pharmacology and Toxicology, Universityof Kragujevac, Svetozara Markovica 69, 34 000 Kragujevac, Serbia.


dx.doi.org/10.1016/j.vhri.2013.07.002

dx.doi.org/10.1016/j.vhri.2013.07.002

dx.doi.org/10.1016/j.vhri.2013.07.002


dx.doi.org/10.1016/j.vhri.2013.07.002

weakness of clinical practice is the lack of consistency in follow-ing guidelines for good clinical practice.

Based on these above-mentioned facts, interventions couldbe made in the future aimed toward more rational prescription,especially diagnostic methods, which would redistribute thelimited resources to the cases in which these are necessary forproper treatment. Similar to our postulates, numerous studieshave shown an increasing trend of the unrealistic use/con-sumption of radiologic diagnostic methods [9–13], and it isspecially the case with new radiologic methods. The downsideof these studies, and thus of ours too, is the short time duringwhich the study was carried out (3–6 years), which will beannulated by the future prolonging of the study time of thisstudy. We consider it necessary to establish a special organ-ization whose purpose would be to monitor the prescription ofradiologic measures and mark the cumulative annual accepteddosage of radiation given to patients, as was the case withpatients who took medicines in some countries, the CzechRepublic for instance [14].

The aim of this article was to establish whether radiologicmethods in diagnostics and therapy have been used rationallyduring the last 4 years and to substantiate the need for making aguideline for the application of radiologic methods in clinicalpractice.

Methods

The 4-year-long retrospective analysis of total expenditure trendsof radiologic services from the spheres of the classical radio-graphic, high-tech imaging diagnostics, interventional radiology,radiation therapy, and procedures of nuclear medicine during theyears 2007, 2008, 2009, and 2010 was provided by the database ofthe Clinical Center in Kragujevac (Figs. 1 and 2). The regularinvoicing of services provided in the daycare service and tohospitalized patients according to International Statistical Classi-fication of Diseases, 10th Revision code of illnesses and the name/surname/personal identification number resulted in a largeadministrative database, which is regularly updated. By thecooperation of the clinics and departments in charge, the previewof the database was obtained.

Authors analyzed services that were used the most frequentlyduring the above-mentioned years (top 10 of the expenditurevolume) and the most expensive services (top 10 of the totalvalue of services), that is, those that by themselves take up 67% to

95% (arithmetic mean � 1 or 2 SD) of the total value of services.The population included in this research amounts to 600,000inhabitants and is situated in central Serbia, in Sumadija, and theclinical center in charge is the one in Kragujevac.

Results

More than 17,000 radiologically examined inpatients per yearhave been noted in the clinical center in Kragujevac (Table 1).During 2007, most of the services were provided to outpatients;during 2008 and 2009, most of the services were provided toinpatients; and in 2010, all the radiological diagnostic andtherapy services in nuclear medicine and interventional radiol-ogy were provided to inpatients exclusively. Total expensesconstantly increased during the 4-year period analyzed (Table 1).

Nine percent of total 4-year expenses belong to nuclearmedicine, 16% to radiotherapy, while the radiodiagnostic serviceincluding interventional radiology spends 75% of the budgetintended for radiological services. The number of inpatientsconstantly increased from 2007 to 2010 (Table 1), but the ratioof patients who received one of the radiological services wasrelatively reduced with periodic oscillations (44.04% in 2007,34.55% in 2008, 35.55% in 2009, and 31.39% in 2010). In 2007, onaverage every second patient hospitalized received one of theradiological services, while in 2010, every third patient receivedone of the radiological services.

The average price of radiological services per patient constantlyincreased during this 4-year period analyzed—in 2007, 10,658.97RSD (€134.89); in 2008, 20,516.81 RSD (€259.77); in 2009, 26,506.07RSD (€283.67); and in 2010, 33,045.99 RSD (€316.08) (Fig. 3).

The total number of hospital admissions with radiologicallyexamined patients slightly reduced from 2007 to 2010 (Table 1).Eighty-three percent of all the first hospitalizations at the Depart-ment of Radiology and Nuclear Medicine belong to radiodiagnos-tics with interventional radiology.

The number of patients examined and nuclear medicineservices provided in this period decreased (Table 1). Costs,however, increased significantly although the number of patientsexamined or services provided reduced, from 32,272,107.84 RSD(€408,404.30) to 50,264,302.32 RSD (€481,597.20) in 2010, with aslight decline in 2008 (Table 1).

The number of patients who received some of the radiodiag-nostic services remained at about 15,000 during the 4-year period(range of 17,000–19,000). The expenses of these services, however,constantly increased, tripling from 2007 to 2010 (Table 1).

The number of patients who received some of the radio-therapy services, and the number of services provided as well,had a slight increase (Table 1). This slight rise in the obtainedservices volume follows the increase in cost (Table 1).

The services of nuclear medicine lowered their expenses from15% in 2007 to 8.6% in 2010 per year. A similar fall was noticed inradiotherapy, from 23% in 2007 to 12% in 2010. However, radio-diagnostic services mark a constant increase in the percentageshare from 62% in 2007 to 79.55% in 2010.

The total number of services given, including the repetitiveone, constantly increased (Table 1), With 81% of the servicesprovided belonging to the field of radiotherapy, 18% belonging toradiology diagnostic services with interventional radiology, andonly 1% belonging to nuclear medicine. The total number ofnuclear medicine services provided was doubled in 2010 com-pared with that in 2007 (Table 1). In contrast to nuclear medicine,the total number of radiology diagnostic services and radiother-apeutic services provided constantly grew (Table 1). The trend ofincrease in the total number of radiology diagnostic servicesprovided from 2007 to 2010 was 30 times bigger (Table 1).Fig. 1 – Division of core services.


The diversity of services offered suffered from a short recessionaldrop in some areas but in total all the three groups (nuclearmedicine, radiology diagnostic, and radiotherapy) recorded anincrease of two to three times from 2007 to 2010 (Table 1).

A total of 3% of the patients received some of the radio-therapeutic services, 16% received nuclear medicine services,while 81% of the patients received some other radiology diag-nostic or emergency radiology services.

The upward trend in diversity is the largest in the field ofinterventional radiology and classical radiographics, while thetrend shows a slight increase in nuclear medicine and radiationtherapy.

The services that are most used in nuclear medicine (Table 2)are those that determine the thyroid gland hormones andthyroid-stimulating hormone. We can notice that the numberof given services reduced by two times from 2007 to 2010. Most of

0

500,000

1,000,000

1,500,000

2,000,000

2,500,000

3,000,000

3,500,000

4,000,000

4,500,000

2007 2008 2009 2010

Nuclear Medicine

Radia�on Therapy

Radiology Diagnos�c Service*(including interven�onal)

Total financial value of spending (€)

Fig. 2 – Expenditure comparison between three clinical branches.

Table 1 – Display of all radiological services analyzed and respective costs incurred.

Number of patients treated 2007 2008 2009 2010

Total number of inpatients admitted (regardless of radiologicalexamination presence/absence)

45,677 50,459 53,433 56,007

Nuclear medicine 4,456 2,041 2,566 2,990Radiology diagnostic service (including interventional) 15,224 14,918 15,903 14,050Radiation therapy 437 477 527 539Total 20,117 17,436 19,996 17,579

Number of single hospital admissions 2007 2008 2009 2010Nuclear medicine 5,193 2,114 2,715 3,267Radiology diagnostic service (including interventional) 18,320 18,227 19,231 17,094Radiation therapy 457 517 563 565Total 23,970 20,858 22,509 20,926

Total frequency of services provided (including repeated procedures) 2007 2008 2009 2010Nuclear medicine 18,145 10,670 18,684 32,002Radiology diagnostic service (including interventional) 53,416 183,540 928,777 1,591,285Radiation therapy 2,975,493 3,104,262 3,196,789 3,468,228Total 3,047,054 3,298,472 4,144,250 5,091,515

Diversity of services offered (number of different services available) 2007 2008 2009 2010Nuclear medicine 73 159 128 169Radiology diagnostic service (including interventional) 216 488 596 658Radiation therapy 277 266 275 348Total 566 913 999 1,175

Total financial value of services consumed (€) 2007 2008 2009 2010Nuclear medicine 408,404.3 339,539.4 432,617.4 481,597.2Radiology diagnostic service (including interventional) 1,691,606.7 3,352,212.6 4,253,573.2 4,424,357.8Radiation therapy 613,562.9 837,489.5 702,228.3 659,968.8Total 2,713,573.99 4,529,387.36 5,388,585.15 5,556,341.35


the expenses in nuclear medicine belong to services related todetermining the thyroid gland hormones (Table 2).

The greatest expenses of radiotherapy (Table 2) during theanalyzed 4-year period were for gentigrey. According to theexpenditure volume, gentigrey is the leader of radiotherapeuticmethods (Table 2).

According to the volume of services, nuclear medicine sig-nificantly decreased the prescription of its services from 2007to 2010.

The most common radiotherapy service, gentigrey, shows theapplication plateau at about 60 units per patient hospitalizedduring the analyzed period, while the accelerator isocentertechnique reduced the volume of consumption, since every 8thpatient received that service, and in 2010, every 12th patientreceived it. Team treatment in 2007 of the patient for radiationtherapy was significantly increased in volume. In 2007, every501st patient received this service while in 2010 every 133rdpatient received it.

Table 2 shows that the lung RO in maximum expiration is themost common method applied among the classical radiologydiagnostic (RO) methods. Its use has a clear trend of increase—anincrease of 3.5 times from 2007 to 2010. It means that almostevery patient who received some of the classical diagnosticmethods of radiology got at least one RO graphy of lungs inmaximum expiration, and out of all hospitalized patients every4th patient got the lung RO graphy in maximum expiration in2010, while in 2007 every 13th patient received it. The use of ROgraphy of the abdomen in the posterior-anterior (PA) position andprofile in the same period increased about 7 times.

Similarly, the consumption of classical methods of radiologyas the most widespread methods is the most expensive, and thetrend of consumption is evident, from 726,269.49 RSD (€9,190.96)

in 2007 to 3,278,406.48 RSD (€31,411.27) in 2010 (a jump of 4.5times in consumption) (Table 2).

Imaging methods are great consumers, and the greatestconsumer is the CT-targeted imaging of particular organs accom-panied with reconstructions, and the expenses of this methodincreased 23 times during the 4-year period. Right afterwardscome the head and neck CT without contrast agents for whichthere was an increase of 1.3 times.

In Table 2, we notice that imaging methods are prescribed veryoften. Targeted CT of particular organs followed by reconstruction isthe most common method, and its frequency of usage represents aclear trend of growth of 18.5 times during the period analyzed. Itmeans that every 3rd patient hospitalized received this service in2010, while in 2007, it was every 52nd patient. It is interesting tonote that the use of the standard abdominal ultrasound examina-tion (examination of liver, gall bladder, pancreas, spleen, andkidney) decreased 1.8 times. The greatest leap (50 times) is noticedfor the head and neck CT without contrast agents.

Methods of interventional radiology are applied to a lesserextent, and the most common method is invasive hemodynam-ics, followed by selective coronary angiography and cardiaccatheterization. Throughout the analyzed period, each patientreceived the service of invasive hemodynamics. Percutaneoustransluminal angioplasty (PTA) revascularization (without stentimplantation) and endovascular treatment of intracranial aneur-ism are the largest consumers, and costs have a growing trend ofabout three times during the period analyzed (Table 2).

The top 10 most expensive disorders to treat were recognizedby analyzing resource use and costs related to the patient'sdiagnosis at hospital discharge. These are provided jointly forradiology diagnostic examinations, nuclear medicine procedures,and radiation therapy in oncology in Table 3.

Fig. 3 – Average cost per patient treated and average cost per admission.


Discussion

In 2007, the largest portion of services was obtained out-patiently; in 2008 and 2009, services were equally obtainedoutpatiently and inpatiently; and in 2010, radiological serviceswere mostly obtained inpatiently because it was noticed that

more strict prescription is important as a way to achieveserious cost-containment, which is already less in comparisonto that in some other countries [15]. By that measure, thenumber of radiological services provided decreased in 2010,but because the price of these services multiplied, and somemore expensive methods started to be used, treatmentexpenses tripled compared with those in 2007. We can notice

Table 2 – Top 10 services according to the volume of unit consumption and according to treatment expenditures.

Radiodiagnostics

Unit consumption Value of consumption (€)

Chest graph in the maximum expiration AP 9,598,728 Target CT images with reconstruction ofsome organs

1,650,655

RO graphy of the abdomen in the PA position and profile 399,191 PTA revascularization (without stentimplantation)

1,531,675

The esophagus, stomach, and duodenum— double volumecontrol methods targeted shooting

351,846 Endovascular treatment of intracranialaneurysm

1,414,823

RO graphy of knee joint, lower leg, ankle, or foot in twodirections

380,503 Coronarography—Catheterization 885,123.9

RO graphy pelvis—AP position 281,978 CT head and neck without contrast media 524,185.6Irrigography double contrast medium 196,383 Selective coronary angiography 312,544.9RO L and LC spine in two directions 184,345 Invasive hemodynamics 291,353.5RO graphy PA skull in profile position 183,697 CT of abdominal organs with contrast

medium21,2791

Lung RO graphy in the D or L decubitus 156,841 CT of abdominal organs without contrastmedia

19,0800

Skull RO graphy in children 71,135 Percutaneous balloon angioplasty ofcoronary artery catheter

56,910.3

Nuclear medicine


TSH—tireotrophic homone RIA 11,005 TSH—tireotrophic hormone RIA 170,801.1Free T4 (FT4)—RIA 10,984 Free T4 (FT4)—RIA 144,161.3Free T3 (FT3)—RIA 3,705 Free T3 (FT3)—RIA 110,124.7Mikrosomic antibody (anti-TMS) (IRMA) 3,256 Microsomic antibodies (anti-TMS) (IRMA) 108,182.5Serum prolactin (RIA) LTH 3,022 Tireoglobulin (RIA) 63,723.5Cortisol determination (RIA)—method of incubation and

separation2,996 Serum prolactin (RIA) LTH 59,135.47

Determination of insulin RIA 2,908 Insulin determination RIA 51,449.5Tireoglobulin (RIA) 1,794 Scintigraphy of the whole body J-131 47,230.93Titer thyroglobulin antibody 1,652 Determination of cortisol (RIA)—method of

incubation and separation40,631.77

Scintigraphy of the whole body J-131 1,193 Titer tireoglobulin antibody 26,918.96

Radiation therapy


Gentigrey (in units) 11,581,975 Gentigrey (in radiation absorbtion units) 1,439,235Accelerator—isocentric technique 19,732 Determination of airfields spelling 26,566.63Supervoltage accelerator radiotherapy with the modified field 12,224 Intracavitary gynecological applications 21,739.46Supervoltage accelerator radiotherapy with wedge-shaped

filter8,028 Team treatment for aerial treatment of the

patient19,564.45

Determination of the airfield graphs 6,320 Determination of markers spelling 5,199.80Determination of markers spelling 1,334 Accelerator—an isocentric technique 735.70Team treatment for aerial treatment of the patient 1,250 Supervoltage accelerator radiotherapy

with the modified field463.62

Radiotherapy—accelerator leaning 579 Supervoltage accelerator radiotherapywith wedge-shaped filter

436.10

After loading the applicator with the applications withsource intensity catheter

20 Radiographic verification using selectron 203.54

Intracavitary gynecological applications 20 Team treatment—selectron 142.78

AP, anterior-posterior; IRMA, Immunoradiometric assay; LTH, luteotropic hormone; PA, posterior-anterior; PTA, percutaneous transluminalangioplasty; RIA, Radioimmunoassay; RO, radiology diagnostic; TMS, thyroid microsomal antibodies.


the disproportion between services volume and treatmentcosts.

Radiology diagnostics and interventional radiology are the larg-est consumers of these three groups of services. This was expectedbecause these services are often prescribed, and the price of theinterventional radiology service was high, and so 75% of the budgetof radiology and nuclear medicine was spent for this purpose.

Because the number of hospitalized patients constantlyincreased during the analyzed period, the expectations of costexpenditure are justified. Although more stringent prescription ofthese services is applied, so that the percentage share of patientswho receive radiological and nuclear services is decreased sig-nificantly, the leap in the expense of these services led to anincrease in total costs.

The average service price increased about three times perpatient and that is a great deal of the budget obtained for thesepurposes. Although recession significantly affected the morestrict prescription, the services price leap led to the growth inaverage service price per patient.

The number of patients hospitalized for the first time whoreceived some service was significantly decreased, and thenumber of patients processed from all three groups decreasedas well. In contrast to reducing the volume of unit consumption,the costs increased many times. Out of all services, the radiologydiagnostic service with interventional radiology recorded a con-stant increase in percentage cost share, from 62% to 79.55%. Theincrease in expenses followed because of the increase in theservice volume of interventional radiology methods, which arevery expensive and their prices are constantly growing on theworld market. The increase in the number of various servicesoffered led to an increase in cost, because the constant promo-tion of institutions and the education of physicians widen the

service assortment available, and therefore covers the greaterrange of diseases that can be treated.

Because the thyroid gland diseases are increasing in Serbia, itis expected that the volume and the price of services used for thethyroid gland examination will increase. This is most evidentwhen we look at nuclear medicine services among the top 10services, and their volume and consumption—six to seven of thetop 10 services are connected with the examination of the thyroidgland. It is believed that the main reason for the poor economicstatus of the population (upper-middle gross national income percapita), and therefore, poor diet, increasing stress, and environ-mental damage caused by the NATO bombing campaign of 1999.In the last decade, a fivefold increase was marked in thyroidgland and other endocrine gland diseases.

Malignant diseases, which require radiotherapy, are rising aswell. Thus, although the percentage share of expenditure ofradiotherapy services in relation to the entire annual consump-tion decreased, customer value of these services increased byone third.

The most number of services was provided in the radio-therapy domain. That is to be expected when every session inthe treatment of some malignant disease means giving a largernumber of cytostatics several times during the treatment.

Almost every patient who received some of the diagnosticmethods of classical radiology got at least one lung RO graphy inmaximum expiration, and out of all hospitalized patients every4th patient got the lung RO graphy in maximum expiration in2010, while in 2007, every 13th patient got the lung RO graphy.This meant that the number of hospitalized patients increased,but the lung RO graphy in maximum expiration was prescribedeven more often in 2010—although it was expected because of itsstringent regulation—rarely an indication of recession. Therefore,this method is the biggest consumer, and marks a jump of 4.5times in the period analyzed.

The important data are that CT-targeted recordings of partic-ular organs with reconstruction are the greatest consumers ofimaging methods and that their consumption increased 23 times,which is affirmed by the data that the volume trend of thismethod marked a clear growth of 18.5 times. The findings in theliterature were also similar [7,16].

The methods of interventional radiology did not record anyincrease in the consumption volume, and the growth in theprices of the services led to a consumption increase of aboutthree times. Although the opening of the catheterization hall inKragujevac should have led to a significant influx of patients inthis field and increase the service volume, it had not happened.

Because each 40th patient in our center received some kind ofa radiology service, it is considered as the nonrational use ofthese diagnostic methods. The tendency of the growth in con-sumption from 2007 till 2008, which dropped in 2009, can beexplained by the means of the global economic crisis, and shouldnot be ascribed to the rationally prescribed radiologic procedures.This is a devastating piece of information and must be taken witha great caution and a guide of good practice should be made inthe future concerning the radiologic procedures and in that waytry to examine the prescription of the treatment and sanctionthose who do not follow this guide. All this should reduce theunnecessary waste of budget assets.

When compared with the Levin [17] study in Pennsylvania,PA, of about 6 million health insurance holders, significantlyhigher expenses were incurred in total on radiologic check-ups,out of which 62% were nonradiologic expenses. It is a relativelysmall consumption for these services, but it still represents anenormous expense for our modest budget assets and it preventsthe allocation of necessary financial assets for other purposes.Sunshine et al. [18] reported that during 1996 and 1997, largefinancial resources were spent during the full-time workload at

Table 3 – Top 10 most expensive diagnoses ofradiologically examined and/or treated patients2007–2010 (€).

ICDcode

Diagnosis Valueconsumed

Z95 Persons with potential health hazardsrelated to family and personalhistory and certain conditionsinfluencing health status—Presenceof cardiac and vascular implants andgrafts

1,365,643.22

C50 Malignant neoplasm of breast 973,285.91C73 Malignant neoplasms of thyroid and

other endocrine glands— Malignantneoplasm of thyroid gland

920,159.01

I21 Ischemic heart diseases—Acutemyocardial infarction

664,869.96

I20 Ischemic heart diseases—Anginapectoris

653,568.84

C01 Malignant neoplasms of lip, oral cavity,and pharynx—Malignant neoplasmof base of tongue

460,959.67

I60 Cerebrovascular diseases—Subarachnoid hemorrhage

442,914.85

I67 Cerebrovascular diseases—Othercerebrovascular diseases

424,082.16

I10 Hypertensive diseases—Essential(primary) hypertension

410,427.79

G81 Cerebral palsy and other paralyticsyndromes—Hemiplegia

373,981.90

ICD, International Statistical Classification of Diseases.


radiology for about 320 radiologic procedures in the United States(including the methods of radiologic oncology).

A great importance is given today to the methods of interven-tional radiology, which replace the classical surgery and greatlyshorten the postoperational recuperation of patients who arehospitalized and in this way reduce the expenses as well. Theseshould be usedmore andmore, and the fall in their use in our studyduring 2009 was, although a partial explanation can be found in thereduction of the health budget, due to the financial hardship.

Average expenses for radiologically examined patientsin all the three analyzed years were in the range from 10,000RSD (€125) to 17,000 RSD (€218), and if this is multiplied withthe size of the population of a million residents, it is a greatexpense.

The devastating fact is that irrespective of the dischargediagnosis each patient gets a lung graph or every 10th gets a CTof the endocranium on average. This is a dubious piece ofinformation when it is compared with other studies, as the oneby Bhargavan and Sunshine [19], in which out of the total 4176radiologic procedures, 49.3% go to all the radiologic check-ups and9.36% to all the CTs. When these two populations are compared,this is a catastrophic information for our health care system andspeaks in favor of the human neglect and the great need for aguide of good clinical practice with strict indications [7].

It is interesting to note that among the 10 most expensivemethods, RO of the lung graph ranked second, the reason for thisbeing its nonrational consumption, which must be stoppedbecause of not only huge expenses but also the unnecessaryradiation to which patients are exposed.

Many studies have shown that there has been a reduction inthe total number of radiologic check-ups. The number of offeredservices using new methods (CT and NMR) has increased, whichmeans that they are often misused and that patients areunnecessarily undergoing shots and radiation [20]. As it is awell-known fact that the dosage during the CT treatment is 300times higher than that during ordinary CT scan graph, here liesthe danger of the possible development of tumor, which origi-nates from mutations caused by such radiation.

When compared with the American study from 2001, in whichout of the 4176 radiologic services there were 215 services ofinterventional radiology [19], the percentage of the interventionradiology service is modest in our case, but at the same time theuse of other conventional methods is very high.

The age distribution, as well as the sex distribution, onlyconfirms that today the presence of the disease is the same forboth sexes and that the older population is much more exposedto the radiologic radiation. The devastating piece of informationis that one tenth of the patients who undergo the radiologiccheck-up are younger than 18 years. With this population oneshould be stricter when prescribing the radiology diagnosticimaging (RDI) methods because of the impact of the radiationon gonad cells.

Our study once again confirmed that the most frequent arestill the cardiovascular diseases and that their prevention is themost important. A special attention should be paid to this pieceof information, and the guide for the prevention of such disordersshould be used.

Because CT ranked eighth among the top 10 services accordingto the unit consumption, it is a worrying information because it isnot possible that all these patients have fulfilled the conditions forcertain indications in which these very powerful but at the sametime very harmful radiologic methods are prescribed. It is also veryimportant that the RDI of lungs is still the most often used of allradiologic methods and that it is unbelievable that each patient getsalong a graph on average. It is an irresponsible prescription of thisRTG method, and strict instructions should be given to stop suchnonrational prescription.

Top 10 lists according to the total consumption have beenshown, which claim that the lungs RDI is ranked second and thatthe remaining ranks are filled with CT methods and two inter-vention methods. Huge resource consumers are also the adjoin-ing expenses of these radiologic procedures: RO films, contrasts,and radioactive isotopes.

A special attention should be paid to lists of most expensivedisorders per clinical radiology branch-department. Nuclear med-icine top 10 list of most expensive diagnosis reveals an antici-pated structure, with endocrine disorders dominating. Oncologyradiation therapy services consumption exhibits unique morbid-ity rates of particular cancer types in the region. A very interest-ing finding in the domain of radiology diagnostics is forcardiovascular and orthopedic surgical implantation proceduresand dealing with their complications, being placed at the very topof the list. Other details can be studied in Table 3.

The trend of a constant rise in expenses as well as the volumeof services in radiology was marked with many services. We havechosen only some of them and shown them on the graph, inwhich a trend of growth can clearly be seen, with a slightdecrease during 2009 due to the reduction in the budget as aresult of the global economic crisis [21].

The data concerning the number of services provided, whichrises from year to year and is tripled every year, is a very goodindicator of how nonrationally the methods of radiologic diag-nostics and therapy are used. At the same time, however, adecrease in the application of the methods of interventionalradiology is discomforting, because the recently opened theatreof interventional radiology should replace the big expenses of thesurgical blocks and contribute to the quick diagnostics andtreatment of these patients [22].

Conclusions

The total level of the value of the consumption of radiologic servicesis comparable (higher/lower) with other tertiary clinics of similarsize among developing countries. However, the structure of theconsumption within the available hospital budget is not the bestand could be significantly improved by redirecting the nonrationallyspent resources toward priority domains. The suggested measurefor the improvement of the structure of the suggested diagnosticand therapeutic intervention is the making of a local protocol ofgood practice by which a desirable and suggested dynamics of thefrequency of doing certain check-ups in certain indications (pneu-monia, stroke, tumors, etc.) is to be established. The investment ofresources for the implementation and application of such a guideand a periodical (annual) estimation of following the protocol,which was agreed upon, would be a good idea. Consumptionpatterns should be transformed in respect to their quality andupgraded, so that the decision is more significantly based onevidence. The recession, macroeconomic events, and current cutsin the budget for the health sector in Serbia will increase theexternal pressure on management bodies and so such changes willprobably be inevitable in the coming years.

Acknowledgments

We express our gratitude to the Ministry of Science Educationand Technological Development of the Republic of Serbia forgrant number 175014, out of which this research project waspartially financed.

Source of financial support: This study was partially financedby the Ministry of Science Education and Technological Develop-ment of the Republic of Serbia (grant no. 175014).


R E F E R E N C E S

[1] Goedert J. The hidden costs of radiology. Health Data Manag2007;15:24–6.

[2] Clevert DA, Stickel M, Jung EM, et al. Cost analysis in interventionalradiology–a tool to optimize management costs. Eur J Radiol2007;61:144–9.

[3] Golder W. Benefits, costs and analysis in diagnostic radiology:definitions and glossary. Rofo 1999;170:73–9.

[4] Kelekis DA, Brountzos EN. How much does radiology contribute toincreasing health care costs? The situation in Greece. Acad Radiol1996;3(Suppl. 1):S121–4.

[5] Gerhardt P. Diagnostic radiology: health care costs and need for therapyrelevant examination strategies. Röntgenpraxis 1994;47:129–38.

[6] Krotz D. Aetna outsources radiology services utilization review. DiagnImaging (San Franc) 1998;20:19.

[7] Khorasani R, Goel PK, Ma’luf NM, et al. Trends in the use of radiologywith inpatients: what has changed in a decade? AJR Am J Roentgenol1998;170:859–61.

[8] Otero HJ, Ondategui-Parra S, Nathanson EM, et al. Utilizationmanagement in radiology: basic concepts and applications. J Am CollRadiol 2006;3:351–7.

[9] Levin DC, Bree RL, Rao VM, Johnson J. A prior authorization program ofa radiology benefits management company and how it has affectedutilization of advanced diagnostic imaging. J Am Coll Radiol 2010;7(1):33–8: (Erratum in: J Am Coll Radiol 2010;7:235).

[10] Maitino AJ, Levin DC, Parker L, et al. Nationwide trends in rates ofutilization of noninvasive diagnostic imaging among the Medicarepopulation between 1993 and 1999. Radiology 2003;227:113–7.

[11] Sunshine JH, Bushee GR, Mallick R. U.S. radiologists’ workload in 1995–1996 and trends since 1991–1992. Radiology 1998;208:19–24.

[12] Zylak CJ, Gafni A. A methodologic overview of the evaluation of costsand benefits in diagnostic radiology. Invest Radiol 1992;27:483–8.

[13] Heilman RS. Costs, benefits, and common sense in radiology.Radiographics 1998;18:849–50.

[14] Biba V, Kubica Z, Strnadova J, et al. Drug consumption statistics in theCzech Republic. Vestnik 1997;(Suppl.):1–68.

[15] Mecozzi B, Pancione L, De Intinis G, et al. Analysis of productionfactors, costs, and process efficacy in the radiology department of alocal health agency in Italy. Radiol Med 2003;105:215–29.

[16] Nisenbaum HL, Birnbaum BA, Myers MM, et al. The costs of CTprocedures in an academic radiology department determined by anactivity-based costing (ABC) method. J Comput Assist Tomogr2000;24:813–23.

[17] Levin DC. The practice of radiology by nonradiologists: cost, quality,and utilization issues: Merrill C. Sosman Lecture. AJR Am J Roentgenol1994;162:513–8.

[18] Sunshine JH, Burkhardt JH, Mabry MR. Practice costs in diagnosticradiology. Radiology 2001;218:854–65.

[19] Bhargavan M, Sunshine JH. Utilization of radiology services in theUnited States: levels and trends in modalities, regions, andpopulations. Radiology 2005;234:824–32.

[20] Matin A, Bates DW, Sussman A, et al. Inapatient radiologyutilization: trends over the past decade. AJR Am J Roentgenol2006;186:7–11.

[21] Jakovljevic MB. Resource allocation strategies in SoutheasternEuropean health policy. Eur J Health Econ 2013;14:153–9.

[22] Rankovic A, Rancic N, Jovanovic M, et al. Impact of imaging diagnosticson the budget – are we spending too much? Vojnosanit Pregl2013;70:709–11.






























































Children Hospitalized for Varicella: Complications and Cost BurdenOzden Turel, MD1,�, Mustafa Bakir, MD2, Ismail Gonen, MD1,3, Nevin Hatipoglu, MD3, Cigdem Aydogmus, MD3,Emine Hosaf, MSc4, Rengin Siraneci, MD3

1Department of Pediatrics, Bezmialem Vakif University Faculty of Medicine, Istanbul, Turkey; 2Department of Pediatrics, Marmara University Faculty of Medicine,Istanbul, Turkey; 3Department of Pediatrics, Bakirkoy Maternity and Children's Educational and Treatment Hospital, Istanbul, Turkey; 4Department ofMicrobiology, Bakirkoy Maternity and Children's Educational and Treatment Hospital, Istanbul, Turkey

A B S T R A C T

Objective: To evaluate the direct medical cost of hospital admissionsfor patients with varicella (i.e., chickenpox) to assess the cost burdenof varicella from a health care perspective for ultimate use in healtheconomics studies in Turkey. Methods: Records of children hospital-ized with varicella at the Bakirkoy Maternity and Children’s Hospitalbetween November of 2006 and June of 2011 were reviewed. Reasonsfor hospitalization, types of varicella-associated complications, anddirect medical cost of hospitalization were noted. Patients withunderlying risk factors were excluded. Data obtained from onehospital were used to estimate the national cost of the disease.Results: During the 4.5-year study period, 234 patients were hospi-talized with varicella. Of these cases, 48 (20%) children previously illwith underlying cancers or chronic diseases were excluded from thestudy. Ultimately, 186 previously healthy children (age range: 14 daysto 159 months, median age: 14 months) were included. The main

reasons for hospitalization were complications related to varicella(79%), the most frequent of which was skin and soft tissue infections,followed by neurological complications and pneumonia. The mediancost of hospitalization per patient was US $283, 50% of which wasattributed to medication costs. The annual cost for varicella hospital-izations in Turkey was estimated at US $396,200. Conclusions: Asignificant number of healthy children are hospitalized for varicellaand associated complications. Descriptions of these complicationsand their related costs provide important data for cost-effectivenessstudies for decisions about the inclusion of the varicella vaccine in achildhood vaccination program.

Keywords: child, complications, cost, varicella.


Introduction

Varicella, otherwise known as chickenpox, that is, the primarymanifestation of a varicella zoster virus (VZV) infection, is gen-erally mild; indeed, severe complications are seldom reported inimmunocompetent children [1]. Nevertheless, hospitalizationsdue to varicella do occur in otherwise healthy children, therebyproducing an economic burden on the health care system [2].Worldwide, the reported incidence of varicella-related hospital-izations involving children varies widely, that is, from 0.9 to 29.4/100,000, depending on the geographic setting and hospital admis-sion policies [3–5]. In Turkey, the exact incidence is unknownbecause varicella is not on the list of tracked diseases; however,estimates indicate that the rate is 6.3/100,000 [6].

A safe and effective vaccine against varicella was developed in1970 and has been made a recommended part of childhoodvaccination programs in several countries. Countries with routinechildhood varicella vaccinations have seen a positive effect ondisease prevention and control [7–13]. In the United States, theannual varicella-related hospitalization rate decreased from 0.5 per10,000 in 1993 to 0.13 per 10,000 in 2001. The incidence of varicella

has also decreased in nonvaccinated groups, including adults andinfants who are too young to be vaccinated, thereby suggesting astrong herd protection effect [7–10]. A widespread varicella vacci-nation program, however, has not yet been introduced everywhere,especially in developing countries.

The health economics of a VZV immunization program is vitalfor decisions on vaccine funding and has been studied in manycountries [14–17]. To facilitate the decision-making processregarding the introduction of a vaccination program, each coun-try needs to collect data on the incidence, complications, and costdue to hospitalizations associated with the particular diseaseunder consideration. Surveillance of varicella complications isalso important to assess the potential impact of a vaccinationprogram. In Turkey, a few studies have evaluated complicationrates; however, knowledge about the cost of varicella hospital-izations is quite limited. A more detailed investigation of the costburden of varicella from a health care perspective can beaccomplished by collecting data about the number and cost ofhospitalizations in a tertiary care hospital in Istanbul, which canthen be extrapolated to the whole country to estimate thenational burden of this disease.




Conflict of interest: The authors report no conflict of interest.


� Address correspondence to: Ozden Turel, Department of Pediatrics, Division of Pediatric Infectious Diseases, School of Medicine,Bezmialem Vakıf University, Adnan Menderes Bulvarı Vatan caddesi 34093 Fatih, Istanbul, Turkey.


dx.doi.org/10.1016/j.vhri.2013.05.003

dx.doi.org/10.1016/j.vhri.2013.05.003

dx.doi.org/10.1016/j.vhri.2013.05.003


dx.doi.org/10.1016/j.vhri.2013.05.003

Methods

Study Design

In this retrospective cohort study, we reviewed the records ofchildren with varicella who were admitted to the BakirkoyMaternity and Children’s Educational and Treatment Hospital(BEH) between November of 2006 and June of 2011. In Istanbul, amajority of patients are treated at secondary and tertiary carehospitals [18]. BEH is one of the main tertiary referral centers forpediatric patients in which 40,000 patients are hospitalized andover 500,000 patients are examined annually [19]. The populationof children aged younger than 15 years in Istanbul is 3,455,049,that is, one-fifth of the population under 15 in Turkey [20].

Data Collection

Enrollment in this study required a discharge diagnosis ofvaricella or its associated complications as defined by the Interna-tional Classification of Disease codes. Further detailed investigationsof medical records of patients hospitalized with varicella wereundertaken to avoid incorrect diagnoses. Data on the demo-graphic features of the patients, their underlying conditions,reasons for hospitalization, types of varicella-related complica-tions, blood culture results, length of hospital stay, outcomes,and costs were collected.

Hospital expenses noted in this study included the cost of theprescribed drugs, doctor visits, nursing care, laboratory andradiological diagnostic tests, bed stay, and other related charges.The records of the patients logged hospital costs in the Turkishlira and were converted into the US dollar.


Version 16 of the Statistical Package for Social Sciences (SPSS forWindows) was used for all statistical analyses. One-way analysisof variance was used to compare continuous data among morethan two groups. Multiple comparisons were analyzed by usingTukey’s honestly significant difference post hoc test, whilePearson’s correlation test was used to assess the relationshipsamong continuous variables.

Results

Of the 684 children with varicella who were examined at BEHduring the study time period, 234 were hospitalized. Of these cases,48 patients (i.e., 20%) had an underlying illness and were thusexcluded from the study (Table 1). Therefore, 186 previously healthychildren (i.e., 55.9% males) were included in this study. Consideringthat BEH provides services to 15% of all children hospitalized inIstanbul and 3% of all children hospitalized in Turkey, the annualnumber of formerly healthy children younger than 15 years who arehospitalized because of varicella in Turkey was estimated at 1400.

The median age of the patients with varicella in our study was14 months (i.e., ranging from 14 days to 159 months). The highestrate of hospitalization occurred in patients younger than age 3years (i.e., 74.15%), of which 64.2% were younger than or equal toage 1 year. The median length of hospital stay for this populationwas 5 days. The majority of the cases were detected in the springand early summer months, with a peak in May (Fig. 1). The mainreasons for hospitalization were complications associated withvaricella (i.e., accounted for 79% of the admissions). Bacterialsuperinfections involving the skin and soft tissues accounted for32.6% of the admissions and were the most frequently observedcomplications; this was followed by neurological complicationsin 29.9% of the admissions and pneumonia in 21.7% of the

admissions (Table 2). The children who were hospitalized forpneumonia were younger than those hospitalized with neuro-logical complications (P o 0.05) (Fig. 2). Most outcomes werefavorable with exception to one child needing thoracal tubeinsertion to treat empyema, one with abduscent nerve paralysis,and one with cellulitis that resulted in severe scarring.

The median cost of hospitalization per patient was US $283, 50%of which was attributed to medication costs (Table 3). The costs forphysician visits were lowest among the cost categories listed inTable 3 because revisits are not billed in accordance with currentTurkish regulations. A positive correlation was observed between thetotal costs and the age of the patients (r ¼ 0.27, P o 0.001). Costs forhematological complications were higher than the costs for any ofthe other complications. For the five patients with the highest totalcosts, three had hematological complications and received bloodtransfusions; for the remaining two, one had septicemia and theother child had severe cellulitis that necessitated 26 days of hospital-ization. The direct annual cost for otherwise healthy childrenhospitalized with varicella in Turkey was estimated to be US $396,200.

Discussion

Although varicella complications are believed to be rare inimmunologically healthy children, related hospitalizations have

Table 1 – Varicella-related hospitalized patientswith underlying illnesses.

Underlying illness n %

MalignancyAcute lymphoblastic leukemia 14 29Lymphoma 2 4.2Rhabdomyosarcoma 1 2.1Spinal tumor 1 2.1Neuroblastoma 1 2.1PNET 1 2.1

MetabolicType 1 diabetes mellitus 5 10.4Congenital adrenal hyperplasia 2 4.2Cystic fibrosis 1 2.1Graves’ disease 1 2.1Niemann Pick 1 2.1

HematologicHereditary spherocytosis 2 4.2Diamond Blackfan anemia 1 2.1Thalassemia major 1 2.1Thrombasthenia 1 2.1Factor 7 deficiency 1 2.1Chronic ITP 2 4.2

Primary immunodeficiencySCID 1 2.1IgA deficiency 1 2.1CVID 1 2.1Cyclic neutropenia 1 2.1Other 2 4.2

OtherHoloprosencephaly 1 2.1CMV hepatitis, hyperphenylalaninemia 1 2.1Echinococcal cyst 1 2.1

Total 48 100

CMV, cytomegalovirus; CVID, common variable immunodefi-ciency; IgA, immunoglobulin A; ITP, idiopathic thrombocytopenicpurpura; PNET, primitive neuroectodermal tumor; SCID, severecombined immunodeficiency.


been reported in children without any identified risk factors[2,21]. Of the 24,488 varicella-related hospitalizations during2000-2006 in the United States, 70% were among healthy personswith no contraindications for vaccination [22]. Similarly, wenoted that 80% of the children with varicella who were hospi-talized during the study period at BEH had no underlyingdiseases. Carapetis et al. [2] reported that immunocompromisedpatients with varicella were admitted earlier in their illness andhad lower complication rates than did otherwise healthypatients. For this reason, we conducted a detailed examinationof hospitalizations for varicella only in previously healthy chil-dren. We estimated that hospitalizations due to varicella inotherwise healthy children have an incidence of 7.7/100,000. A

multicenter study in Turkey also showed that 73.3% of hospi-talized patients were previously healthy [23]. The estimatedoverall incidence was 5.29 to 6.89/100,000 in all children aged 0to 15 years; therefore, considering that hospitalized patientsconstitute 1% of all cases of varicella, the overall incidence ofvaricella was estimated as 466 to 768/100,000 [23].

The median age of children hospitalized with varicella-relatedcomplications in the United Kingdom and Ireland was reported as3 years [5]. In our study, children hospitalized for varicella tendedto be younger; indeed, 47.6% of the patients in our study were 1year or younger. Seroprevalence studies in Turkey have shownthat after the decline in maternal antibodies present in thebloodstream at birth, seropositivity rates for VZV were low untilthe end of the first year (i.e., 16.6%) and gradually increased to41.2% at age 5 years [24,25]. Protection of children younger than 1year, however, can be achieved only by increased herd immunitythat is established via a widespread vaccination program becausethe vaccine is not recommended in children younger than 1 year.

A multicenter study in France demonstrated a strong inversecorrelation between levels of circulating anti-VZV maternal anti-bodies in full-term infants and the occurrence of varicella com-plications in children who contract the disease at age youngerthan 1 year [26]. During the first 3 months of life, maternalantibodies against VZV protect most infants, and unless thisimmunity is absent, newborns with mild chickenpox should notrequire antiviral therapy [27]. In our study population, weidentified 21 patients aged 14 days to 2 months who werehospitalized solely for varicella without any complications, andmost of them received antiviral therapy. This indicates thatmany physicians still view the disease as a serious illness duringearly infancy.

The majority of varicella complications identified in our studywere bacterial superinfections, which is similar to the findings ofprevious studies [6,21]. Staphylococcus aureus and group A β-hemolytic streptococci are the main pathogens responsible forbacterial complications related to varicella [28,29]. Pathogenswere isolated from cultures of tissue samples in only five of ourstudy patients (i.e., two had coagulase-negative staphylococciand one had α-hemolytic streptococci from blood cultures,whereas one had methicillin-sensitive S. aureus from abscessmaterial, respectively). The low yield of microorganisms in ourstudy is likely caused by insufficient culturing practices that mustbe improved in future studies and treatment of patients withvaricella.

Neurological complications were the second most commoncomplication in patients who were hospitalized for varicella in

0

5

10

15

20

25

30

35

Num

ber o

f pat

ient

s

Months

Fig. 1 – Number of hospitalized varicella cases according to months of the year.

Table 2 – Reasons for hospitalization in childrenwith varicella.

Cause n (%)

Complications 147 (79)Skin, mucosa and soft tissue infections 48 (25.8)

Pyoderma 25Cellulitis 12Abscess 5Cervical lymphadenitis 2Stomatitis 4

Neurologic 44 (23.7)Cerebellar ataxia 19Convulsion 17Meningoencephalitis 5Papilloedema, optic neuritis 1

Pneumonia 32 (17.2)Other 23 (12.3)

Hematologic (thrombocytopenia) 7Sepsis 8Upper respiratory tract infection 4Arthritis 2Gastrointestinal 2

Other causes 39 (21)Ageo2 months 21Fever 8Generalized skin lesions 3Poor feeding 3Poor general appearance 2

Total 186 (100)


our study. Bozzola et al. [30] reported that although neurologicalcomplications did not usually result in permanent sequelae, theycould lead to prolonged hospital stays and other indirect costs.

The indirect costs from absenteeism and loss of associatedproductivity are more important than the direct health care costs[31]. The proportion of indirect costs in the total cost of varicellawas reported to range from 42% to 98% in different sudies [32].The highest percentage reported by Lieu et al. [33] was explainedby the inclusion of the cost of death or prolonged disabilityresulting from the disease. Only direct costs were evaluated inour study.

The costs of varicella-related hospitalizations may varyaccording to the epidemiologic features, management strat-egies of the disease, and the type of complications. Zhou et al.[17] estimated that in 2006, the direct per case medical cost ofhospitalization for uncomplicated varicella, varicella pneu-monia, and varicella encephalitis was US $3,317, US $4,213,and US $12,064, respectively. In Australia, the cost of hospital-ization for children with varicella was reported to be US $3,272per case in 2004 [2]. In Spain, hospitalization cost per case was

US $5,113, with the total associated national cost excludingsymptomatic treatment at US $516,531 [34]. We estimated thedirect national cost of varicella-related hospitalizations as US$396,200 in Turkey. Differences in health service fees such ashospital and physician costs may explain why hospital care inTurkey costs less than in the United States, Australia, andSpain. For Brazil, a country with a socioeconomic backgroundsimilar to that of Turkey, Valentim et al. [14] reported that thehospital cost for a case of varicella was US $439. In China,inpatient care cost was US $640 per case [35]. In previousstudies, the costs of hospitalizations for varicella were sig-nificantly higher in children who had underlying diseases[15,23]. Only otherwise healthy children were included inour study.

Decisions on vaccine funding are often based on a number ofvariables, such as immunogenicity of the vaccine and the cost-effectiveness of the immunization program. The varicella vaccineprevents the disease in 85% of immunized children but offers a97% protection against its most severe forms [36]. Moreover, theuniversal two-dose immunization has been shown to be cost-effective in Western temperate countries [37]. In Brazil, the cost-effectiveness of a universal vaccination program against varicellawas dependent on the vaccine price and the required number ofdoses [14].

In Turkey, one dose of varicella vaccine costs US $45 andthe national birth cohort is 1.2 million. As a preventivemeasure, a universal vaccination program will require a largeinvestment. In this study, however, we confirmed that hospi-talization in varicella cases contributes to a significant costburden even in previously healthy children. Our estimates ofcosts should contribute to future cost-effectiveness research inTurkey.

Conclusions

In Turkey, a significant number of otherwise healthy childrenhave been hospitalized because of varicella and its complications.Data gathered about hospitalization expenses provide important

P<0.05

(hematologic, sepsis, upper respiratory tract infection, arthritis, gastrointestinal)

Fig. 2 – Distribution of complications according to age of the patients.

Table 3 – Total hospital costs (US dollar).

Variable Cost

Mean SD Median Min Max

MED 206 469 108 0 4529PHY 7 12 4 0 91LAB 60 80 30 0 515NC 28 30 20 0 185HOS 106 73 88 0 696Total 407 552 283 15 4886

MED, medications; PHY, physicians’ fees for examinations andconsultations; LAB, diagnostic tests and radiological examinations;NC, nurse care; HOS: bed stay; SD, standard deviation.


knowledge for future cost-effectiveness studies. This informationwill, therefore, assist policymakers in decisions about whether toinclude the varicella vaccine in childhood vaccination programs.Rates of complications may provide significant backgroundknowledge for evaluating the impact of vaccinations againstchildhood varicella.

Source of financial support: The authors have no otherfinancial relationships to disclose.

R E F E R E N C E S

[1] Heininger U, Seward JF. Varicella. Lancet 2006;368:1365–76.[2] Carapetis JR, Russell DM, Curtis N. The burden and cost of

hospitalized varicella and zoster in Australian children. Vaccine2004;23:755–61.

[3] Liese JG, Grote V, Rosenfeld E, et al. The burden of varicellacomplications before the introduction of routine varicella vaccinationin Germany. Pediatr Infect Dis J 2008;27:119–24.

[4] Dubos F, Grandbastien B, Hue V, et al. Epidemiology of hospitaladmissions for pediatric varicella infections: a one-year prospectivesurvey in the pre-vaccine era. Epidemiol Infect 2007;135:131–8.

[5] Cameron JC, Allan G, Johnston F, et al. Severe complications ofchickenpox in hospitalized children in the UK and Ireland. Arch DisChild 2007;92:1062–6.

[6] Koturoglu G, Kurugol Z, Cetin N, et al. Complications of varicella inhealthy children in Izmir, Turkey. Pediatr Int 2005;47:296–9.

[7] Davis MM, Patel MS, Gebremariam A. Decline in varicella-relatedhospitalizations and expenditures for children and adults afterintroduction of varicella vaccine in the United States. Pediatrics2004;114:786–92.

[8] Hambleton S, Phill D, Gershon AA. The impact of varicella vaccinationin the United States. Semin Pediatr Infect Dis 2005;16:38–43.

[9] Guris D, Jumaan AO, Mascola L, et al. Changing varicella epidemiologyin active surveillance–United States. 1995–2005. J Infect Dis 2008;197(Suppl. 2):S71–5.

[10] Marin M, Watson TL, Chaves SS, et al. Varicella among adults: datafrom an active surveillance project, 1995–2005. J Infect Dis 2008;197(Suppl. 2):S94–100.

[11] Quian J, Rüttimann R, Romero C, et al. Impact of universal varicellavaccination on 1-year-olds in Uruguay: 1997–2005. Arch Dis Child2008;93:845–50.

[12] Cenoz MG, Catalán JC, Zamarbide FI, et al. Impact of universalvaccination against chickenpox in Navarre, 2006–2010. An Sist SanitNavar 2011;34:193–202.

[13] Pozza F, Piovesan C, Russo F, et al. Impact of universal vaccination onthe epidemiology of varicella in Veneto, Italy. Vaccine 2011;29:9480–7.

[14] Valentim J, Sartori AM, de Soárez PC, et al. Cost-effectiveness analysisof universal childhood vaccination against varicella in Brazil. Vaccine2008;26:6281–91.

[15] Azzari C, Massai C, Poggiolesi C, et al. Cost of varicella-relatedhospitalizations in an Italian pediatric hospital: comparison withpossible vaccination expenses. Curr Med Res Opin 2007;23:2945–54.

[16] Tseng HF, Tan HF, Chang CK. Varicella epidemiology and cost-effectiveness analysis of universal varicella vaccination program inTaiwan. Southeast Asian J Trop Med Public Health 2005;36:1450–8.

[17] Zhou F, Ortega-Sanchez IR, Guris D, et al. An economic analysis of theuniversal varicella vaccination program in the United States. J InfectDis 2008;197(Suppl 2):S156–64.

[18] Istanbul Health Directorate. Istanbul health statistics 2010. Availablefrom: http://www.istanbulsaglik.gov.tr/w/anasayfalinkler/pano3.asp.[Accessed April 10, 2013].

[19] Istanbul Health Directorate. Kanuni Sultan Süleyman Educational andTreatment Hospital. Available from: http://www.istanbulsaglik.gov.tr/w/sb/tedk/bakirkoy.asp. [Accessed April 10, 2013].

[20] Turkish Statistical Institute. Address-based population registrationsystem. Available from: www.tuik.gov.tr/VeriTabanlari.do?vt_id=9&ust_id=5. [Accessed April 10, 2013].

[21] Grimprel E, Levy C, de La Rocque F, et al. Pediatric varicellahospitalizations in France: a nationwide survey. Clin Microbiol Infect2007;13:546–9.

[22] Lopez AS, Zhang J, Brown C, et al. Varicella-related hospitalizations inthe United States, 2000–2006: the 1-dose varicella vaccination era.Pediatrics 2011;127:238–45.

[23] Dinleyici EC, Kurugol Z, Turel O, et al. The epidemiology and economicimpact of varicella-related hospitalizations in Turkey from 2008 to2010: a nationwide survey during the pre-vaccine era (VARICOMPstudy). Eur J Pediatr 2012;171:817–25.

[24] Alp H, Altinkaynak S, Ertekin V, et al. Seroepidemiology of varicella-zoster virus infection in a cosmopolitan city (Erzurum) in the easternTurkey. Health Policy 2005;72:119–24.

[25] Savas S, Dallar Y, Arikan I, et al. Varicella-zoster virus seroprevalencein children between 0-15 years old. Mikrobiyol Bul 2004;38:69–75.

[26] Pinquier D, Lécuyer A, Levy C, et al. Inverse correlation between varicellaseverity and level of anti-varicella zoster virus maternal antibodies ininfants below one year of age. Hum Vaccin 2011;7:534–8.

[27] Lécuyer A, Levy C, Gaudelus J, et al. Hospitalization of newborns andyoung infants for chickenpox in France. Eur J Pediatr2010;169:1293–7.

[28] Lesko SM, O’Brien KL, Schwartz B, et al. Invasive group A streptococcalinfection and nonsteroidal antiinflammatory drug use among childrenwith primary varicella. Pediatrics 2001;107:1108–15.

[29] Peterson CL, Mascola L, Chao SM, et al. Children hospitalized forvaricella: a prevaccine review. J Pediatr 1996;129:529–36.

[30] Bozzola E, Tozzi AE, Bozzola M, et al. Neurological complications ofvaricella in childhood: case series and a systematic review of theliterature. Vaccine 2012;30:5785–90.

[31] Preblud SR. Varicella: complications and costs. Pediatrics 1986;78:728–35.[32] Soárez PC, Novaes HM, Sartori AM. Impact of methodology on the

results of economic evaluations of varicella vaccination programs: is itimportant for decision-making? Cad Saude Publica 2009;25(Suppl. 3):S401–14.

[33] Lieu TA, Cochi SL, Black SB, et al. Cost-effectiveness of a routinevaricella vaccination program for US children. JAMA 1994;271:375–81.

[34] Piqueras Arenas AI, Otero Reigada MC, Pérez-Tamarit D, et al.Hospitalizations for varicella in the Hospital Infantil La Fe, Valencia,Spain, 2001–2004. An Pediatr (Barc) 63:120–4.

[35] Da YP, Luo LY, Song LZ. Economic burden of inpatient of varicella inShandong, Gansu, and Hunan provinces, 2007. Zhongguo Yi Miao HeMian Yi 2009;15:438–42.

[36] Gershon AA, Takahashi M, Seward J. Varicella vaccine. In: PlotkinSA, Orenstein WA, eds., Vaccines (4th ed.). Philadelphia: WB Saunders,2004.

[37] Flatt A, Breuer J. Varicella vaccines. Br Med Bull 2012;103:115–27.


















































http://www.istanbulsaglik.gov.tr/w/anasayfalinkler/pano3.asp


http://www.istanbulsaglik.gov.tr/w/sb/tedk/bakirkoy.asp

http://www.istanbulsaglik.gov.tr/w/sb/tedk/bakirkoy.asp


www.tuik.gov.tr/VeriTabanlari.do?vt_id=9&ust_id=5






















































PATIENT-REPORTED OUTCOMES

The Methodological Challenges for the Estimation of Quality of Life inChildren for Use in Economic Evaluation in Low-Income CountriesTravor Mabugu1,*, Paul Revill2, Bernard van den Berg2

1Clinical Research Centre, University of Zimbabwe, Harare, Zimbabwe; 2Centre for Health Economics, University of York, York, UK

A B S T R A C T

Objectives: The assessment of quality of life (QOL) in children has beenunderresearched in high- and low-income countries alike. This is partlydue to practical and methodological challenges in characterizing andassessing children’s QOL. This article explores these challenges andhighlights considerations in developing age-specific instruments for chil-dren affected by HIV and other health conditions in Africa and other low-income settings. Methods: A literature search identified works that have1) developed, 2) derived utilities for, or 3) applied QOL tools for use ineconomic evaluations of HIV interventions for children. We analyzed theexisting tools specifically in terms of domains considered, variations in agebands, the recommended respondents, and the relevance of the tools toAfrican and also other low-income country contexts. Results: Only limitedQOL research has been conducted in low-income settings on either adultsor children with HIV. A few studies have developed and applied tools for

children (e.g., in Thailand, Brazil, and India), but none have been in Africa.The existing methodological literature is inconclusive on the appropriatewidth or depth by which to define pediatric QOL. The existing instrumentsinclude QOL domains such as “physical functioning,” “emotional andcognitive functioning,” “general behavior (social, school, home),” “healthperception,” “coping and adaptation,” “pain and discomfort,” “extendedeffects,” “life perspective,” and “autonomy.” Conclusions: QOL assessmentin children presents a series of practical andmethodological challenges. Itsapplication in low-income settings requires careful consideration of anumber of context-specific factors.

Keywords: children, HIV/AIDS, low-income, quality of life.


Introduction

The assessment of quality of life (QOL) is crucial to informcomparisons of the effectiveness of health care and public healthinterventions, comparative health research, performance meas-urement, purchasing decisions, and economic evaluations inhigh- and low-income countries alike (see Smith et al. [1] for aconceptual overview of the QOL literature). Much less progresshas been made in QOL assessment in children than in adulthealth. The unique challenges of child health state assessmenthave been highlighted elsewhere [2,3]. Ungar and Gerber [4]explain that children start life as vulnerable infants and developtoward independent persons. Throughout this process the childexperiences change in dependency relations with parents, rela-tives, friends, teachers, neighbors, and people from the commun-ity and encounters various types of health care workers such asdoctors, nurses, and community health workers. The nature ofthe child’s relationships and his or her encounters with healthcare workers are likely to significantly affect patterns of healthcare use, which then impacts his or her health status.

The development of QOL instruments for children involvestackling a series of practical and methodological issues that in

many cases are simpler to address for adults. From a practicalviewpoint, it is simply less straightforward to ask children todescribe the domains that make up their QOL or to assess theirown QOL because they might find it hard to clearly communicatetheir thoughts and feelings. They can also face difficulties indistinguishing between good and ill health or well-being, andpermanent and transitory health problems. Approaches thathave been proposed to overcome these difficulties include usinginnovative instruments to aid understanding (e.g., use of pictures[5]), varying questions and tools by age (e.g., see PaediatricQuality of Life Inventory [PedsQL] in Overview of Children’s QOLInstruments section) and stage of development of children, anddelivering questions to adult proxies responsible for children(such as parents or carers). Obviously, the responses of childrenand their proxies can vary [6]. They have been shown to differmost widely with respect to domains around social and emo-tional functioning, whereas they seem more similar withinphysical activity, functioning, and symptoms domains [7].

Effectiveness research on child health interventions in low-income countries has usually relied solely on intermediateclinical markers (used primarily to inform narrow clinical deci-sions) instead of QOL measures, which aim at assessing the wider




Conflict of interest: The authors have indicated that they have no conflicts of interest with regard to the content of this article.


� Address correspondence to: Travor Mabugu, Clinical Research Centre, University of Zimbabwe, Parirenyatwa Hospital Annexe, Cnr JTongogara/Mazowe Street, Harare, Zimbabwe.


dx.doi.org/10.1016/j.vhri.2013.07.005

dx.doi.org/10.1016/j.vhri.2013.07.005

dx.doi.org/10.1016/j.vhri.2013.07.005


dx.doi.org/10.1016/j.vhri.2013.07.005

impact of a disease on a child’s life and his or her socialenvironment (see Wilson and Cleary [8] on the conceptualdistinction between clinical and QOL outcomes). The usefulnessof these narrow clinical indicators to evaluate a range of medicalalternatives is limited. Instead, particularly when adopting a publichealth approach, there is a need to understand the effect of healthcare on a person’s QOL in whatever way this is defined. Theestimation of children’s QOL should therefore be a central priority.

In addition to the practical challenges in assessing children’sQOL, there are considerable methodological challenges. This articleaims to contribute to the literature by debating methodologicalchallenges related to pediatric QOL measurement in low-incomesettings. By doing so we challenge existing instruments based on thescope and perspective of existing pediatric QOL tools. We also aim toinform how pediatric QOL estimation can be used within economicevaluation. Although the article has a general focus—contributing tooutcome research in general—special attention is paid to addressingthese challenges of QOL assessment for use in the economicevaluation of health care, and especially HIV/AIDS interventions.

The article proceeds as follows. A literature search for QOLassessment in children is outlined. Based on the literature searchconsiderations of the QOL domains, suggested respondents andage bands are described. Next, methodological considerations fortransforming QOL outcomes for economic evaluation are presented

together with a general typology on the forms of economicevaluation that QOL estimation can inform. The article concludesafter a general discussion on the scope of the implications of ourarticle and a reflection on our work compared with other studies.

Literature Search

Search

The literature search was conducted by the first author. Hefollowed the Preferred Reporting Items for Systematic Reviewsand Meta-Analyses guidelines. The primary search term used was“quality of life” and the following secondary search terms wereused: “children,” “HIV,” “AIDS,” “instrument,” and “measurement.”The PubMed and PubMed Central search engines were used. Inaddition, the following selected journals were searched: Quality ofLife Research, Pharmacoeconomics, Journal of Health Economics, AIDS &STD Patient Care, and Cost-Effectiveness and Resource Allocation.

Eligibility Criteria

Included articles 1) developed QOL instruments for, 2) derivedutilities (values) for, or 3) applied tools in economic evaluations of

Ar�cles iden�fied through selected Journals

Pharmaco-economics (n = 300)Health Economics (n = 312)

Cost-effec�veness & Resource Alloca�on (n = 35)AIDS Pa�ent Care & STDs (n = 424)Quality of Life research (n = 113)

Ar�cles iden�fiedthrough PMC

database searching(n = 159)

Ar�cles iden�fiedthrough PubMed

databasesearching(n = 393)

Iden

�fic

a�on

Records a�er duplicates removed and addi�on of informal search ar�cles(n = 1495)

Records excluded a�erini�al screening by �tle

and abstract(n = 1444)

Full-text ar�cles assessedfor eligibility

(n = 51)

Full-text ar�cles excluded,with reasons

(n = 35)

Studies included inqualita�ve synthesis

(n = 13)

Scre

enin

gEl

igib

ility

Incl

uded

Fig. 1 – Flow chart of literature search events. Thirty-five articles were excluded for a number of reasons, with the main reasonsfor exclusion being as follows: assessed the QOL of adults with HIV-infected children (13 articles), commentaries (3) or letters (4)that highlighted measurement issues of QOL in HIV-infected children, compared the type of interview (researcher vs. self-administered interviews) (2 articles), the tools used applied to children older than 18 years (8 articles), and reviews thathighlighted general measurement of health outcomes in both children and adults (5 articles). We ensured that tools used ormentioned in these excluded articles were captured in the final 13 articles used in the analysis. QOL, quality of life.


HIV interventions in children. Articles published in any languageother than English were excluded. The focus population waschildren in low-income settings, but particular attention paid tothe QOL of children with HIV in Africa. See Figure 1 for moredetails.

Overview of Children’s QOL Instruments

The overview revealed an appreciable body of work on QOLresearch in HIV/AIDS conducted internationally [9], includingsome studies in sub-Saharan Africa [10,11], as well as for children[12–14], but not for children in Africa.

Table 1 shows all articles for which pediatric QOL tools forHIV/AIDS have been developed or applied and names the toolsthat were taken forward for qualitative analysis. The QOL instru-ments are either generic health status instruments (in whichscores can be compared for many diseases, e.g., Child Quality ofLife Scale [CQLS], PedsQL, and Quality of Life Assessment Ques-tionnaire [QLAQ]) or HIV condition/disease-specific instruments(which measure narrowly focused patient views on the impact ofHIV/AIDS). The literature search yielded a mixture of these toolsbut mostly generic instruments.

The identified tools have all been developed for use primarily inhigh- and upper-middle-income countries. A few studies haveapplied these tools in middle-income countries such as Thailand,Brazil, and India. Zambia is the only lower-middle income countryin Africa that has reported application of child utilities in economicevaluation [15]. In this study, adult utilities were extrapolated foruse in children, and the authors highlight lack of child-specific QOLweights appropriate to lower-income settings.

The dimensions or domains of the retrieved tools are sum-marized in Table 2. The number of domains varies from 4 (inPedsQL 4.0) to 12 (Child Health Questionnaire 28). The mostcommon domains included in the tools were variants of physical,emotional, and social functioning, and general health perception.The CQLS contains leisure and family domains, and the QLAQhas a measure of physical resilience. Despite requests we wereunsuccessful in obtaining access to the Thai Quality of Life forHIV Infected Children tool, and therefore it was not included insubsequent analyses.

Table 3 shows the origin of the tools considered in thisanalysis. It also gives a summary of the key factors that wereconsidered in the elicitation of responses, how the summaryscores were obtained, and which age groups were considered inthe studies analyzed.

An important issue when assessing child QOL is the determi-nation of an appropriate respondent—in particular whether itshould be a child and/or a proxy (parent or carer). Only the PedsQL,Thai Quality of Life for Children, and Child Health Questionnaire 28allow for both adult proxy responses and child self-responses. Otherinstruments recommend either adult proxies (e.g., the QLAQ andGeneral Health Assessment for Children) or child self-response (e.g.,the CQLS). Table 3 also contains a subanalysis on the age sensitivityof the identified tools. We also note the width of age variation, andthis is further discussed below. Most instruments were developedin what would be categorized as middle- to high-income countriesusing standard World Bank classifications [16]. Application of thesetools in lower-income settings can raise a number of concernsassociated with the diverging socioeconomic and cultural contextsof these settings, although some similarities can also be drawnbetween these different economies. We believe that three primaryquestions need to be addressed: 1) which dimensions and 2) whichage bands apply in the African context, and finally 3) whoseresponse is most appropriate to elicit (the child’s or a proxy’s).

Key Issues to Consider in Assessment of QOL inChildren

What Dimensions to Consider?

Table 4 presents the domains that have been used to assess childQOL in existing instruments.

This table generally suggests, as observed above, that majordomains such as physical, psychological, and social functioningand general health perception form basic and fundamentalpillars of a child’s QOL. Physical functioning is an assessmentof a child’s ability to perform daily tasks and includes sitting,walking, running, and playing. It is about general motility.Psychological functioning assesses aspects such as cognitiveand attentive abilities, emotions, personality, behavior, and inter-personal relationships. Social wellness assesses internal andexternal engagement at all levels (such as at the micro level ofindividual agency and the macro level of systems at school,home, and other social structures). Finally, general health per-ception assesses sentiments relating to current health status. It isa function of an individual’s perception of the extent of deviationin health status from a desired or aspired level, and is oftenaffected by changes in health status over time.

In addition, other descriptors of QOL that need to be exploredfor importance and relevance include domains such as lifeperspective; autonomy, pain, and discomfort; extended effects;and coping and adaptation strategies.

Whose Response to Consider?

It is inevitable that different individuals report QOL differently,especially if they are proxy respondents. Various potentialchoices exist for pediatric study populations. A key distinctioncan be made between children themselves and adult proxies—either a parent or caregiver, or a health care provider. Theliterature highlighted here is inconclusive on this issue. A moreprecise response can be obtained from self-responses, althoughyounger children face more cognitive difficulties in expressingthemselves than do older children. Therefore, because of devel-opmental and cognitive changes that occur as the child grows(see Is Age Sensitivity an Issue? section), it may be useful toobtain either child and/or adult proxy responses depending onthe age of the child. Adult proxy responses represent closesubstitutes for younger children who are not able to comprehendand express themselves clearly. As children grow older, however,their responses can then be considered.

Is Age Sensitivity an Issue?

A closely related consideration to determining the appropriaterespondent is to reflect childhood developments in QOL measure-ment based on a child’s age. A child’s life trajectory is typicallycharacterized by development in cognitive abilities and changingdependency relations as the level of autonomy increases. Thestudies highlighted in the review do not offer one consistentapproach to address the age-dependent elements of pediatricQOL measurement (see Table 3).

If we consider the ecological model of influences on childdevelopment [4], different systems exert differing effects on thedevelopment of children and are each introduced at varying ages(and with some degree of overlap). This results in unpredictableimpacts on the cognitive abilities of a child. Different healthsystems in Africa have categorized children by age to informclinical practice. In Malawi, for example, children are groupedinto the following age groups: 0 to 2 years, 3 to 5 years, and 6 to 15years. Appropriate tools should therefore be age sensitive toreflect key developmental milestones that affect pediatric QOL.


Table 1 – Articles included for qualitative data analysis.

Code Title Authors Year Journal Country Objectives Populationgroup

(samplesize)

Eligibilitycriteria

Instrumentused

GNI percapita

Setting

AA22

Oral health relatedquality of life ofpaediatric patientswith AIDS

Massarenteet al. [33]

2011 BMC OralHealthJournal

Brazil Assessment of oralHRQOL of HIV/AIDS-infected children andassociated factors

10–15 y (88) Valueelicitation

OHR-QoL 11,420.00 MIC

AA24

Poor quality of lifeamong untreatedThai & Cambodianchildren withoutsevere HIVsymptoms

Bunupiradahet al. [34]

2011 AIDS Care Thailand,Cambodia

To evaluate QOL inuntreated Thai andCambodian childrenwith HIV who do nothave severe HIV

1–11 y (294) Tooldevelopmentand valueelicitation

GCHA 8,360.00and2,230.00

UMIC/LMIC

AA25

Health related qualityof life assessmentquestionnaire forchildren aged 5-11years with HIV/AIDS:cross-culturaladaptation forPortuguese language

Costa et al.[14]

2011 Cad. SaúdePública,Rio deJaneiro

Brazil To assess cross-culturaladaptation of theQLAQ used tomeasure the HRQOLin Brazilian childrenaged 5–11 y with HIV/AIDS

5–11 y (35) Tooldevelopment

QLAQ 11,420 MIC

AA29

Development of theEQ-5D-Y—a childfriendly version ofthe EQ-5D

Wille et al.[31]

2010 Quality ofLifeResearch

Germany,Italy,Spain,andSweden

To develop a self-reportversion of the EQ-5Dquestionnaire foryounger respondents(EQ-5D-Y), and testcomprehensibility ofchildren andadolescents withHIV/AIDS

48 y Tooldevelopment

EQ-5D-Y 40,170.00/32420.00/31,440.00/42,210.00

HIC (all)

AA17

Quality of life ofchildren living inHIV/AIDS-affectedfamilies in ruralareas in Yunman,China

Xu et al. [13] 2010 AIDS Care China To explore factorsinfluencing theHRQOL of childrenliving in HIV-affectedfamilies in ruralareas in Yuunan,China.

8–17 y (225) Valueselicitation

PedsQL 4.0 8,390.00 MIC

AA18

Health-related qualityof life in HIV-infected childrenusing PedsQL™ 4.0and comparisonwith uninfectedchildren

Banerjeeet al. [35]


India Assess reliability andvalidity of PedsQL 4.0in children with HIV,and the associationof HIV infectiontreatment regimens

8–12 y (300) Valueselicitation

PedsQL 4.0 3,620.00 LMIC

VALUE

INH

EALTH

REGIO

NAL

ISSUES

2(2

013)231–239

234

and type of carereceived on QOL

AA19

Quality of life andpsychosocialfunctioning of HIV-infected children

Das et al. [36] 2010 IndianJournal ofPaediatrics

India To assess the QOL andpsychosocialproblems of HIV-infected children, incomparison tochildren with cysticfibrosis

Z6 y(71) Valueelicitation

PedsQL 4.0 3,620.00 LMIC

AA20

Health-related qualityof life of Thaichildren with HIVinfection: acomparison of theThai Quality of Lifein children (TlQLC)with the PedsQL™4.0 generic corescales

Punpanichet al. [37]


Thailand Assess reliability andvalidity of the ThQLCinstrument incomparison with thePedsQL 4.0 inchildren on long-term HIV care

8–16 y (292) Valueelicitation

PedsQL 4.0and ThQLC

8,360.00 MIC

AA21

Development of aculturallyappropriate health-related quality of lifemeasure for HIV-infected children inThailand

Punpanichet al. [38]

2010 Journal ofPeadiatricsand ChildHealth

Thailand To develop a reliableand valid self-reported HRQOLinstrument for HIV-infected children inThailand

Z8 y (292) Tooldevelopment

ThQLHC 8,360.00 MIC

AA27

Impact of disclosure ofHIV infection onHRQoL amongchildren andadolescents withHIV infection

Butler et al.[39]

2009 Paediatrics USA, PourteRica

To examine the impactof HIV disclosure onpediatric QOL

45 y (395) Valueselicitation

GHAC 48,820.00/- HIC/HIC

AA26

Cost- effectiveness ofcotrimoxazoleprophylaxis in HIV-infected children inZambia

Ryan et al.[15]

2008 AIDS Zambia Assess the CEA ofcotrimoxazoleprophylaxis in HIV-infected children inZambia

1–14 y (534) Utilitiesequated

VAS, TTO, SG 1,490.00 LMIC

AA28

Psychometrics of childhealth questionnaireparent short form(CHQ-28) used tomeasure quality oflife in HIV-infectedchildren in complexanti-retroviraltherapy

Byrue et al.[40]


USA To expand thepsychometricproperties of theChild HealthQuestionnaire 28(CHQ-28) parentreport short form as ameasure of well-being for childrenwith chronic illness

5–18 y(33) Tooldevelopment/adjustment

CHQ-28 48,820.00 HIC

VALUE

INH

EALTH

REGIO

NAL

ISSUES

2(2

013)231–239

235

Children aged 2 years or younger generally require the full-timeattention of the mother or another adult. The mother or carer isthen fully responsible for assessing the welfare of the child byusing means such as interpreting their expressions. Childrenaged 3 to 5 years old have some increased autonomy and willlikely spend some time away from the major caregiver, be givenopportunity to play alone or with others, and will perhaps spendtime in day care activities. They also have an improved ability toreport ill health and discomfort, even though they may not beable to describe this in detail. Children aged 6 to 12 yearsexperience life increasingly away from family members. Theyare often introduced to formal education (although not always inthe African context), and the expectations placed on them aresomewhat higher—for instance, they may need to do homeworkor help with housework. In the age range 13 to 15 years, childrenhave their final years of childhood. They start to experiencehormonal changes and transition to life as adults.

Using QOL Measurement in Economic Evaluation

After the outcomes of health care interventions have beenmeasured, using a context-sensitive QOL tool, there is a need to

Table

1–co

ntinued

Cod

eTitle

Auth

ors

Yea

rJourn

alCou

ntry

Objectives

Population

grou

p(sam

ple

size

)

Eligibility

criteria

Instru

men

tuse

dGNIper

capita

Setting

AA 23

Validation

studyofa

scaleoflife

quality

evaluation

ina

groupofpae

diatric

patients

infected

with

HIV

Ferreira

etal.

[41]

2005

Ciência

Saúd

eColetiva

Brazil

Tova

lidateth

esc

aleof

children’s

QOLin

agroupofch

ildren

infected

with

HIV

rece

ivingclinical

care

inBrazil

4–12

y(100

)Tool

dev

elopmen

tan

dap

plica

tion

CQLS

11,420

.00

LMIC

CEA

,cost-effec

tive

nes

san

alys

is;C

HQ-28,

ChildHea

lthQues

tionnaire

28;C

QLS

,ChildQualityofLife

Scale;

EQ-5D,E

uro

Qolfive

-dim

ensional;G

HAC,G

eneral

Hea

lthAsses

smen

tforChildren;H

IC,

highinco

meco

untry;

HRQOL,

hea

lth-related

qualityoflife;L

MIC

,lower

middle

inco

meco

untry;

OHR-Q

oL,

oralhea

lthrelated-q

ualityoflife;P

edsQ

L,Pa

ediatric

QualityofLife

Inve

ntory;Q

LAQ,

QualityofLife

Asses

smen

tQues

tionnaire;QOL,

qualityoflife;SG

,standardga

mble;

ThQLC

,Thai

QualityofLife

forChildren;ThQLH

C,Thai

QualityofLife

forHIV

Infected

Children;TTO,time

trad

e-off;UMIC

,upper

middle

inco

meco

untry;

VAS,

visu

alan

alogu

esc

ale.

Table 2 – List of domains found in identified tools.

Number Instrument Domains considered

1 PedsQL4.0 Physical, emotional, social, andschool functioning

2 ThQLC Physical, emotional, social, life, andschool functioning

3 ThQLHC HIV-targeted tool used to assesssymptoms associated with HIV inchildren

4 CQLS Autonomy, leisure, functions, andfamily

5 QLAQ Physical, psychological, social/rolefunctioning, health perceptions,health care utilization, andphysical resilience

6 EQ-5D-Y Mobility, self-care, pain/discomfort,anxiety/depression, and usualactivities

7 GHAC General health perception,symptom distress, psychologicalstatus, and physical functioning

8 CHQ-28 Physical function; limitation tosocial role related to emotional orphysical function; bodily pain;general behavior; mental health;self-esteem; general healthperception; change in health overpast year; impact on parentaltime; impact on parentalemotions; family activities; andfamily cohesion

CHQ-28, Child Health Questionnaire 28; CQLS, Child Quality of LifeScale; EQ-5D-Y, EuroQol five-dimensional questionnaire foryounger respondents; GHAC, General Health Assessment for Chil-dren; PedsQL, Paediatric Quality of Life Inventory; QLAQ, Quality ofLife Assessment Questionnaire; ThQLC, Thai Quality of Life forChildren; ThQLHC, Thai Quality of Life for HIV-Infected Children.�This is an HIV-specific tool and could not access the tool forcomparison purposes.


value these outcomes if these are to be used within economicevaluation.

Economic evaluations can guide decision makers on theallocation of their budgets, whether this is in health care or morespecifically for HIV/AIDS interventions. There are variousapproaches to conducting economic evaluations depending onthe scope and perspective of the study. Literature however pointsbroadly to two predominant forms of analyses: cost-utility (some-times called cost-effectiveness) analyses and cost-benefit analy-ses [17]. The key challenge in all forms of economic evaluation isto capture the impact of various interventions in affecting theQOL of patients, or potential future patients, and their associatedcosts. For pediatric evaluations, there is therefore a need tomeasure the QOL impacts of pediatric interventions, using someform of instrument, and then for these impacts to be valued.These values can then be compared directly to costs to inform anassessment of value for money.

There are then broadly two ways of incorporating QOLestimation into economic evaluation, and the appropriateapproach to valuation relates to the form of analysis adopted:(1) for cost-utility/cost-effectiveness analyses, effects of interven-tions have to be translated into “utilities,” and (2) for cost-benefitanalyses, it is necessary to translate effects of interventions intomonetary values. These two approaches are considered below.

Calculating Utilities for Use within Cost-Utility/EffectivenessAnalyses

Utilities value QOL on a scale in which 1 represents “full health”and 0 represents “death.” This has two major advantages: first,they enable comparisons across different interventions andtherapeutic areas, and second, they can enable reduced mortalityand reduced morbidity to be assessed concurrently.

The applied literature on the evaluation of pediatric inter-ventions has often struggled to determine appropriate utilities inboth high- and low-income country settings. For instance, in anevaluation of prophylaxis for HIV-infected children, Ryan et al.[15] had to apply adult utility values to child health outcomes.The authors explain that there were no utilities available for HIV-infected children and instead they have to rely on the use of adultutilities. This clearly is a limitation of this and other pediatriceconomic evaluations in HIV/AIDS.

Two of the most widely used methods to determine prefer-ence or utility scores are the standard gamble and time trade-off[17,18]. These are choice or preference-based methods in whichstandard gamble involves eliciting preferences from respondents

between health states that are uncertain, whereas time trade-offinvolves making choices on the basis of the length of time spentin alternative health states. Both approaches are relativelyexpensive to undertake on a case-by-case basis. One alternativeis the visual analogue scale, which can be administered alongsideQOL measurement tools to inform utilities [19].

Table 4 – Description of new domains for QOL inchildren.

Domain Description

Physical functioning Ability to sit, walk, run, play,participate in general activities thatrequire use of body, general motility

Emotional andcognitivefunctioning

Mental wellness (neurologicalimpairment, anxiety, depression),emotions (sad, anger, fear,happiness, self-esteem, confidence),and cognitive (intellectual mind,comprehension, etc.)

General behavior(social, school,home)

Behavior at home (e.g., familyparticipation and cohesion), school(class participation, schoolwork,group work, school absenteeism orpreseentism), and social/roleplaying (among other kids)

Life perspective Views on life events, life expectationsHealth perceptions Overall perception of health (includes

perception of current health,changes in health over the previousperiod, health care utilization as aproxy to overall health status)

Copying/adaptation Ability to cope or adapt to changes inhealth or to changes not related tohealth

Pain and discomfort Pain and discomfortExtended effects The external effects of caring for the

sick child, e.g., time, emotions,burden, effort, and impact on yourcurrent health and well-being

Autonomy Examines the level of independence ofthe child, self-reliance

QOL, quality of life.

Table 3 – Origins of tools, income settings, variation in respondents, and age sensitivity of tools.

Tool Country originally developed (and tested) Respondent Age sensitive

Development Tested Income Self Adultproxy

Yes/no

Specify

PedsQL4.0

San Diego, CA San Diego, CA High income Yes Yes Yes 2–18 y

ThQLC Thailand Thailand Upper-middle income Yes Yes Yes 5–16 yCQLS Brazil Brazil Upper-middle income Yes No Yes 4–12 yQLAQ Brazil Brazil Upper-middle income No Yes Yes 5–12 yEQ-5D-

YGermany, Italy, Spain,

and SwedenGermany, Spain, and

South AfricaHigh- and upper-

middle incomeYes No Yes Z8 y

GHAC USA USA High income No Yes Yes Z5 yCHQ-28 USA USA High income Yes Yes Yes Z8 y

CHQ-28, Child Health Questionnaire 28; CQLS, Child Quality of Life Scale; EQ-5D-Y, EuroQol five-dimensional questionnaire for youngerrespondents; GHAC, General Health Assessment for Children; PedsQL, Paediatric Quality of Life Inventory; QLAQ, Quality of Life AssessmentQuestionnaire; ThQLC, Thai Quality of Life for Children.


Calculating Monetary Values for Use within Cost-BenefitAnalyses

A second means of assessing value from HIV/AIDS and healthcare interventions is to assign monetary values to health out-come measures directly.

Monetary valuation can be assigned by using either revealedpreferences or stated preferences (SP) of subjects. The well-beingvaluation method is a relatively new approach in the revealedpreferences tradition to calculate monetary valuations (seePowdthavee and van den Berg [20] or Deaton et al. [21]). In theabsence of a functioning market for particular commodities, suchas is the case for health gains resulting from interventions, an SPapproach usually has to be adopted by an analyst.

One common way to elicit SP to calculate monetary values isto use contingent valuation. This method elicits values by usinghypothetical questions contained in a survey [22,23]. Individualsare asked to directly state their willingness to pay for medicinesor treatments associated with an illness, or for inclusion ofinterventions in health insurance packages or reimbursementlists [24–28].

The related approach of undertaking discrete choice experi-ments asks similar questions but over a whole range of attrib-utes, including health domains and monetary outcomes,enabling monetary values to be placed on the outcomes associ-ated with the receipt of interventions [29].

Discussion and Conclusions

The idea of this article arose during the development of aninstrument for QOL assessment in children with HIV in sub-Saharan Africa. During the process of developing that instru-ment, it became clear that there were quite a few methodologicalchallenges that remained unresolved in the literature—in partic-ular relating to choices of domains, recommended respondents,and child age bands. It also became clear that these are ofrelevance not only to the evaluation of HIV-related outcomes inAfrica but also to pediatric QOL estimation in low-incomecountries more generally. This article therefore is not meant asa research manuscript but instead has outlined some of themethodological challenges. It is hoped this will encourage furtherdebate on methodological issues relating to QOL assessment inchildren. The article has also outlined how QOL outcomes can bevalued for use within economic evaluation studies.

Identifying existing instruments on the basis of the literaturesearch suggested that quite a broad range of QOL domains shouldbe considered for inclusion in any instrument. These could belabeled as “physical functioning,” “emotional and cognitive func-tioning,” “general behavior (social, school, home),” “health per-ception,” “coping and adaptation,” “pain and discomfort,”“extended effects,” “life perspective,” and “autonomy.” We donot claim that our search was exhaustive, for example the CHU9Dtool [42] did not appear in our search. However, the domains arebroadly similar.

This finding is particularly interesting because it suggests thattwo of the most notable existing instruments (PedsQL and Euro-Qol five-dimensional questionnaire for young respondents [EQ-5D-Y]) may have a focus that is too narrow to fully capturechildren’s QOL. The PedsQL is considered one of the mostpromising QOL instruments for children [13]. It includes domainscovering physical, emotional, social, and school functioning. Itdoes not, however, contain domains found in other tools such ashealth perception, coping and adaptation, and pain and discom-fort. The EQ-5D-Y stems from the well-known adult EQ-5Dquestionnaire, which is the main outcome measure in cost-utility analysis [30]. It assesses childhood QOL by using domains

of mobility, self-care, pain/discomfort, anxiety/depression, andusual activities. These match domains used for adults although itis expected that the types of considerations within each maychange between adults and children—for instance, usual activ-ities for adults will typically include work and leisure, whereas forchildren they are more likely to focus around things such asschool and play.

The scope of both the PedsQL and the EQ-5D-Y may thereforenot be wide enough to capture what is really important forchildren’s QOL. This has already notably been acknowledged,for instance, by Willie et al. [31]. The implication, however, is thatthe search for appropriate measurement tools for pediatric QOL,even in high-income countries, is not a concluded task.

This article has also highlighted other considerations relating towhether tools should be catered to particular age groups and therelated question of whose responses to consider (the child’s or anadult proxy’s). As a result of developmental and cognitive changesthat occur as the child grows, we believe that catering tools todifferent age groups is required. This also has implications on whoshould respond to questions. Adult proxy responses represent closesubstitutes for younger children who are not able to comprehend andexpress themselves clearly. As children grow older, however, theirresponses can be considered. Others have discussed QOL measure-ment in pediatric patients with HIV on the basis of demonstratingadequate psychometric properties of existing instruments [12].

Finally, this article has outlined how pediatric QOL estimationcan be used within economic evaluation. This requires valuationof the effects of interventions—either using utilities (for use incost-utility/effectiveness analyses) or assigning monetary values(for use in cost-benefit analysis). The applied QOL literature hastended to use adult health profiling instruments in child healthstate assessment and/or adult preferences for health utilities.This is a major limitation that risks leading to inappropriatepolicy advice on the use of pediatric health care interventions.Although there is a substantial literature on parents’ willingnessto pay even in low-income settings that we may not havecaptured in our search, it is worth emphasizing that in thisliterature it is generally parents who value children’s QOL. Itseems therefore fair to conclude that the more accurate inform-ing of policy requires either both child-specific measurement andvaluation of QOL or better involvement of parents in valuation ofchildren’s QOL. When parents are going to value children’s QOL,the analyst should consider basing QOL estimations on interde-pendent utility functions [32]. It also requires measurement andvaluation to be appropriately culturally sensitive [14]. Regardlessof the valuation approach, it is necessary to be sensitive tocultural variations in the conception of QOL that will ultimatelyaffect the value put on the QOL estimated. Cultural variationsoften correlate with levels of economic well-being, although thisis not necessarily always true. QOL estimation should thereforebe based on a culturally sensitive tool, and valuation should alsobe sensitive to cultural differences and economic conditions.

In conclusion, QOL estimation in children presents a series ofpractical and methodological challenges. Its application in low-income countries requires careful consideration of a number ofcontext-specific factors. The article challenges existing instru-ments to capture a broader range of domains to assess pediatricQOL for effectiveness research. We have also informed howpediatric QOL outcomes can be valued for use within economicevaluation studies.

Acknowledgment

We gratefully acknowledge the suggestions and comments of threeanonymous reviewers. We thank the Department for International


Development and the Medical Research Council Clinical Trials Unitfor supporting this work.

Source of financial support: This work was funded by theDepartment for International Development, through the MedicalResearch Council Clinical Trials Unit, as part of the ARROW:Young Lives project and the Lablite Project.

R E F E R E N C E S

[1] Smith KW, Avis NE, Assmann SF. Distinguishing between quality of lifeand health status in quality of life research: a meta-analysis. Qual LifeRes 1999;8:447–59.

[2] Halfon N, Nweacheck PW. Characterizing the social impact of asthmain children. In: Weiss KB, Buist AS, Sullivan SD,eds., Asthma’s Impacton Society, the Social and Economic Burden. New York: Marcel Dekker,Inc, 2000.

[3] Sameroff A, Chandler M. Reproductive risk and the continuum ofcaretaking casualty. In: Horowitz F,ed., Review of Child DevelopmentResearch. Chicago: University of Chicago Press, 1975.

[4] Ungar WJ, Gerber A. The uniqueness of child health and challenges tomeasuring costs and consequences. In: Ungar WJ,ed., EconomicEvaluation in Child Health. New York, NY: Oxford University Press,2010.

[5] De Civita M, Regier D, Alamgir AH, et al. Evaluating health-relatedquality-of-life studies in paediatric populations: some conceptual,methodological and developmental considerations and recentapplications. Pharmacoeconomics 2005;23:659–85.

[6] Britto MT, Kotagal UR, Chenier T, et al. Differences betweenadolescents’ and parents’ reports of health-related quality of life incystic fibrosis. Pediatr Pulmonol 2004;37:165–71.

[7] Eiser C, Morse R. Quality of life measures in chronic diseases ofchildhood. Health Technol Assess 2001;5:1–157.

[8] Wilson IB, Cleary PD. Linking clinical variables with health-related qualityof life: a conceptual model of patient outcomes. JAMA 1995;273:59–65.

[9] Skevington SM, O’Connell KA. Measuring quality of life in HIV andAIDS: a review of the recent literature. Psychol Health 2002;18(3):331–50.

[10] Robberstad B, Olsen JA. The health related quality of life of peopleliving with HIV/AIDS in sub-Saharan Africa – a literature review andfocus group discussion. Cost Effect Res Alloc 2011;8:5.

[11] Jansen van Rensburg MS. Measuring the quality of life of residents inSADC communities affected by HIV. AIDS Care 2007;21(9):1132–40.

[12] Garvie PA, Lawford J, Banet MS, West RL. Quality of life measurement inpaediatric and adolescent populations with HIV: a review of theliterature. Child Care Health Dev 2009;35:440–53.

[13] Xu T, Wu Z, Rou K, et al. Quality of life of children living with HIV/AIDS-affected families in rural areas in Yunnan, China. AIDS Care PsycholSocio-Med Asp AIDS/HIV 2010;22:390–6.

[14] Costa LS, Dias de Oliveira Latorre MDR. Health related quality of lifeassessment questionnaire for children aged 5-11 years with HIV/AIDS:cross-cultural adaptation for the Portuguese language. CadSaudePublica, Rio de Janeiro 2011;27:1445–9.

[15] Ryan M, Griffin S, Chitah B, et al. The cost-effectiveness of cotri-moxazole prophylaxis in HIV-infected children in Zambia. AIDS2008;22:749–57.

[16] World Bank. How we classify countries [online]. Available from: http://data.worldbank.org/about/country-classifications. [Accessed April 21,2013].

[17] Drummond MF, O’Brien B, Stoddart GL, Torrance GW. Methods for theEconomic Evaluation of Healthcare Programmes (3rd ed.). New York,NY: Oxford University Press, 2005.

[18] Torrance GW. Measurement of health utilities for economic appraisal:a review. J Health Econ 1996;5:1–30.

[19] Parkin D, Devlin N. Is there a case for using visual analogue scalevaluations in cost utility analysis? Health Econ 2006;15:653–64.

[20] Powdthavee N, van den Berg B. Putting different price tags on the samehealth condition: re-evaluating the well-being valuation approach.J Health Econ 2011;30:1032–43.

[21] Deaton A, Forston J, Tortora R. Life (Evaluation), HIV/AIDS, and Death inAfrica. NBER Working Papers 14637. Cambridge, MA: National Bureau ofEconomic Research, Inc., 2009.

[22] O’Brien B, Ganfi A. When do the ‘dollars’ make sense? Toward aconceptual framework for contingent valuation studies in health care.Med Decis Making 1996;16:288–99.

[23] Tsuchiya A, Williams A. Welfare economics and economic evaluation.In: Drummond M, McGuire A,eds., Economic Evaluation in Health Care.New York, NY: Oxford University Press, 2001.

[24] Donaldson C, Shackley P, Addalla M, Miedzybrodzka Z. Willingness topay for antenatal carrier screening for cystic fibrosis. Health Econ1995;4:439–52.

[25] Liu J, James K, Hammitt JW, Liu JL. Mother’s willingness to pay for herown and her child’s health: a contingent valuation study in Taiwan.Health Econ 2000;9:319–26.

[26] Amin M, Khondoker F. A contingent valuation study to estimatethe parental willingness-to-pay for childhood diarrhoea and genderbias among rural households in India. Health Res Policy Systems2004;2:3.

[27] Pinto-Prades J, Loomes G, Brey R. Trying to estimate a monetary valuefor the QALY. J Health Econ 2009;28:553–62.

[28] Van der Star SM, van den Berg B. Individual responsibility and health-risk behaviour: a contingent valuation study from the ex ante societalperspective. Health Policy 2011;101:300–11.

[29] Ryan M. Discrete choice experiments in healthcare. BMJ 2004;328:360.[30] Richardson J, McKie J, Bariola E. Review and Critique of Health Related

Multi Attribute Utility Instruments. Research Paper 64, Centre forHealth Economics, Monash University, Melbourne, Australia, 2011.

[31] Willie N, Badia X, Bonsel G, et al. Development of the EQ-5D-Y: a childfriendly version of the EQ-5D. Qual Life Res 2010;19:875–86.

[32] Van den Berg B, Bleichrodt H, Eeckhoudt L. The economic value ofinformal care: a study of informal caregivers’ and patients’ willingnessto pay and willingness to accept for informal care. Health Econ2005;14:363–76.

[33] Massarente DB, Domaneschi C, Marques HHS, et al. Oral health-relatedquality of life of paediatric patients with AIDS. BMC Oral Health2011;11:2.

[34] Bunupuradah T, Puthanakit T, Kosalaraksa P, et al. Poor quality of lifeamong untreated Thai and Cambodian children without severe HIVsymptoms. AIDS Care. DOI:10.1080/09540121.2011.592815.

[35] Banerjee T, Pensi T, Banerjee D. HRQoL in HIV-infected children usingPedsQLTM 4.0 and comparison with uninfected children. Quality of LifeResearch 2010;19:803–12.

[36] Das S, Mukherjee A, Lodha R, Vatsa M. Quality of Life and PsychosocialFunctioning of HIV Infected Children. Indian J Pediatr 2010;77:633–7.

[37] Punpanich W, Boon-yasidhi V, Chokephaibulkit K, et al. Health-relatedQuality of Life of Thai children with HIV infection: a comparison of theThai Quality of Life in Children (ThQLC) with the Pediatric Quality ofLife Inventory™ version 4.0 (PedsQLTM 4.0) Generic Core Scales. Qualityof Life Research 2010;19:1509–16.

[38] Punpanich W, Hays RD, Detels R, et al. Development of a culturallyappropriate health-related quality of life measure for humanimmunodeficiency virus-infected children in Thailand. Journal ofPaediatrics and Child Health 2011;47:27–33.

[39] Butler AM, Williams PL, Howland LC, et al. Impact of Disclosure of HIVInfection on Health-Related Quality of Life Among Children andAdolescents With HIV Infection. Pediatrics 2009;123:935–43.

[40] Byrne MW, Honig J. Psychometrics of child health questionnaire parentshort form (CHQ-28) used to measure quality of life in HIV-infectedchildren on complex anti-retroviral therapy. Quality of Life Research2005;14:1769–74.

[41] Ferreira D, RomeroLealPassos M, dePaulaMottaRubini N, et al.Validation study of a scale of life quality evaluation in a group ofpediatric patients infected by HIV. Ciência & Saúde Coletiva2011;16:2643–52.

[42] Stevens K, Rathcliffe J. Measuring and Valuing Health Benefits forEconomic Evaluation in Adolescence: An Assessment of the Practicalityand Validity of the Child Health Utility 9D in the Australian AdolescentPopulation. Value in Health 2012;15:1092–9.















































http://data.worldbank.org/about/country-classifications

http://data.worldbank.org/about/country-classifications

http://refhub.elsevier.com/S2212-1099(13)00084-8/otheref0010


















































































The Impact of Pharmaceutical Care Intervention on the Quality of Life ofNigerian Patients Receiving Treatment for Type 2 DiabetesMaxwell O. Adibe, PhD, MPharm, BPharm1,2,�, Chinwe V. Ukwe, PhD, MPharm, BPharm1,2, Cletus N. Aguwa, PharmD1,2

1Department of Clinical Pharmacy and Pharmacy Management, University of Nigeria, Nsukka, Enugu, Nigeria; 2Pharmacotherapeutic Group, Department ofClinical Pharmacy and Pharmacy Management, University of Nigeria, Nsukka, Enugu, Nigeria

A B S T R A C T

Objectives: To evaluate the impact of pharmaceutical care (PC)intervention on health-related quality of life (HRQOL) of patients withtype 2 diabetes. Methods: This study was a randomized, controlledstudy with a 12-month patient follow-up. The study protocol wasapproved by the Research Ethical Committees of the institutions inwhich this study was conducted. A total of 110 patients wererandomly assigned to each of the “intervention” (PC) and “control”(usual care [UC]) groups. Patients in the UC group received the usual/conventional care offered by the hospitals. Patients in the PC groupreceived UC and additional PC for 12 months. The HUI23S4EN.40Q(developed by HUInc - Mark index 2&3) questionnaire was used toassess the HRQOL of the patients at baseline, 6 months, and 12months. Two-sample comparisons were made by using Student’s ttests for normally distributed variables or Mann-Whitney U tests fornonnormally distributed data at baseline, 6 months, and 12 months.

Comparisons of proportions were done by using the chi-square test.Results: The overall HRQOL (0.86 � 0.12 vs. 0.64 � 0.10; P o 0.0001)and single attributes except “hearing” functioning of the patients weresignificantly improved at 12 months in the PC intervention arm whencompared with the UC arm. The HRQOL utility score was highlynegatively (deficit ≥10%) associated with increasing age (≥52 years),diabetes duration (44 years), emergency room visits, comorbidity ofhypertension, and stroke in both PC and UC groups. Conclusion:Addition of PC to UC improved the quality of life in patients with type2 diabetes.Keywords: HRQOL, patients with diabetes, pharmaceutical careintervention, quality of life, usual care.


Introduction

Chronic medical conditions can impact multiple dimensions ofhealth-related quality of life (HRQOL) [1]. Given that diabetes ispart of a metabolic syndrome that increases the risk of heartdisease and stroke [2], it is not uncommon for these conditions tooccur as comorbidities in individuals with diabetes. Becausecomorbidities are prevalent in diabetes, it is unlikely that theHRQOL deficits associated with diabetes would be limited to thecondition itself. Indeed, the presence and severity of complica-tions or comorbidities have been associated with depression,anxiety, and impairment on multiple dimensions of HRQOL indiabetes [3]. The presence of cardiovascular complications ascomorbidity with diabetes also leads to deficit in HRQOL [4].

The national standardized prevalence rate of diabetes mellitusin Nigeria is 2.2%, while the crude prevalence rate is 74% in thoseaged 45 years and above who live in urban areas [5]. Globalestimates of the prevalence of diabetes for 2010 and 2030 showedthat the prevalence of diabetes in Nigeria in 2010 was 4.7% andthat it would be 5.5% in 2030 when compared with worldpopulation [6]. The complex nature of diabetes managementprompted the Nigerian Ministry of Health to come up with a

standard treatment guideline to streamline the process of diabetesmanagement and what service the patients should receive [7].

Several research studies have been carried out on health deficitassociated with diabetes comorbidities. For instance, the workdone by Maddigan et al. [8] to assess the impact of comorbid heartdisease, stroke, and arthritis on HRQOL in people with diabetes inthe general Canadian population concluded that “The illnessburden experienced by individuals with diabetes is not onlyassociated with diabetes itself, but largely with co-morbid medicalconditions.” Also, Westaway [9] reported that chronic diseasestatus and comorbidities were more important determinants ofhealth and well-being than were ethnicity, age, language, gender,and marital status. Quality of life (QOL) is also increasinglyrecognized as an important health outcome in its own right,representing the ultimate goal of all health interventions [10].

The health utilities index Mark 3 (HUI3) classification systemcomprises eight attributes: vision, hearing, speech, ambulation,dexterity, emotion, cognition, and pain—each with five or sixlevels of ability/disability. Most of these attributes can be neg-atively affected by diabetes and its complications.

Pharmaceutical care (PC) is the direct, responsible provision ofmedication-related care with the purpose of achieving definite





E-mails: [email protected]; [email protected].

� Address correspondence to: Maxwell O. Adibe, Department of Clinical Pharmacy and Pharmacy Management, University of Nigeria,Nsukka, Enugu, Nigeria.


dx.doi.org/10.1016/j.vhri.2013.06.007

dx.doi.org/10.1016/j.vhri.2013.06.007

dx.doi.org/10.1016/j.vhri.2013.06.007



dx.doi.org/10.1016/j.vhri.2013.06.007

outcomes that improve a patient’s QOL [10]. The principalelements of PC are that it is medication related; it is care that isdirectly provided to the patient; it is provided to produce definiteoutcomes; these outcomes are intended to improve the patient’sQOL; and the provider accepts personal responsibility for theoutcomes [10]. It is also the determination of the drug needs for agiven individual and the provision of not only the required drugbut also the necessary services (before, during, or after treatment)to ensure the optimally safe and effective drug therapy [11].

Diabetes is a disease that desperately needs more pharmacistinvolvement. Pharmacists who are specialized in this growingchronic condition can make a significant, positive impact on thepatient, the health care system, and themselves [12]. Health careprofessionals are becoming increasingly aware of the need toassess and monitor the QOL as an important outcome of diabetescare. The QOL is an important outcome in its own right andbecause it may influence the patient’s self-care activities, whichmay consequently have an impact on the diabetes control [13].Many PC programs have been established in various countries toenhance clinical outcomes and the HRQOL. These programs wereimplemented by pharmacists, with the cooperation of physiciansand other health care professionals. PC and the expanded role ofpharmacists are associated with many positive diabetes-relatedoutcomes, including improved clinical measures [14], improvedpatient and provider satisfaction [15,16], and improved cost ofmanagement [15,17]. The pharmacist can, therefore, in collabo-ration with physicians and other health care professionals, con-tribute to an improvement in the QOL of patients with diabetes byinforming and educating patients, answering their questions, and,at the same time, monitoring the outcomes of their treatment[18]. Such interventions, however, are not very common inNigeria. Therefore, this research was aimed at evaluating theimpact of the PC intervention on the QOL of patients with type 2diabetes mellitus in a tertiary hospital setting in Nigeria.

Methods

Research Design

This study was a randomized, controlled, and longitudinal pro-spective study with a 12-month patient follow-up. The studyprotocol was approved by the Research Ethical Committees of theUniversity of Nigeria Teaching Hospital, ItukuOzalla, and NnamdiAzikiwe University Teaching Hospital, Nnewi, in which this studywas conducted. These hospitals are tertiary hospitals that serveas referral centers to most of the hospitals in Nigeria.

Patients with type 2 diabetes mellitus who fulfilled theentrance criteria were identified and included in the study. Theinclusion criteria were as follows:

1. patients who were diagnosed with type 2 diabetes mellitus,2. patients with type 2 diabetes who were receiving oral hypo-

glycemic and/or insulin therapy,3. patients who provided written informed consent,4. patients who expressed willingness to abide by the rules of

the study, and5. patients who were certified fit for the study by their consulting

doctors.

Exclusion criteria were as follows:

1. patients who were diagnosed with type 1 diabetes (to avoidcomplexity in the study scope),

2. patients who were younger than 18 years (they are legallyregarded as dependents and consequently they cannot takedecisions of their own),

3. patients who were pregnant (they are generally not allowed toparticipate in a study of this nature by the institutions usedfor the study), and

4. patients who expressed willingness to withdraw from thestudy (participation is voluntary).

These criteria were according to the guiding principles of theinstitutional review boards of the hospitals used in this study.Following sample size determination, a sample size of at least 104patients in each of the control and intervention groups wasrequired. Based on these data, to ensure sufficient statisticalpower and to account for “dropouts” during the study, a targetsample size of 220 patients was recruited (110 patients from eachof the hospitals). The folders of the 110 selected patients in eachhospital were assigned numbers 1 to 110, which represented anindividual patient, and patients were randomly assigned to oneof two groups (intervention group or control group) on the basisof the number on their folders by using online “random sequencegenerator” [19] with sequence boundaries of 1 to 110 (boundariesinclusive) set in a two-column format: the first column was prioridesignated to the intervention group (55 patients) and the secondcolumn to the control group (55 patients).

Patients in the usual care (UC) group received the usual/conventional care offered by the hospitals, which includedhospital visits on appointment or on a sick day, consultationswith doctors, prescription of drugs and routine laboratory tests,review of diagnosis and medications, refilling of prescriptions bypatients, and referral. This UC was offered with no education/training of the patients on their diseases and drugs and withoutempowerment of the patients to be fully involved in the self-management of their illnesses. Patients in the PC group receivedUC and PC for 12 months. This additional PC included a stepwiseapproach: setting priorities for patient care, assessing patients’specific educational needs and identification of drug-relatedproblems, development of a comprehensive and achievable PCplan in collaboration with the patient and the doctor, implemen-tation of this plan, and monitoring and review of the plan fromtime to time [10]. The nurses collaborated with the pharmacist interms of organizing the patients and patients’ folders, takingpoint-of-care testing, counseling the patients, and reinforcing theinformation given to the patients during training sections. Thephysicians provided the visitation/appointment schedule for thepatients, and prescription of laboratory tests. They were alsoinvolved in the implementation of consensus strategies in man-aging drug-related problems in areas of changing, substitution,and withdrawal of medications.

The educational/training program for the patients consistedof four sessions of 90 to 120 minutes. The program covered thefollowing areas: diabetes overview and its complications, self-monitoring blood glucose techniques and interpretation ofdiabetes-related tests, medications and their side effects, lifestylemodification, counseling, and effective interaction with healthproviders. PC provided ground for the patients to monitor andreact to changes in their blood glucose levels, allowing them tointegrate their diabetes into the lifestyle they preferred.

Data Collection

The HUI23S4EN.40Q (developed by HUInc - Mark index 2&3)questionnaire was used to assess the HRQOL of the patients.HUI23S4EN.40Q questionnaires were interviewer-administered tothe patients in the intervention group and the control group atbaseline, 6 months, and 12 months.

The HUI3 classification system comprises eight attributes. Itdefines 972,000 unique health states. Single-attribute scores ofmorbidity are defined on a scale such that the worst level has ascore of 0.00 and the best level has a score of 1.00. Multiattribute


utility functions convert comprehensive health state descriptions(i.e., vectors of one level for each attribute defined by a HUIclassification system) into preference measures of overall HRQOL.The multiattribute scales of overall HRQOL are defined such thatthe score for dead is 0.00 and the score for perfect health is 1.00.Both HUI2 and HUI3 allow for negative scores of HRQOL thatrepresent health states considered worse than dead. The lowestpossible HRQOL scores are −0.03 for HUI2 and −0.36 for HUI3 [20].Also collected were data on patients’ demographics character-istics, lifestyles, and medical conditions, as outlined in Table 1.

Because we used two hospitals, we initially made comparisonsof the groups (UC and PC) across the hospitals to determine theirsimilarity, or, more specifically, to uncover any problems related toselection, history, or maturation effects. If the groups were found tobe essentially similar in these respects, we planned to combine thegroups for baseline, 6-month, and 12-month assessments of theeffects of PC. If major differences were identified, we planned toanalyze and report the group findings separately [21].

Data Analysis

Statistical analyses were performed by using the SPSS version 16.An intention-to-treat approach was used. Two-sample compari-sons were made by using Student’s t tests for normally distributedvariables or Mann-Whitney U tests for nonnormally distributeddata. Comparisons of proportions were done by using chi-square,Fisher’s exact, or McNemar’s tests. The differences in PC and UCwere assessed at baseline, 6 months, and 12 months. An a priorisignificance level of P less than 0.05 was used throughout.

Results

The medical and educational content of the training course wasrated positively by the 17 doctors and 29 nurses: the majority 38(82.6%) rated the content as “excellent” and the remaining 8 ratedthe content as “very good” or “good”; only 3 (6.5%) of them

Table 1 – Baseline characteristics of the patients in PC and UC arms.

Demographic data UC (n ¼ 110) PC (n ¼ 110) P

Mean age � SD (y) 52.8 � 8.2 52.4 � 7.6 0.708Grouped age: 4 53 y, n (%) 81 (73.64) 75 (68.18) 0.373Sex: Male, n (%) 49 (44.55) 44 (40) 0.495Level of education, n (%) 0.406Primary school 3 (2.72) 6 (5.45)Secondary school 71 (64.55) 63 (57.27)University 36 (32.73) 41 (37.27)

Marital status, n (%) 0.409Currently married 37 (33.64) 46 (41.82)Widowed 71 (64.54) 63 (57.27)Single 2 (1.82) 1 (0.91)

Occupation, n (%) 0.611Self-employed 37 (33.64) 34 (30.91)Employee 35 (31.82) 42 (38.18)Retired 38 (34.54) 34 (30.91)

Smoking status: Smoker, n (%) 34 (30.91) 21 (19.09) 0.043�

Duration, mean � SD 4.5 � 2.2 4.8 � 2.8 0.378Duration: ≥5 y, n (%) 62 (56.36) 71 (64.55) 0.215Family history of diabetes, n (%) 71 (64.55) 62 (56.36) 0.214Physical activity/exercise, n (%) 18 (16.36) 23 (20.91) 0.387Comorbidities, n (%)Hypertension 60 (54.55) 73 (66.36) 0.073Congestive heart failure 11 (10.00) 15 (13.64) 0.404Ischemic heart disease 7 (6.36) 8 (7.27) 0.789Arthritis 37 (33.64) 43 (39.09) 0.400≥2 comorbidities 72 (65.45) 81 (73.64) 0.187

Overnight hospitalization, n (%) 9 (8.18) 7 (6.36) 0.604Emergency room, n (%) 1(0.91) 2 (1.82) 0.561Use of insulin, n (%) 17 (15.45) 13 (11.82) 0.432Oral antidiabetic medication, n (%) 103 (93.64) 107 (97.27) 0.1954Other medications, n (%)Daily aspirin 43 (39.09) 57 (51.82) 0.058Diuretics 71 (64.55) 84 (76.36) 0.055Antihypertensives 98 (89.91) 78 (70.91) 0.0007�

Lipid-lowering 23 (20.91) 14 (12.73) 0.105Complications, n (%)Myocardial infarction 2 (1.82) 4 (3.64) 0.408Stroke 9 (8.18) 6 (5.45) 0.422Foot ulcer 2 (1.82) 3 (2.73) 0.651Blindness 1 (0.91) 1 (0.91) 1.000Renal failure 3 (2.73) 8 (7.27) 0.122

PC, pharmaceutical care; UC, usual care.� P ≤ 0.05


suggested little modification or changes. With the exception ofthe number of participants taking hypertensive drugs and smok-ing, we found no other significant differences at baseline in bothUC and PC groups. The number of patients who completed thestudy and whose data were analyzed at 6 months and 12 monthsin UC and PC groups were 98 (8.09%) versus 102 (92.73%) and 93(84.55%) versus 99 (90.0%), respectively (Table 1).

The overall HRQOL of the patients was significantly improved at 6months and 12 months in the PC arm when compared with the UCarm (0.79 � 0.07 vs. 0.65 � 0.05; P o 0.0001 and 0.86 � 0.12 vs. 0.64 �

0.10; P o 0.0001, respectively). The following single attributes weresignificantly improved in the PC arm over the UC arm at 6 months:vision (0.94 � 0.10 vs. 0.79 � 0.09; P o 0.0001), dexterity (0.92 � 0.09vs. 0.84 � 0.15; P o 0.0001), cognition (0.96 � 0.14 vs. 0.88 � 0.11; P o0.0001), and pain (0.82 � 0.15 vs. 0.70 � 0.14; P o 0.0001). There wasno significant improvement in hearing, speech, ambulation, andemotion attributes. At 12 months, there were significant improve-ments in all the single attributes except hearing (Table 2). Increasingage (≥52 years, the overall mean age) had high negative impact(deficit ≥10 %) on overall HRQOL in both PC and UC. A greaterpercentage of patients had ages greater than the overall mean age ofabout 52.6 years; this is reflected in more than one third of thepatients being retirees. Patients older than 52 years, that is, olderpatients, had clinically significant lower QOL than did youngerpatients in both UC and PC groups. A majority of the patients hadsecondary education, and about a third were self-employed. Familyhistory of diabetes was reported by a majority of the patients in bothUC and PC groups. Those without a family history of diabetes hadhigher QOL in both groups. This result was consistent with what wasexpected, but it should be understood that some of the patients didnot know their families’ full medical history in detail. Diabetesduration (44 years), emergency room visits, and comorbidity ofhypertension and stroke had a high negative impact (deficit ≥10%) onthe overall HRQOL in both PC and UC groups (Table 3). These factorswere also associated with a low single-attribute score recorded inboth arms of the study (Table 4). The changes in the overall HRQOLscore and single-attribute scores of patients in the PC group werehigher than in the UC group, which showed that the PC interventionhad an overriding effect over the patients’ characteristics than didUC. The percentage changes (improvements) in the PC group werehigher than in the UC group (Table 3). Addition of PC to UC resultedin a significant gain in quality-adjusted life-years (QALYs) (PC vs. UC)(0.7625 � 0.15 vs. 0.6425 � 0.13; P o 0.0001), with 0.12 (0.07 to 0.1601;95% confidence interval) QALY gained at the end of 12 months.

Discussion

Impact of the PC Intervention on the HUI3 Overall HRQOL

The overall HRQOL of the patients was significantly improved at 6months and 12 months in the PC arm when compared with theUC arm. All the single attributes were significantly improved inthe PC arm when compared with the UC arm at 12 months excepthearing. The PC intervention impacted positively on patientsassigned to it, and this change was both clinically (difference≥0.03) and statistically significant [22,23]. PC interventions canhave a positive change on the overall HRQOL and single attrib-utes, that is, how patients with diabetes are able to cope withdaily activities [24–26]. The improvement in the HRQOL may inpart be attributed to the increased contact of patients withdiabetes with the clinical pharmacist, but it is also likely to beassociated with improved adherence to lifestyle advice.

The results of this intervention were similar to that of aprospective study on the impact of PC on QOL in patients withtype 2 diabetes mellitus that was conducted in a private tertiarycare teaching hospital in South India for a period of 8 months,

Table

2–Changesin

overa

llHRQOLand

single

attribute

sbetw

een

thePC

and

UC

gro

upsafter12-m

ofo

llow-u

p.

Outcom

esBas

eline

6th

mon

th12

thmon

th12

-moch

ange

P

UC(n

¼11

0)PC

(n¼

110)

UC(n

¼98

)PC

(n¼

102)

UC(n

¼93

)PC

(n¼

99)

PCminusUC

(95%

CI)

Bas

eline

6mo

12mo

Ove

rall

HRQOL

0.63

�0.04

0.61

�0.08

0.65

�0.05

0.79

�0.07

0.64

�0.10

0.86

�0.12

0.22

(0.188

4–0.25

16)

0.10

29o0.00

01�

o0.00

01�

Vision

0.81

�0.10

0.77

�0.11

0.79

�0.09

0.94

�0.10

0.83

�0.10

0.97

�0.08

0.14

(0.114

3–0.16

57)

0.00

5�o0.00

01�

o0.00

01�

Hea

ring

0.90

�0.13

0.86

�0.21

0.91

�0.18

0.96

�0.19

0.92

�0.12

0.97

�0.23

0.05

(−0.00

27to

0.10

27)

0.09

080.05

770.06

29Sp

eech

0.92

�0.21

0.89

�0.12

0.92

�0.09

0.94

�0.12

0.91

�0.08

0.95

�0.09

0.04

(0.015

7–0.06

43)

0.19

470.18

530.00

14�

Ambu

lation

0.82

�0.27

0.81

�0.19

0.83

�0.16

0.88

�0.29

0.84

�0.07

0.94

�0.10

0.10

(0.075

3–0.12

47)

0.75

100.15

26o0.00

01�

Dex

terity

0.79

�0.19

0.81

�0.12

0.84

�0.15

0.92

�0.09

0.80

�0.10

0.95

�0.22

0.15

(0.100

8–0.19

92)

0.35

16o0.00

01�

o0.00

01�

Emotion

0.87

�0.12

0.85

�0.15

0.86

�0.20

0.90

�0.08

0.86

�0.15

0.96

�0.10

0.10

(0.063

9–0.13

61)

0.27

600.06

29o0.00

01�

Cogn

ition

0.89

�0.12

0.91

�0.18

0.88

�0.11

0.96

�0.14

0.85

�0.08

0.98

�0.12

0.13

(0.100

8–0.15

92)

0.33

33o0.00

01�

o0.00

01�

Pain

0.67

�0.12

0.71

�0.10

0.70

�0.14

0.82

�0.15

0.68

�0.15

0.84

�0.10

0.16

(0.123

9–0.19

61)

0.00

8�o0.00

01�

o0.00

01�

CI,co

nfiden

ceinterval;HRQOL,

hea

lth-related

qualityoflife;PC

,pharmac

eutica

lca

re;UC,usu

alca

re.

�P≤

0.05

.


Table 3 – Impact of patients’ characteristics on overall utility score of the patients.

Patients’ characteristicsOverall HRQOL score after 12 mo (with characteristics minus without characteristics)

Change in overall HRQOL score after12 mo (PC minus UC)

UC PC Deficit� Change† (%)

With condition Without condition With condition Without condition UC PC UC PC

SociodemographicAge (≥52 y) 0.54 0.62 0.61 0.73 −0.08 −0.12 0.07 (13.0) 0.11 (17.7)Family history of diabetes 0.59 0.65 0.67 0.72 −0.06 −0.05 0.08 (13.6) 0.07 (10.8)Smoking status 0.58 0.62 0.64 0.72 −0.04 −0.08 0.06 (10.3) 0.10 (16.1)

SeverityUse of insulin and oral medications 0.51 0.63‡ 0.56 0.74‡ −0.12 −0.18 0.05 (9.80) 0.11 (17.5)Use of oral medication only 0.58 0.63‡ 0.65 0.74‡ −0.05 −0.09 0.07 (12.1) 0.11 (17.5)Diabetes duration 44 y 0.56 0.62 0.67 0.74 −0.06 −0.07 0.11 (19.6) 0.12 (19.4)

Resource utilizationOvernight hospitalization 0.58 0.60 0.62 0.73 −0.02 −0.11 0.04 (6.9) 0.13 (21.7)Contact with physician in ER 0.59 0.61 0.68 0.76 −0.02 −0.08 0.09 (15.3) 0.15 (24.6)Doctor visit more than 12 times 0.61 0.67 0.76 0.79 −0.06 −0.03 0.15 (24.6) 0.12 (17.9)

ComorbidityHypertension 0.62 0.64 0.74 0.78 −0.02 −0.04 0.12 (19.4) 0.14 (21.9)Stroke 0.42 0.59 0.57 0.82 −0.17 −0.25 0.15 (35.7) 0.23 (39.0)Eye problems 0.63 0.66 0.69 0.73 −0.03 −0.04 0.06 (9.5) 0.07 (10.6)Number of medical conditions ≥2 0.58 0.67 0.63 0.71 −0.09 −0.08 0.05 (8.6) 0.04 (6.0)

Note. Mean difference is the actual deficit associated with a particular condition (comorbidity, severity, and resource utilization). Negative values (−) indicate that the characteristic impactednegatively (lower utility scores) on the patients. ER, emergency room; PC, pharmaceutical care; UC, usual care.� Deficit, utility scores of patients with the characteristic minus utility scores of patients without the characteristics within the UC or PC group.† Change, utility scores of patients in the PC group minus utility scores of patients in the UC group. “Deficit” and “change” values ≥0.03 or ≤0.03 are clinically significant.‡ Noninsulin nonantidiabetes medications (NINM).

VALUE

INH

EALTH

REGIO

NAL

ISSUES

2(2

013)240–247

244

which concluded that the PC program was effective in improvingthe clinical outcome and QOL of patients with type 2 diabetesmellitus [27]. Also, another 1-year observational study thatevaluated the QOL in patients at the Medical University of SouthCarolina Family Medicine Clinic, who were followed by a clinicalpharmacist diabetes educator, showed that patients rated theirQOL high after the follow-up [28]. A 1-year study by Correr et al.[29] concluded that pharmacotherapy follow-up of patients withtype 2 diabetes in community pharmacies can improve theHRQOL and satisfaction of patients.

Impact of Patients’ Characteristics on Overall HRQOL

The health deficit imposed by old age cut across both groups, butthe patients in the PC group has a significantly higher QOL thandid those in the UC group, which showed that the interventionhad an overriding effect over the effect of age on QOL. Loweducation and unemployment are associated with low income,which consequently affects the QOL of patients. This result isconsistent with a study carried out in a Nigeria teaching hospital

by Issa and Baiyewu [30] who concluded that lower income, loweducation level, and low-rated employment affect the QOL ofNigerian patients with diabetes. The overall HRQOL scores ofnonsmokers in both UC and PC groups were greater than scores ofthe smokers. This result was consistent with what was expectedbut could be attributed to the fact that the PC group had morefemale patients and there is sociocultural stigma associated withfemales smoking in Nigeria. The comorbidities considered in thestudy were eye problem, heart disease, and stroke. The respond-ents without eye problem had clinically significant higher QOLthan did those with eye problem, in both UC and PC groups. Thiseye problem could be attributed to diabetes retinopathy, which isa complication of diabetes. This is consistent with a studyperformed in Canada [8] that concluded that the illness burdenexperienced by individuals with diabetes was associated not onlywith diabetes itself but largely with the comorbid medical con-dition. Lloyd et al. [31] concluded that the presence of even milddiabetes complications had a significant impact on patients’ QOL.Early diagnosis and treatment is essential to help prevent thedeterioration in the HRQOL of these patients.

Table 4 – Impact of patients’ characteristics on single-attribute utility score for HUI3.

Single attributes PC Mean difference PC Mean difference PC Mean difference

Sociodemographic Smoking status Family history of diabetes Age (≥52 y)Vision 0.8681 0.0707* 0.7977 0.0302 0.9295 0.1686*Hearing 0.9013 −0.0380 0.9270 0.0063 0.9670 0.0534*Speech 0.9962 0.0303 0.9671 0.0110 0.9890 0.0234Ambulation 0.8928 0.0867* 0.8126 0.0257 0.9465 0.1752*Dexterity 0.8189 0.0433 0.8198 0.0629* 0.8494 0.0928*Emotion 0.9028 −0.0378 0.8746 0.0014 0.9154 0.0607*Cognition 0.8157 0.0935* 0.7215 0.0420 0.8606 0.1699*Pain 0.7096 0.1008* 0.6283 0.0077 0.7389 0.1554*

Resource utilization Overnight hospitalization Contact with physician in ER Physician visit ≥ 12 visitsVision 0.8302 0.0315 0.8216 0.0445 0.8331 0.0487Hearing 0.8522 0.0423 0.9378 0.0437 0.9471 0.0419Speech 0.9861 0.0254 0.9825 0.0557* 0.9830 0.0249Ambulation 0.8909 0.1251* 0.8688 0.2488* 0.8892 0.1541*Dexterity 0.7996 0.0268 0.8199 0.2005* 0.8475 0.1948*Emotion 0.8948 0.0407 0.8884 0.0855* 0.8648 �0.0217Cognition 0.8035 0.1152* 0.7933 0.2886* 0.8254 0.199*Pain 0.7059 0.1426* 0.6799 0.279* 0.7522 0.2923*

Severity Use of insulin Use of medication Duration of diabetes 4 4 yVision 0.8319 0.0377 0.9750 0.1695* 0.8653 0.1278*Hearing 0.9413 0.023 1.0000 0.0733* 0.8478 0.0436Speech 0.9886 0.0328 1.0000 0.0284 0.9911 0.0448Ambulation 0.9010 0.1568* 0.9670 0.1479* 0.9119 0.2098*Dexterity 0.8531 0.1417* 0.9880 0.2128* 0.8403 0.1266*Emotion 0.8858 0.0252 0.9280 0.057* 0.8956 0.055*Cognition 0.8179 0.1552* 0.9760 0.244* 0.8424 0.2444*Pain 0.7542 0.2563* 0.8610 0.2406* 0.7589 0.3029*

Comorbidity Eye problem Heart disease StrokeVision 0.9047 0.1909* 0.9190 0.1228* 0.8298 0.1091*Hearing 0.9269 0.0074 0.9462 0.0185 0.9399 0.0654*Speech 0.9914 0.0388 0.9300 0.0316 0.9849 0.0821*Ambulation 0.9171 0.1908* 0.9190 0.1082* 0.8883 0.4262*Dexterity 0.8790 0.1962* 0.8138 0.0327 0.8492 0.4371*Emotion 0.9024 0.0601* 0.9521 0.0915* 0.8932 0.1332*Cognition 0.8400 0.2017* 0.8352 0.1064* 0.7964 0.3598*Pain 0.7591 0.2667* 0.7383 0.1238* 0.7158 0.5748*

Note. Mean difference represents the single-attribute utility score for patients in the PC group minus patients in the UC group. Positive value(þ) indicates high single-attribute utility score for the PC group (i.e., HRQOL gained).ER, emergency room; HRQOL, health-related quality of life; HUI3, health utilities index Mark 3; PC, pharmaceutical care; UC, usual care.� Clinically significant for HUI3 single-attribute utility score (value ≥0.05).


A majority of the patients in this study had comorbidities.Stroke imposed the greatest HRQOL deficit among all othercomorbidities in both PC and UC groups but more in the UCgroup. The HRQOL deficit associated with stroke and hypertensionhad been reported by Ekwunife et al. [32]. Their study demon-strated that patients with hypertension alone had a higher overallmean utility score while hypertensive patients with stroke had thelowest overall mean score. Other similar studies [2,9,33] revealedthat stroke and other comorbidities can impose considerablehealth deficits on patients and concluded that the illness burdenexperienced by individuals with diabetes is associated not onlywith diabetes itself but largely with comorbid medical conditions.The results of this study also conform to a study that found thatthe presence of cardiovascular complications as comorbidity withdiabetes also leads to deficit in HRQOL [3]. This emphasized theconsiderable public health impact that all these chronic condi-tions have on the HRQOL, particularly when they occur together.

This study has shown that diabetes complications have aprofound effect on the HRQOL of patients with type 2 diabetes.Even the presence of mild diabetes complications has a significantimpact on the HRQOL. In this study, the following four parameterswere used to assess the severity of the disease: use of insulin anddiabetes medications, use of diabetes medication only, duration ofdiabetes, and number of absenteeism fromwork or school in the last1 year. Those who use a combination of insulin and oral antidia-betes medications and only oral antidiabetes medication had severehealth deficits when compared with patients using noninsulinnonmedications in both UC and PC groups. This is consistent withthe results of a study by Rubin [34] who assessed the feelings andcomplaints made by patients with diabetes using insulin and foundthat these could impact negatively on their QOL. This shows that asthe disease progresses to a more severe situation, treatment optionchanges from diet to the use of oral medication only and deterio-rates to the use of insulin in combination with oral medication.

This study revealed that those with a duration of diabetes ofmore than 4 years also had a considerable health deficit whencompared with patients with a duration of diabetes of 4 years or lessin both UC and PC groups. This result is consistent with a report fromthe American Diabetes Association [35] that stated that the longerthe duration of diabetes the higher the chances of a patientdeveloping overt nephropathy, which, in turn, lowers the HRQOL ofthe patients. To improve the HRQOL of patients with type 2 diabetes,early diagnosis of the disease and aggressive management of riskfactors are necessary to prevent or delay the development of diabetescomplications. Resource utilization was assessed as overnight hos-pitalization, contact with doctor in emergency room, and number ofdoctors’ visit in the past 1 year. These resources included physicians,pharmacists, and nurses’ time, hospital bed space, and spent moremoney. There was a considerable health deficit associated withresource utilization factors in both UC and PC groups. Patients whohad overnight hospitalization and contact with physician or nurse inemergency room in the past 12 months had more health deficitwhen compared with patients who were not in their categories inboth UC and PC groups.

Despite the enormous burdens imposed by the above-outlinedpatients’ characteristics on the overall HRQOL, the patients in thePC group had significantly higher overall HRQOL scores than didtheir counterparts in the UC group, implying that the PC inter-vention had an overriding effect over the debilitating powers ofthe patients’ characteristics.

Impact of Patients’ Characteristics on the HUI3 SingleAttributes

Comorbid stroke and old age (≥52 years) were associated with arelatively large mean difference between PC and UC groups in allsingle attributes. This indicates that patients in the PC intervention

who were 52 years or older and had stroke benefited most in alltheir single-attributes domains. Patients with diabetes in the UCgroup who were with comorbid medical conditions had lowerutility scores on the anticipated diabetes-relevant single attributesthan did their counterparts in the PC group, which resulted in alarge mean difference. Based on the literature, it was anticipatedthat diabetes would affect the vision, dexterity, ambulation, emo-tion, and pain attributes of the HUI3 [4,36,37]. The considerableburden associated with diabetes and its comorbidities on thesespecific single attributes was therefore not surprising. Associationof all sociodemographic, severity, comorbiditiy, and resource uti-lization factors, except family history, with the “pain” attribute wasclinically and statistically significant. Ischemic heart disease,angina, and congestive heart failure are all associated with pain[38,39]. This is consistent with a study conducted by Ekwunife et al.[32] that demonstrated that the pain attribute had the lowesthealth state utility score among Nigerian patients with hyper-tension alone, hypertension and heart failure, hypertension andcoronary heart disease, and hypertension and stroke. For the painattribute in this study, the utility score recorded by the patientswith stroke in the PC group far exceeded the score in the UC groupafter 12 months, resulting in a large mean difference. This meansthat the PC intervention was able to counter and override theutility-reducing power of comorbidities within 12 months. It ispossible that some of the pain and discomfort (low score) found inthe UC group was not attributable to stroke, but rather to anotherpainful comorbid condition associated with diabetes and heartdiseases, such as peripheral vascular disease [40,41].

Functioning on the “ambulation” attribute was similar to that ofpain. All the sociodemographic, severity, comorbidity, and resourceutilization factors were significantly associated with ambulationclinically. This means that the ambulation attribute of patients inthe PC intervention group was highly improved compared withthat of patients in the UC group. Also, the PC intervention hadoverriding benefits over all these factors that impair the QOL. Theburden on the ambulation attribute associated with diabetes,comorbidities, and other factors was quite visible, not peculiar toone but all the factors. Clinically important burden on the emotionattribute was apparent for comorbidities and older patients. Thelargest emotional benefits of the PC intervention were observed inpatients with diabetes with stroke compared with the UC group.These findings are consistent with a report of Maddigan et al. [8]that demonstrated that diabetes and its comorbidities could impactnegatively on the overall utility state and single attributes.

QALYs associated with PC were significantly higher than thoseof UC. This indicates that extending this study beyond 1 year couldoffer more benefits to patients with diabetes in terms of QALYsgained. Some studies had demonstrated that the extension of PCbeyond 1 year could offer extra benefits to patients with diabetes[42,43]. This study lends support to the use of the HUI3 for theevaluation of HRQOL as an outcome in PC intervention programsfor patients with diabetes, as suggested by Feeny et al. [20].

Although the patients’ characteristics affected the singleattributes of the patients’ HRQOL adversely, the interventionnot only had overriding effects of these patients’ characteristicsthat lower the overall HRQOL but also improved the singleattributes of the QOL of these patients.

Limitations

Our study was subject to the following limitations, and the resultswere interpreted in this light: Selection bias was a threat becauseparticipation was voluntary though the groups were randomized.It remains possible that patients who chose to participate in theprogram may have differed in some important way from thosewho did not participate, which could pose a threat to externalvalidity or generalizability. Given the difficulty in blinding in this


type of study, the clinicians involved in providing care and theinterviewers administering the HUI3 questionnaire were notblinded to patient allocation. Recall bias was possible in thisstudy because HUI3 was a 4-week recall questionnaire. Attritionbias or loss during follow-up was also a serious threat but wasavoided by using an intention-to-treat design. The data were self-reported; however, self-reported data about diabetes status havebeen established to be both valid and reliable [44].

Conclusions

The addition of PC to UC resulted in improving the QOL of patientswith type 2 diabetes. The overall HRQOL of the patients wassignificantly improved at 6 months and 12 months in the PC armwhen compared with the UC arm. Vision, cognition, and pain singleattributes were significantly improved in the PC arm over the UCarm after 6 months and all single attributes except hearing weresignificantly improved after 12 months. Stroke imposed the greatestHRQOL deficit among all other patients’ characteristics in both PCand UC arms but more in the UC arm. The results of this studyillustrate a convincing rationale for improving standards of care forpatients with type 2 diabetes through the PC intervention. Furtherresearch, however, is needed to improve on the current PCintervention strategies such that the recorded improvements inHRQOL will be sustained for a very long time after an intervention.

Acknowledgment

We acknowledge Health Utility, Inc., for granting and awarding usHUI23S4En.40Q and HUI23.40Q.MNL used in this study.

Source of financial support: Funding for this project wasprovided by Science and Technology Education Post Basic(STEP-B) through the University of Nigeria. The views expressedin this article are those of the authors, and no official endorse-ment by STEP-B is intended or should be inferred.

R E F E R E N C E S

[1] Stewart M, Brown J, Boon H, et al. Evidence on patient-doctorcommunication. Cancer Prev Control 1999;3:25–30.

[2] Beckman J, Creager M, Libby P. Diabetes and atherosclerosis: epidemiology,pathophysiology, and management. JAMA 2002;287:2570–81.

[3] Peyrot M, Rubin R. Levels and risks of depression and anxietysymptomatology among diabetic adults. Diabetes Care 1997;20:585–90.

[4] de-Visser C, Bibo H, Groenier K, et al. The influence of cardiovasculardisease on quality of life in type 2 diabetics. Qual Life Res2002;11:249–61.

[5] Nyenwe E, Odia O, Ihekwala A, et al. Type 2 diabetes in adult Nigerians:a study of its prevalent and risk factors in Port Harcourt, Nigeria.Diabetes Res Clin Pract 2003;62:177–85.

[6] Shaw J, Sicree R, Zimmet P. Global estimates of the prevalence ofdiabetes. Diabetes Res Clin Pract 2010;87:4–14.

[7] Federal Ministry of Health in collaboration with WHO, EC, DFID.Standard treatment guidelines. Nigeria 2008:90–7.

[8] Maddigan S, Feeny D, Johnson J. Health related quality of life deficitassociated with diabetes and comorbidities in a Canadian NationalPopulation Health Survey. Qual Life Res 2005;14:1311–20.

[9] Westaway M. Effects of ageing, chronic disease and co-morbidities onthe health and well being of older residents of Greater Tshwane. S AfrMed J 2010;100:1–3.

[10] Hepler C, Strand L. Opportunities and responsibilities inpharmaceutical care. Am J Hosp Pharm 1990;47:533–43.

[11] Polonsky WH. Understanding and assessing diabetes-specific quality oflife. Diabetes Spectr 2000;13:36–41.

[12] Davis TM, Clifford RM, Davis WA, Batty KT. The role of pharmaceuticalcare in diabetes management. Br J Diabetes Vasc Dis 2005;5:352–6.

[13] Khan CR, Weir GC, King GL, Moses AC. Joselin’s Diabetes Mellitus. (14thed.). Philadelphia: Lippincott Williams & Wilkins, 2004.

[14] Jaber L, Halapy H, Fenret M, et al. Evaluation of a pharmaceutical caremodel on diabetes management. Ann Pharmacother 1996;30:238–43.

[15] Sadur C, Moline N, Costa M, et al. Diabetes management in a healthmaintenance organization: efficacy of care management using clustervisits. Diabetes Care 1999;22:2011–7.

[16] Majumdar S, Guirguis L, Toth E, et al. Controlled trial of a multifacetedintervention for improving quality of care for rural patients with type 2diabetes. Diabetes Care 2003;26:3061–6.

[17] Coast-Senior E, Kroner B, Kelley C, Trili L. Management of patients withtype 2 diabetes by pharmacists in primary care clinics. AnnPharmacother 1998;32:636–41.

[18] Hawkins D, Bradberry JC, Cziraky MJ, et al. National PharmacyCardiovascular Council treatment guidelines for the management oftype 2 diabetes mellitus: toward better patient outcomes and new rolesfor pharmacists. Pharmacotherapy 2002;22:436–44.

[19] Mads-Haahr. Random Sequence Generator (1998–2011). Available from:http://www.random.org/sequences/. [Accessed February 15, 2012].

[20] Feeny D, Furlong W, Torrance G, et al. Multi-attribute and single-attribute utility functions for the health utilities index Mark 3 System.Med Care 2002;40:113–28.

[21] Cranor C, Christensen D. The Asheville Project: short-term outcomes ofa community pharmacy diabetes care program. J Am Pharm Assoc2003;43:149–59.

[22] Horsman J, Furlong W, Feeny D, Torrance G. The Health Utilities Index(HUI): concepts, measurements properties and applications. HealthQual Life Out 2003;1:54.

[23] Grootendorst P, Feeny D, Furlong W. Health Utilities Index Mark 3evidence of construct validity for stroke and arthritis in a populationhealth survey. Med Care 2000;38:290–9.

[24] Clifford R, Batty K, Davis T, et al. A randomised controlled trial of apharmaceutical care programme in high-risk diabetic patients in anoutpatient clinic. Int J Pharm Pract 2002;10:85–9.

[25] Clifford R, Batty K, Davis T, Davis W. Effect of a pharmaceutical careprogram on vascular risk factors in type 2 diabetes (The FremantleDiabetes Study). Diabetes Care 2005;28:771–6.

[26] Posey L. Proving that pharmaceutical care makes a difference incommunity pharmacy [editorial]. J Am Pharm Assoc 2003;43:136–9.

[27] Srirama S, Chack LE, Ramasamy R, et al. Impact of pharmaceutical careon quality of life in patients with type 2 diabetes mellitus. JRMS 2011;16(Special Issue):412–8: Available from: www.mui.ac.ir. [AccessedDecember 16, 2011].

[28] Jennings DL, Ragucci KR, Chumney ECG, Wessell AM. Impact of clinicalpharmacist intervention on diabetes related quality-of-life in anambulatory care clinic. Pharm Pract 2007;5:169–73.

[29] Correr CJ, Pontarolo R, Souza RAP, et al. Effect of a pharmaceutical careprogram on quality of life and satisfaction with pharmacy services inpatients with type 2 diabetes mellitus. BJPS 2009;45:809–17.

[30] Issa B, Baiyewu O. Quality of life of patients with diabetes mellitus in aNigerian teaching hospital. Hong Kong J Psychiatry 2006;16:27–33.

[31] Lloyd A, Sawyer W, Hopkinson P. Impact of long-term complications onquality of life in patients with type 2 diabetes not using insulin. ValueHealth 2001;4:392–400.

[32] Ekwunife OI, Aguwa C, Adibe M, et al. Health state utilities of a populationof Nigerian hypertensive patients. BMC Res Notes 2011;4:528.

[33] Sacco R, Boden-Albala B, Abel G. Race-ethnic disparities in the impactof stroke risk factors: the northern Manhattan stroke study. Stroke2001;32:1725–31.

[34] Rubin R. Diabetes and quality of life. Diabetes Spectr 2000;13:21.[35] ADA. Standards of medical care in diabetes [Position Statement].

Diabetes Spectr 2011;34(Suppl.):S11–61.[36] Maddigan S, Feeny D, Johnson J. A comparison of the Health Utilities

Indices Mark 2 and Mark 3 in type 2 diabetes. Med Decis Making2003;23:489–501.

[37] Maddigan S, Feeny D, Johnson J. Construct validity of the RAND-12 andHealth Utilities Index Mark 2 and Mark 3 in type 2 diabetes. Qual LifeRes 2004;13:435–48.

[38] Mayou R, Blackwood R, Bryant B, Garnham J. Cardiac failure: symptomsand functional status. J Psychosom Res 1991;35:399–407.

[39] Brown N, Melville M, Gray D, et al. Quality of life four years after acutemyocardial infarction: short form 36 scores compared with a normalpopulation. Heart 1999;81:352–8.

[40] Adler A, Stratton I, Neil H, et al. Association of systolic blood pressure withmacrovascular and microvascular complications of type 2 diabetes (UKPDS36): prospective observational study. BMJ 2000;321:412–9.

[41] Belch J, Topol E, Agnelli G, et al. Critical issues in peripheral arterialdisease detection and management: a call to action. Arch Intern Med2003;163:884–92.

[42] Cranor C, Bunting B, Christensen D. The Asheville Project: long-termclinical and economic outcomes of a community pharmacy diabetescare program. J Am Pharm Assoc 2003;43:173–84.

[43] Neto P, Marusic S, Júnior DP, et al. Effect of a 36-month pharmaceutical careprogram on coronary heart disease risk in elderly diabetic andhypertensive patients. J Pharm Pharm Sci 2011;14:249–63.

[44] West J, Goldberg K. Diabetes self-care knowledge among outpatients at aVeterans Affairs medical center. Am J Health-Syst Ph 2002;59:849–52.

















































http://www.random.org/sequences/























www.mui.ac.ir















































An Audit of Diabetes-Dependent Quality of Life (ADDQOL) in OlderPatients with Diabetes Mellitus Type 2 in SloveniaEva Turk, MA, MBA1,�, Valentina Prevolnik Rupel, PhD2, Alojz Tapajner, BSc3, Stephen Leyshon, RN, MSc1, Arja Isola, RN, PhD4

1Det Norske Veritas, Healthcare & Biorisk, DNV Research and Innovation, Høvik, Norway; 2Institute for Economic Research, Ljubljana, Slovenia; 3Faculty ofMedicine, University of Maribor, Maribor, Slovenia; 4Faculty of Medicine, Institute of Health Sciences, University of Oulu, Oulu, Finland

A B S T R A C T

Objective: This article reports a study to measure diabetes-dependentquality of life (QOL) in older Slovenian patients with diabetes mellitustype 2 (DMT2). Methods: A cross-sectional study of older (age ≥ 65years) patients with DMT2 at outpatient diabetic centers was con-ducted in all regions in Slovenia. The Audit of Diabetes-DependentQuality of Life questionnaire was carried out between January andMay 2012. Statistical analysis was performed by using IBM SPSSStatistics software, version 18.0. Results: After exclusion of noneli-gible respondents, a total of 285 respondents were included in theanalysis, which represented a 57% response rate. Lower QOL wassignificantly connected to a heart attack episode (odds ratio 2.42; 95%confidence interval 1.06–5.20) and to the perception of not havingdiabetes under control (odds ratio 0.36; 95% confidence interval 0.18–0.69). Eleven (3.9%) patients reported no impact of DMT2 on their QOL

at all, while in the remaining respondents, particular reference wasput to the effects on freedom to eat, dependency on others, and familylife. There was no significant difference between the older peopleliving in urban and rural areas. Conclusions: The findings of thepresent study highlight the impact of DMT2 on QOL. DMT2 imposes apersonal burden on individuals. Information on the QOL of olderpatients with diabetes is important to Slovenian policymakers andfamily physicians to identify and implement appropriate interven-tions for achieving better management of diabetes and ultimatelyimproving the QOL of patients with diabetes.Keywords: ADDQOL, DMT2, elderly, patient-reported outcomes, qualityof life.


Introduction

Diabetes is a chronic metabolic disease that can have a profoundimpact on the health status and quality of life (QOL) of patients interms of physical, social, and psychological well-being [1–3].Diabetes is now a global health concern: affecting both indus-trialized and transitioning countries. The number of people withdiabetes is increasing because of population growth, aging,urbanization, and increasing prevalence of obesity and physicalinactivity. Diabetes mellitus (DM) currently affects about 285million adults worldwide, and it is projected to rise to 366 millionin 2030 [4,5]. The most important demographic change to dia-betes prevalence across the world appears to be the increase inthe proportion of people aged 65 years or older [4]. In Europealone, more than 50 million individuals are affected by diabetes,90% of whom have diabetes mellitus type 2 (DMT2) [6].

Slovenia does not differ significantly from other EuropeanUnion countries with regard to the prevalence of diabetes. Theestimates of the National Institute of Public Health [7] amount toapproximately 125,000 patients with diabetes in Slovenia, whichis 6.3% of the population. Of these, 22.2% are aged 75 years orolder, and 16% are aged between 65 and 74 years, with the meanage of patients with DMT2 being 65 years. Similarly, as in the rest

of Europe, the population in Slovenia is ageing and populationhealth improvement is an increasingly important component ofcoordination and collaboration among patients and health careproviders [8,9].

Internationally, there has been a marked shift in thinkingabout what health is and how it is measured [10]. Traditionalclinical ways of measuring health and the effects of treatment areeither accompanied by or even replaced by patient-reportedoutcome measures (PROMs), which present an entirely subjectivereport of the status of a patient’s health condition. Research hasshown that patients with diabetes are more concerned aboutphysical and social function, emotional and mental health, aswell as the burden of illness and treatments on daily life thanwith clinical biomarkers such as hemoglobin A1c, blood pressure,or lipid levels [11,12]. PROMs are thus meaningful and relevantoutcomes. Furthermore, there is evidence that when the health-related quality of life (HRQOL) of individuals with diabetes isproperly measured and the results are incorporated into healthcare management, improvements in patient outcomes occur[13,14]. Improvements in glycemic control and QOL, as well asreduction in short-term complications including the incidence ofsevere hypoglycemia, can be observed in combination of treat-ment and education of patients [15–17].




Conflict of interest: The authors have indicated that they have no conflicts of interest with regard to the content of this article.


� Address correspondence to: Eva Turk, Healthcare Programme, DNV Research and Innovation, Veritasveien 1, 1322 Høvik, Norway.


dx.doi.org/10.1016/j.vhri.2013.05.001

dx.doi.org/10.1016/j.vhri.2013.05.001

dx.doi.org/10.1016/j.vhri.2013.05.001

mailto:[email protected]<?A3B2 show .?>

dx.doi.org/10.1016/j.vhri.2013.05.001

Many valid instruments to measure PROMs in diabetes havebeen developed and are already used in industrialized countries[18–20]. Among diabetes-dependent QOL measures, the Audit ofDiabetes-Dependent Quality of Life (ADDQOL) is a widely usedinstrument [21–24]. In Slovenia, however, despite the high prev-alence of diabetes, so far no studies evaluating patient-reportedoutcomes, such as HRQOL, have been conducted. According tothe literature, less research was conducted on how various riskfactors influence the QOL of patients with diabetes [25–27]. In thismanner, the objective of the current study was to measurediabetes-dependent QOL in the older Slovenian patients withDMT2 and to assess its relationships with sociodemographic andhealth factors.

Methods

Instrument

The ADDQOL consists of two overview items; one measuresgeneric overall QOL and a further 19 items are concerned withthe impact of diabetes on specific aspects of life. The 19 lifedomains are as follows: leisure activities, working life, local orlong-distance journeys, holidays, physical health, family life,friendships and social life, close personal relationships, sex life,physical appearance, self-confidence, motivation to achievethings, people’s reactions, feelings about the future, financialsituation, living conditions, dependence on others, freedom toeat, and freedom to drink. These 19 domains ask the respondentsto evaluate how their life would be if they did not have diabetes.

The scales range from −3 to þ1 for 19 life domains (impactrating) and from 0 to þ3 in attributed importance (importancerating). A weighted score for each domain is calculated as amultiplier of impact rating and importance rating (ranging from−9 to þ3). Lower scores reflect poorer QOL. Finally, a meanweighted impact score (ADDQOL score) is calculated for the entirescale across all applicable domains [21,23,28]. Apart from theperceived QOL, data on patients’ demographic characteristics,duration of diabetes, and existing diabetic complications weremeasured. The linguistic validation and cultural adaptation ofthe original English ADDQOL into Slovenian version is describedelsewhere (E. Turk, V. Prevolnik-Rupel, A. Tapajner, et al., unpub-lished data, 2013).

Study Design and Participants

A cross-sectional study was conducted between January and May2012 by using a structured questionnaire.

Patients from the 12 participating outpatient diabetic centerswere recruited by using the convenience sampling method. Theregions selected were defined by the Statistical office of theRepublic of Slovenia (E. Turk, V. Prevolnik-Rupel, A. Tapajner,et al., unpublished data, 2013). For recruitment, we used thelargest outpatient center in each region in consideration thatpatients were approximately half from urban and rural areas.Each outpatient center recruited from 20 to 80 patients accordingto region size and diabetes prevalence [29]. All the study patientshad an established relationship with the outpatient centers.Patients who met our inclusion criteria were asked to participatein this study. The inclusion criteria were as follows: physician-diagnosed DMT2, noninsulin treatment, and age 65 years orolder. Patients who were diagnosed as suffering from type 1diabetes, secondary diabetes, or gestational diabetes wereexcluded. All patients were diagnosed by physicians in light ofdiagnostic criteria recommended by the World Health Organiza-tion in 1999 [31].

A total of 500 patients with DMT2 were invited to participatein the research. Of them, 391 agreed and after exclusion ofincomplete questionnaires, our sample included 285 patientswith DMT2. The response rate was 57%.

After informed consent was obtained, all prospective partic-ipants were given the questionnaire. Where assistance wasneeded in completing the questionnaire, this was given bymedical students, who were trained in the use of the ADDQOLquestionnaire prior to the launch of this study.

Ethical Considerations

The study was approved by the National Medical Ethics Commit-tee of the Republic of Slovenia. The data obtained through thequestionnaires were anonymous and based on participantconsent.


The sample data were expressed as frequencies and percentagesfor categorical variables or by mean values and SDs for contin-uous variables. Binary logistic regression analysis was used toassess the influence of sociodemographic and health character-istics of patients with DMT2 on their QOL by using the ADDQOL.The calculation included Wald chi-square, odds ratio (OR), 95%confidence interval (95% CI), and P value. Nagelkerke’s R2 wasused to indicate goodness of fit. Patients were divided into twogroups by using the ADDQOL score by using quartiles; the firstgroup in the lower quartile was considered as having lower QOL.Such a cutoff strategy was previously applied in the literature[26,31]. Statistical analysis was performed with the SPSS 18.0software (SPSS, Inc., Chicago, IL). A P value of less than 0.05 wasconsidered statistically significant.

Results

Sociodemographic characteristics of the studied population arepresented in Table 1. The age ranged from 65 to 84 years, with amean of 70.0 � 4.9 years. Among the 285 respondents, less thanhalf were female (135, 47.4%). The majority of the respondentswere married (191; 67.0%), owned their own house (171; 60.0%),and lived in an urban area (243; 85.3%).

The body mass index (kg/m2) ranged from 16.9 to 53.0, with amean value of 29.6 � 5.0. A majority of the respondents havebeen living with DMT2 for 11 years or more (56.5%), and manyhad problems with hypertension (78.9%) and high cholesterol(59.6%). More details about respondents’ health characteristicsare shown in Table 2. A vast majority of the respondents (230,80.7%) reported to be satisfied with professional health supportprovision, and 114 (40.0%) were of the opinion that their diabeteswas under control.

The ADDQOL score of 285 patients with DMT2 was calculatedin a range of −8.3 to 0.0 on a defined range from −9 to þ3. Themedian ADDQOL score was calculated at −1.6, lower quartilecutoff was calculated at −3.0, 213 (74.7%) patients with DMT2reported an ADDQOL score of −3.0 or more, and 72 (25.3%)patients had an ADDQOL score of less than −3.0 (lower QOL).Eleven patients (3.9%) reported an ADDQOL score of 0, whichmeans that their QOL was not affected by diabetes at all.

Table 3 shows the logistic regression model results of thepredictors of QOL according to the ADDQOL score. Lower QOL wassignificantly connected to a heart attack episode (OR 2.42; 95% CI1.06–5.20). From a patient perspective, being of the opinion thattheir diabetes was under control decreased the likelihood of alower QOL (OR 0.36; 95% CI 0.18–0.69). Living in a rural environ-ment was not significantly connected to a lower QOL. Results inFigure 1 show that only 13.6% of the patients without heart attack


and being of the opinion that their diabetes was under controlreported a lower QOL.

The distribution of responses and the weights assigned to theimpact ratings are shown in Table 4. Diabetes had the greatestimpact on “freedom to eat” (mean impact rating: −1.5 � 1.0) andthe least impact on “people’s reaction” (mean −0.4 � 1.0).“Dependence on others” was rated as the most important (mean2.5 � 0.7), and “freedom to drink” was rated as the leastimportant to them (mean 1.2 � 1.0). After considering weighting,“freedom to eat” remained as the most (mean −3.2 � 2.9) and“people’s reaction” as the least (−0.8 � 1.5) affected QOL domains,respectively.

The ADDQOL instrument includes five domains that respond-ents can choose not to score. If no answer is provided, theADDQOL score is calculated without these domains. Respondentsin this study showed less interest in working life (the “notavailable” [NA] response was 76.5%) and sex life (the NA response

was 54.0%) domains (Table 4). NA responses are important whenthe reliability and construct validity of the ADDQOL instrumentare considered. In our study, because of the high NA response inworking life and sex life, reliability and instrument validitycalculation was possible only on 17 domains. Even then, thesample size for reliability and construct validity was n ¼ 180because holidays, family life, and personal relationship domainsin combination excluded n ¼ 105 respondents. The reliability ofthe ADDQOL instrument according to Cronbach’s alpha was 0.91(weighted impact score). For validation purposes, we used factoranalysis with forced one-factor solution. This condition wasimposed because the ADDQOL was intended to provide a singlesummary score [21]. In the forced one-factor solution, alldomains with the exception of “freedom to drink” had factorloadings of more than 0.4. Freedom to drink loaded with a valueof 0.285 into this factor. The forced one-factor solution explained48.8% of the total variance.

Discussion

This study provides detailed information about diabetes-dependent QOL and its assessment among older patients withDMT2 in Slovenia by using the ADDQOL, which is a widely useddiabetes-specific scale in the literature [21–25,32,33]. To the bestof our knowledge, the current study is the first to measure theHRQOL of older patients with DMT2 in Slovenia. WeightedADDQOL domain scores reliability was similar to that in previousstudies [26,27,32]. Structure validity results supported the one-factor scale structure of the ADDQOL, and the “freedom to drink”domain was calculated as the only possible domain that may notcontribute to the ADDQOL instrument. The forced one-factorsolution explained 48.8% of the total variance, which was alsosimilar to that in previous studies [27,34]. The findings of thepresent study highlight the impact of DMT2 on QOL. An interest-ing finding in the current study was that a few patients report noimpact of DMT2 on their QOL at all. In the rest, however,particular emphasis was put on the impact of “freedom to eat,”“dependency on others,” and “family life.” Consistent with earlierstudies [22–24,28,32], we found that the greatest negative impactobserved was for the domain “freedom to eat,” indicating thestrong influence of dietary restrictions on the QOL. Similarly, theleast affected domain was “people’s reaction.” Relative to theoverall negative effects of diabetes on the QOL, the effect ofspecific sociodemographic and clinical factors was fairly modest[25,31].

The results in the present study show that lower QOL wassignificantly connected to the presence of additional healthproblems (i.e., heart attack). Other studies show that the influ-ence of comorbidities or health complications in diabetes-dependent QOL seems unclear. Collins et al. [25] measured theamount of diabetes complications and concluded that theincreased number did not result in lower QOL. Similarly, inChung et al. [26], increased microvascular complications showedno association. Conversely, in Wang and Yeh [27], complicationsresulted in lower QOL, yet comorbidities provided no association.In concordance with Collins et al. [25], the study presented hereprovided no association of demographic data on diabetes-dependent QOL. Wang and Yeh [27], however, concluded thatmore education has negatively affected the QOL, and Chung et al.[26] found a positive association between older age and higherQOL. In the study presented here, we also measured the influenceof diabetes duration on the QOL and found no association, whichwas in concordance to the results of Wang and Yeh [27].

Respondents who reported that they managed their diseasewell and have it under control showed a decreased likelihood oflower QOL, which suggests the importance of self-management

Table 2 – Health characteristics.

N ¼ 285 %

Duration of diabetes mellitus (y)≤4 51 17.95–10 73 25.6≥11 161 56.5

Hypertension 225 78.9High cholesterol 170 59.6Poor eye vision 26 9.1Kidney dialysis 0 0.0Foot amputation 8 2.8Brain stroke 23 8.1Heart attack 39 13.7

Table 1 – Sample data.

N ¼ 285 %

GenderMale 150 52.6Female 135 47.4

EducationPrimary education 87 30.5Secondary education 163 57.2College or higher 35 12.3

Marital statusMarried, in partnership 191 67.0Widowed 71 24.9Divorced 11 3.9Alone 12 4.2

ResidenceOwn house 171 60.0Own apartment 92 32.3Renting 11 3.9Relatives 8 2.8Nursing home 3 1.1

Monthly income in euro≤365 36 12.6366–730 162 56.8731–1100 60 21.1≥1101 or above 27 9.5

Region≤200 per km2 (rural) 42 14.74200 per km2 (urban) 243 85.3

Age (y), mean � SD, range 70.0 � 4.9 65–84Body mass index, mean � SD, range 29.6 � 5.0 16.9–53.0


of the disease [16,36–37]. In the present study, despite the factthat more than 50% of the patients have been living with DMT2for 11 or more years, only 40% reported that they have theirdisease under control. This finding suggests that work is neededto increase patient empowerment and DMT2 self-management inSlovenia.

Open access to primary health care is ensured for all healthcare–insured individuals in Slovenia, in terms of both economicand geographical accessibility. Despite this, differences betweenregions in Slovenia have been reported, mainly due to aninadequate distribution in the number of primary health carepersonnel in some of the more remote rural areas, in which therecan be a lack of doctors [38]. Therefore, we hypothesized thatolder patients with DMT2 living in rural areas would show alower QOL than would patients living in urban areas. The presentstudy, however, showed no connection with place of living andQOL. This finding, together with the finding of the parallelresearch on diabetes knowledge [39], implies that accessibilityto chronic disease care provision and information does notdepend on the place of living in Slovenia. This is in line withthe European Health Interview Survey study [29,40], whichimplies that the use of primary care is relatively evenly spreadacross socioeconomic classes in Slovenia. In addition to the QOL,

our results suggest that the older patients with DMT2 aresatisfied with the delivery of care.

Although the research presented here shows no difference inQOL between rural and urban areas, this is not the same assaying that both groups receive high-quality care or that it cannotbe improved. Health services should aspire to improve the QOL ofolder patients with DMT2. This is going to become an increas-ingly important issue as the prevalence and economic burden ofdiabetes among the Slovenian older population rises. A lack ofcritical assessment at a system level may hamper the attempt toimprove the QOL. Hence, a well-translated and culturally adapteddisease-specific HRQOL measure such as the ADDQOL couldcontribute to the more accurate assessment of the effectivenessof disease management programs.

Study Strengths and Limitations

The study gives an overview of the self-perceived QOL of olderpopulations with DMT2. The main strength of the ADDQOL is thatit measures the QOL in various areas of people’s lives. This,however, can have a consequence that not all areas are appli-cable to all respondents. As a result, some respondents did notprovide complete data for all ADDQOL domains, which may have

Table 3 – Predictors for lower QOL according to the ADDQOL score (prediction model: χ2: 48.697, df ¼ 22, P o0.001).

ADDQOL score Wald χ2 OR (95% CI) P

Higher Lower≥−3.0 o−3.0n ¼ 213 (in %) n ¼ 72 (in %)

Living in rural areas (≤200 /km2) 12.7 20.8 2.38 1.96 (0.83–4.63) 0.123Female 49.8 40.3 1.54 0.66 (0.34–1.28) 0.215Age (y)65–74 (old) 72.3 75.0 1.00 (ref)75–84 (old-old) 27.7 25.0 0.32 0.81 (0.38–1.71) 0.573

Body mass indexo25 13.1 13.9 1.00 (ref)≥25–o30 50.7 43.1 1.99 0.48 (0.14–1.12) 0.156≥30 36.2 43.1 1.36 0.56 (0.21–1.48) 0.244

EducationPrimary education 29.6 33.3 1.00 (ref)Secondary education 57.7 55.6 0.01 0.96 (0.46–2.01) 0.922College or higher 12.7 11.1 0.73 0.55 (0.14–2.16) 0.394

Monthly income (euro)≤365 11.3 16.7 1.00 (ref)366–730 56.8 56.9 0.28 0.78 (0.31–1.97) 0.598731–1100 22.5 16.7 1.09 0.55 (0.17–1.70) 0.296≤1101 9.4 9.7 0.22 1.46 (0.30–7.13) 0.639

Residence in own house 61.5 55.6 1.73 0.65 (0.34–1.24) 0.189Single/divorced/widowed 32.9 33.3 0.02 1.05 (0.50–2.21) 0.888Years of diabetes≤4 17.8 18.1 1.00 (ref)5–10 27.7 19.4 2.10 0.50 (0.20–1.28) 0.147≥11 54.5 62.5 0.48 0.75 (0.33–1.69) 0.487

Hypertension 77.9 81.9 0.02 0.95 (0.44–2.02) 0.888High cholesterol 56.8 68.1 2.24 1.64 (0.86–3.15) 0.134Poor eye vision 7.0 15.3 3.61 2.53 (0.97–6.61) 0.057Foot amputation 1.4 6.9 1.34 2.62 (0.51–13.38) 0.247Brain stroke 9.4 4.2 3.23 0.28 (0.07–1.12) 0.072Heart attack 11.7 19.4 4.53 2.42 (1.06– 5.20) 0.036General health care satisfaction 82.2 76.4 0.10 0.89 (0.41–1.91) 0.757Being of opinion to manage the disease 45.1 25.0 9.25 0.36 (0.18–0.69) 0.002

Note. Nagelkerke R2 ¼ 0.215.ADDQOL, Audit of Diabetes-Dependent Quality of Life; CI, confidence interval; OR, odds ratio; QOL, quality of life.


introduced unintended biases into the analyses. The reason forthe low response rate of the item “working life” might be areflection of the fact that all the responders were aged 65 years orolder and mainly retired.

The lack of a randomized sampling and use of a conveniencesampling limit the ability to generalize the results. Because of

patients’ willingness to participate in the study, a response biasmight have occurred. A larger sample would provide more powerto detect significant relationships between study variables anddifferences between groups.

In the literature, we were not able to find many studies thatresearched the influence of various characteristics on diabetes-dependent QOL. Generally, all studies included demographic dataand health problems together with some specific characteristicsof their research interest; in our case, this was urban or ruralliving area. Among demographic data, age and gender were usedalways; other data such as education, marital status, or employ-ment (income) were used optional. Health problems were pre-sented as sum variables called complications and once separatelyas complications and comorbidities. In our study, we decided toshow the influence of some complications and comorbidities bydisplaying the exact type of the disease (e.g., hypertension, heartattack, and brain stroke). This can provide an added insight thatthe existing literature is missing and lead to improved diabetes-dependent QOL knowledge.

Conclusions

The results of the current study are similar to findings in priorresearch conducted in other countries. This study also demonstratesthat many of the factors related to diabetes-dependent HRQOL areapplicable regardless of the country and health care system.

DMT2 is of growing public health concern in Slovenia. Itimposes a personal burden on individuals and consumes asignificant portion of society’s scarce health care resources.Information on the QOL of older patients with diabetes is there-fore important to Slovenian policymakers and health workers. Itis essential in helping to identify and implement appropriate

Table 4 – Distribution of response (N ¼ 285) by impact and importance rating together with weightedimpact score.

Domain Not available response (%) Mean � SD

Impact rating Importance rating Weighted impact scoreLeisure activities −1.1 � 1.0 1.9 � 0.7 −2.2 � 2.2Working life 218 (76.5) −1.4 � 1.0 2.1 � 0.7 −3.0 � 2.8Journeys −1.3 � 1.0 1.8 � 0.8 −2.5 � 2.5Holidays 98 (34.4) −1.1 � 1.0 1.8 � 0.8 −2.0 � 2.3Physical health −1.2 � 1.0 1.9 � 0.7 −2.5 � 2.4Family life 8 (2.8) −0.9 � 0.9 2.4 � 0.6 −2.3 � 2.5Friendship and social life −0.9 � 1.0 2.0 � 0.8 −1.9 � 2.5Personal relationship 74 (26.0) −0.8 � 1.0 2.3 � 0.7 −1.9 � 2.4Sex life 154 (54.0) −1.1 � 1.0 1.8 � 0.9 −2.3 � 2.7Physical appearance −0.7 � 0.9 1.5 � 0.9 −1.4 � 2.1Self-confidence −0.8 � 1.0 2.0 � 0.7 −1.8 � 2.4Motivation −0.9 � 1.0 1.9 � 0.7 −2.0 � 2.6People’s reaction −0.4 � 0.7 1.5 � 0.9 −0.8 � 1.5Feelings about future −1.1 � 1.0 1.9 � 0.7 −2.5 � 2.5Financial situation −0.6 � 0.9 2.0 � 0.7 −1.3 � 2.2Living conditions −0.9 � 0.9 2.1 � 0.7 −2.0 � 2.5Dependence on others −0.6 � 0.9 2.5 � 0.7 −1.5 � 2.3Freedom to eat −1.5 � 1.0 1.8 � 0.9 −3.2 � 2.9Freedom to drink −0.9 � 1.0 1.2 � 1.0 −1.5 � 2.3

Notes. Impact rating (conditions without diabetes): −3, very much better; −2, much better; −1, a little better; 0, the same; þ1, worse.Importance rating: 0, not at all important; 1, somewhat important; 2, important; 3, very important.Weighted impact score ¼ impact rating (−3 to þ1) � importance rating (0–3) ¼ −9 (maximum negative impact of diabetes) to þ3 (maximumpositive impact of diabetes).

Fig. 1 – Percentage of patients in the lower ADDQoL quartile(o�3.0) according to heart attack (HA) status and theirperception of disease control. ADDQOL, Audit of Diabetes-Dependent Quality of Life; QOL, quality of life.


interventions for achieving the better management of diabetesand ultimately improving the QOL of patients with diabetes.

Acknowledgments

The authors thank Jelka Zaletel, MD, PhD, for her suggestions andadvice. We gratefully acknowledge the outpatient diabetic cen-ters and the diabetic patients in the regions for their willingnessto participate in the study. E.T. and A.I. were responsible for thestudy conception and design. E.T. performed the data collection.E.T. and A.T. performed the data analysis and were responsiblefor the drafting of the manuscript. A.I., S.L., and V.P.R. madecritical revisions to the manuscript for important intellectualcontent. A.T. provided statistical expertise. A.I. and V.P.R. pro-vided administrative, technical, or material support. S.L. editedthe manuscript. A.I. supervised the study.

Source of funding support: The project did not receive anyexternal funding.

R E F E R E N C E S

[1] Cox WM, Blount JP, Crowe PA, Singh SP. Diabetic patients’ alcohol useand quality of life: relationships with prescribed treatment complianceamong older males. Alcohol Clin Exp Res 1996;20:327–31.

[2] Boye KS, Yurgin N, Dilla T, et al. Health-related quality of life of patientswith type 2 diabetes mellitus in primary care in Spain: self-reportedand proxy assessment using the EQ-5D. J Med Econ 2007;10:41–58.

[3] Glasgow RE, Ruggiero L, Eakin EG, et al. Quality of life and associatedcharacteristics in a large national sample of adults with diabetes.Diabetes Care 1997;20:562–7.

[4] Wild SH, Roglic G, Green A, et al. Global prevalence of diabetes:estimates for the year 2000 and projections for 2030 response toRathman and Giani. Diabetes Care 2004;27:1047–53.

[5] Shaw J, Sicree R, Zimmet P. Global estimates of the prevalence ofdiabetes for 2010 and 2030. Diabetes Res Clin Pract 2010;87:4–14.

[6] IDF. Diabetes atlas. Available from: http://archive.diabetesatlas.org/content/europe. [Accessed July 17, 2012].

[7] Moravec Berger D, Zupanič T. Epidemiologija in prevalenca sladkornebolezni v Sloveniji. In: Lovšin D, ed., 1 nacionalna konferenca odiabetesu 6–7. november 2008; Ljubljana: Zavod za izobraževanje odiabetesu, 2008.

[8] Yong PL, Olsen LA, McGinnis JM. Value in Health Care: Accounting forCost, Quality, Safety, Outcomes, and Innovation: Workshop Summary.Washington, D.C. National Academies Press, 2010.

[9] Ali MK, Weber MB, Narayan K. The global burden of diabetes. In: HoltRIG, Cockram CS, Flyvbjerg A, et al., eds., Textbook of Diabetes. (4th ed.).Oxford, UK: Wiley-Blackwell, 2010:69–84.

[10] Devlin NJ, Appleby J. Getting the Most Out of PROMs. London: KingsFund, 2010.

[11] Krumholz HM. Outcomes research: generating evidence for bestpractice and policies. Circulation 2008;118:309–18.

[12] Barr J. The outcomes movement and health status measures. J AlliedHealth 1995;24:13.

[13] Magwood GS, Zapka J, Jenkins C. A review of systematic reviewsevaluating diabetes interventions. Diabetes Educ 2008;34:242–65.

[14] Tapp RJ, O'Neil A, Shaw JE, et al. Is there a link between components ofhealth-related functioning and incident impaired glucose metabolismand type 2 diabetes? Diabetes Care 2010;33:757–62.

[15] Aghamolaei T, Eftekhar H, Mohammad K, et al. Effects of a healtheducation program on behavior, hbA1c and health-related quality oflife in diabetic patients. Acta Medica Iranica 2005;43:89–94.

[16] Khanna A, Bush AL, Swint JM, et al. Hemoglobin A1c improvementsand better diabetes-specific quality of life among participants

completing diabetes self-management programs: a nested cohortstudy. Health Qual Life Outcomes 2012;10:48.

[17] Norris SL, Nichols PJ, Caspersen CJ, et al. Increasing diabetes self-management education in community settings: a systematic review.Am J Prev Med 2002;22:39–66.

[18] Garratt A, Schmidt L, Fitzpatrick R. Patient‐assessed health outcomemeasures for diabetes: a structured review. Diabetic Med 2002;19:1–11.

[19] El Achhab Y, Nejjari C, Chikri M, Lyoussi B. Disease-specific health-related quality of life instruments among adults diabetic: a systematicreview. Diabetes Res Clin Pract 2008;80:171–84.

[20] Watkins K, Connell CM. Measurement of health-related QOL indiabetes mellitus. Pharmacoeconomics 2004;22:1109–26.

[21] Bradley C, Todd C, Gorton T, et al. The development of anindividualized questionnaire measure of perceived impact of diabeteson quality of life: the ADDQoL. Quali Life Res 1999;8:79–91.

[22] Speight J, Woodcock A, Plowright R, Bradley C. Design of anindividualised measure of the impact of diabetes on the quality of lifeof elderly people: the ADDQoL Senior. Qual Life Res 2003;12:1509.

[23] Costa FA, Guerreiro JP, Duggan C. An audit of diabetes dependentquality of life (ADDQoL) for Portugal: exploring validity and reliability.Pharm Pract 2006;4:123–8.

[24] Holmanová E, Žiaková K. Audit Diabetes‐Dependent Quality of Lifequestionnaire: usefulness in diabetes self‐management education inthe Slovak population. J Clin Nurs 2009;18:1276–86.

[25] Collins MM, O'Sullivan T, Harkins V, Perry IJ. Quality of life and qualityof care in patients with diabetes experiencing different models of care.Diabetes Care 2009;32:603–5.

[26] Chung J, Cho D, Chung D, Chung M. Assessment of factors associatedwith the quality of life in Korean type 2 diabetic patients. Intern Med(Tokyo, Japan) 2012;52:179–85.

[27] Wang H-F, Yeh MC. The quality of life of adults with type 2 diabetes ina hospital care clinic in Taiwan. Qual Life Res 2012;22:577–84.

[28] Bradley C, Speight J. Patient perceptions of diabetes and diabetestherapy: assessing quality of life. Diabetes/Metab Res Rev 2002;18(Suppl. 3):S64–9.

[29] SURS. Slovenske regije v številkah. Statistični urad Republike Slovenije:Statistični urad Republike Slovenije, 2011.

[30] Buzeti T, Gobec M. Health inequalities in Slovenia. Ljubljana, Slovenia:National Institute of Public Health. 2012.

[31] World Health Organization. Definition, diagnosis and classification ofdiabetes mellitus and its complications. Geneva, Switzerland: WHO,Department of Noncommunicable Disease Surveillance, 1999.

[32] Trief PM, Wade MJ, Pine D, Weinstock RS. A comparison of health-related quality of life of elderly and younger insulin-treated adults withdiabetes. Age Ageing 2003;32:613–8.

[33] Kong D, Ding Y, Zuo X, et al. Adaptation of the Audit of Diabetes-Dependent Quality of Life questionnaire to patients with diabetes inChina. Diabetes Res Clin Pract 2011;94:45–52.

[34] Akinci F, Yildirim A, Gözü H, et al. Assessment of health-related qualityof life (HRQoL) of patients with type 2 diabetes in Turkey. Diabetes ResClin Pract 2008;79:117–23.

[35] Wee HL, Tan CE, Goh SY, Li SC. Usefulness of the Audit of Diabetes-Dependent Quality-of-Life (ADDQoL) questionnaire in patients withdiabetes in a multi-ethnic Asian country. Pharmacoeconomics2006;24:673–82.

[36] Peyrot M, Rubin RR, Funnell MM, Siminerio LM. Access to diabetes self-management education. Diabetes Educ 2009;35:246–63.

[37] Chodosh J, Morton SC, Mojica W, et al. Meta-analysis: chronic diseaseself-management programs for older adults. Ann Intern Med2005;143:427–38.

[38] Norris SL, Lau J, Smith SJ, et al. Self-management education for adultswith type 2 diabetes. Diabetes Care 2002;25:1159–71.

[39] Albreht T, Turk E, Toth M, et al. Slovenia: health system review. In:Health Systems in Transition (Vol. 11, No. 3). Copenhagen, Denmark:World Health Organization, 2009.

[40] Turk E, Palfy M, Prevolnik Rupel V, Isola A. General knowledge aboutdiabetes in the elderly diabetic population in Slovenia. Slovenian Med J2012;81:517–25.

[41] National Institute of Public Health. European Health Interview Survey(EHIS). Ljubljana, Slovenia: National Institute of Public Health, 2007.
















http://archive.diabetesatlas.org/content/europe

http://archive.diabetesatlas.org/content/europe



































































































Patient-Reported Quality of Life During Antiretroviral Therapy in aNigerian HospitalAzuka C. Oparah, PhD�, Jeffrey S. Soni, PharmD, Herbert I. Arinze, PharmD, Ifeanyi E. Chiazor, PharmD

Department of Clinical Pharmacy and Pharmacy Practice, University of Benin, 300001, Benin City, Nigeria

A B S T R A C T

Objectives: We assessed the reported quality of life of patients withHIV/AIDS and explored the impact of patients’ sociodemographicprofile on the quality-of-life domains. Methods: Consenting outpa-tients who met criteria were consecutively selected in a secondaryhealth care facility in Benin City, Nigeria. Quality of life was deter-mined in the nine domains of HIV/AIDS Targeted Quality of Life (HAT-QOL) instrument. Quality-of-life scores were computed on the scale of0 to 100 and triangulated with a rated interval scale of 1 to 5 suited forquantitative analysis. Association between rated scores and eachdomain was explored by using Students’ t test and analysis ofvariance at 95% confidence interval. Results: Out of the 403 patients,82.1% were females; 147 (36.1%) belonged to the modal age group of 20to 30 years; the mean age for grouped data was 39.2 years. About 239(58.7%) were not married. Also, 338 (83.0%) earned below $1500 perannum and 303 (74.40%) had basic education (grade 1 to grade 9). HAT-QOL scores indicated the following: overall function (89.96 � 5.62); lifesatisfaction (91.94 � 3.62); health worries (87.06 � 4.28); financial

worries (81.00 � 3.95); medication worries (91.65 � 4.47); HIV mastery(71.00 � 3.11); disclosure worries (27.50 � 7.57); provider trust (91.63 �1.96); and sexual function (70.25 � 3.51). Likert-type rated scores werein agreement with HAT-QOL scores. Provider trust was associatedwith gender, employment status, and educational level. Sexualfunction was associated with gender and age (P o 0.05). Conclusions:Patients reported satisfactory quality of life in the domains of overallfunction, life satisfaction, health worries, financial worries, medica-tion worries, HIV mastery, provider trust, and sexual function. Qualityof life was low in the domain of disclosure worries, indicatingconcerns for discrimination and stigmatization. Age, level of educa-tion, and employment status had a strong impact on the quality of lifeof patients with HIV/AIDs.

Keywords: HAT-QOL, HIV/AIDS, Nigeria, quality of life.


Introduction

The HIV epidemic has resulted in history’s single sharpestreversal in human development. Sub-Saharan Africa remainsthe region most heavily affected by HIV, accounting for 67% ofall people living with HIV and for 75% of AIDS deaths in 2007. Interms of disease burden, Nigeria rank second in sub-SaharanAfrica, behind South Africa, and third in the world, behind India[1]. HIV/AIDS continues to contribute significantly to publichealth problems in Nigeria. Although HIV infection was initiallylimited to people with risky behaviors, such as commercial sexworkers, adolescents, youths, prisoners, and people with multi-ple sexual partners, the currently available evidence suggeststhat this infection has permeated all strata of the Nigerianpopulation [2].

With the alarming increase in HIV/AIDS pandemic in devel-oping countries, and the limited accessibility and availability ofhighly active antiretroviral therapy (ART), majority of the peopleliving with HIV/AIDS (PLWHA) continue to suffer the disease,with a serious impact on their quality of life (QOL). Quality of lifeis a term that is popularly used to convey an overall sense of

well-being and includes aspects such as happiness and satisfac-tion with life as a whole. QOL relates both to the adequacy ofmaterial circumstances and to personal feelings about thesecircumstances with overall subjective feelings of well-being thatis closely related to morale, happiness, and satisfaction [3–5]. QOLhas recently been scientifically defined, and it has been consid-ered synonymous with health status, functional status, psycho-logical well-being, happiness with life, satisfaction of needs, andassessment of one’s own life [6].

Several instruments for measuring QOL have been developedand described [7]. A study on the QOL of PLWHA in São Paulo,Brazil, reported that despite differences in sex, skin color,income, and mental and immunological status, PLWHA havebetter (physical and psychological) QOL than do other patients,but lower quality in the social relationships domain [8]. A similarstudy in South India also showed that patients had the worst QOLin the social domain, indicating that the patients’ social contactsand sexual activity were affected markedly to a great extent [9].Fatiregun et al. [2], in their own study of PLWHA in Kogi State,Nigeria, suggested that stigma and discrimination as well aspoor living conditions in the physical environment accounted






� Address correspondence to: Azuka C. Oparah, Department of Clinical Pharmacy and Pharmacy Practice, University of Benin, 300001Benin City, Nigeria.


dx.doi.org/10.1016/j.vhri.2013.07.004

dx.doi.org/10.1016/j.vhri.2013.07.004

dx.doi.org/10.1016/j.vhri.2013.07.004


dx.doi.org/10.1016/j.vhri.2013.07.004

for lower QOL in the environment and social relationshipsdomain [2].

An investigation of the determinants of QOL of PLWHA identi-fied gender, mental status, CD4 cell count, and stage of the diseaseto be important factors associated with the QOL of patients [6]. In astudy of the QOL of HIV-infected Brazilians, the authors reportedthat unemployment was associated with worse QOL in all thedomains measured, except the domain of spirituality [10]. Otherstudies concluded that education, income, occupation, family sup-port, and clinical categories were significantly linked to patients’QOL [11,12]. In a sample of highly active ART-naive asymptomaticHIV-infected subjects, high viral loads and low CD4 count weresignificantly associated with poorer scores in the psychological andsocial domains [13]. Self-reporting indicates that QOL is severelycompromised in PLWA in stages 3 and 4 and limitations in the fourdomains of mobility, usual activities, pain/discomfort, and anxiety/depression constitute major problems for PLWA [14]. Focus groupdiscussions have also revealed that QOL weights are stronglycorrelated to disease stage. Furthermore, clinical experts consis-tently report that ART has a strong positive impact on the HIV/AIDSTargeted Quality of Life (HAT-QOL) of patients, although this effectappears to rebound in cases of drug resistance [15]. Cotton et al. [16]tested six new items addressing personal and social domains ofreligiosity and spirituality. The authors reported that many patientshad become more spiritual or religious by virtue of having HIV/AIDS, and half of the patients believed that their spirituality/religionwas helping them to live longer.

In the setting of this study, patients with HIV/AIDS receivefree medications on a donor-funded project that has been inexistence for over 5 years. No investigation has been carried outto determine the QOL of patients, which is an important treat-ment outcome. We assessed the reported QOL of patients withHIV/AIDS and explored the impact of the patients’ sociodemo-graphic profile on the QOL domains.

Methods

Design and Setting

This study used a cross-sectional descriptive design and wascarried out at Central Hospital Benin City, Edo State. The facilityhas 432 bed spaces and an average daily outpatient attendance of900. This hospital is a secondary public health care facility thatoffers comprehensive HIV care services, including ART. HIV/AIDSoutpatients were accessed at the Antiretroviral Pharmacy, asubunit of the main pharmacy in the hospital.

Data Collection

Data on patients’ QOL were gathered by using the HIV/AIDS-Targeted Quality of Life (HAT-QOL) instrument [17]. The HAT-QOLconsists of 34 item stems in nine domains: overall function, lifesatisfaction, health worries, financial worries, medication con-cerns, HIV mastery, disclosure worries, provider trust, and sexualfunction. Response to the questionnaire items was anchored on ascale of 5 to 1 as follows: all of the time ¼ 5, a lot of the time ¼ 4,some of the time ¼ 3, a little of the time ¼ 2, and none of the time¼ 1. Eight of the 34 items were negatively worded; these itemswere reverse scored for analysis so that higher scores indicatedhigher QOL in each domain.

Consenting outpatients, 18 years and above, visiting the facilityduring the period of August 1 to 30, 2011, were consecutivelyapproached. A brief introduction of the questionnaire was given tothe participants, who were asked to complete the assessment.Patients who were literate completed the questionnaire unaided,while those unable to read were aided by an interpreter, until 403

participants were recruited within the targeted period. Sociodemo-graphic characteristics of the patients were obtained from theircase notes.

Ethical Considerations

The ethical approval for this study was obtained from the EthicalCommittee of the Central Hospital, Benin City, Nigeria. Partic-ipants gave a verbal consent and were assured of the confiden-tiality of the information that they volunteered.

Data Analysis

Usable responses from the 34-item questionnaire with five-pointresponse scale were entered into a Microsoft Excel spreadsheetand rechecked for accuracy prior to analysis.

Data were loaded into the Statistical Package for SocialSciences software version 17.0 for descriptive statistical analysis.All HAT-QOL domains were scored so that the final domain scorewas transformed into a linear 0 to 100 scale, where 0 was theworst score possible and 100 was the best score possible.Obtaining this final transformed domain score was done in foursteps indicated below.

1. A value for all subject responses, item by item, asnoted below was imputed. Most item responseswere valued by using Code A, while others werevalued by using Code B.

Response Option Code A Code B

“All of the time” 1 5“A lot of the time” 2 4“Some of the time” 3 3“A little of the time” 4 2“None of the time” 5 12. Code A was used for the following items: 1b, 1c, 1d,1e, 1f, 3a, 3b, 3c, 3d, 4a, 4b, 4c, 5a, 5b, 5c, 5d, 5e, 6a,6b, 7a, 7b, 7c, 7d, 7e, 9a, and 9b; while

3. Code B was used for the following items: 1a, 2a, 2b,2c, 2d, 8a, 8b, and 8c.

The mean score of items within each domain was used tocalculate the domain scores and subsequently transformed to a0 to 100 scale by using the questionnaire scoring guide as follows:

� Overall function: OVFXN100 ¼ (100/(30�6)) � (OVFXN-6)

� Life satisfaction: LISAT100 ¼ (100/(20�4)) � (LISAT-4)Health worries: HEAWO100 ¼ (100/(20�4)) � (HEAWO-4)

� Financial worries: FINWO100 ¼ (100/(15�3)) �(FINWO-3)Medication worries: MEDWO100 ¼ (100/(25� 5)) � (MEDWO-5)

� HIV mastery: HIVMA100 ¼ (100/(10�2)) � (HIVMA-2)Disclosure worries: DISWO100 ¼ (100/(25�5)) �(DISWO-5)

� Provider trust: PROTR100 ¼ (100/(15�3)) � (PROTR-3)Sexual function: SXFXN100 ¼ (100/(10�2)) � (SXFXN-2)

Furthermore, the summated rated scores (range of 1–5) for eachdomain were computed in terms of mean and the SD, andCronbach’s alpha. Association between rated scores and eachdomain was explored by using Students’ t test and one-way


analysis of variance at 95% confidence interval with the aid ofGraphPad Instat version 3.0. The level of statistical significancewas set at P less than .05.

Results

Patients’ Demographic Characteristics

Of the 403 patients, 72 (17.7%) were males and 331 (81.3%) werefemales. Majority of the patients (147, 36.1%) were aged 20 to 30 years,and 239 (58.7%) were not married. Also, 338 (83.0%) earned below$1500 per annum, and 303 (74.40%) had basic education (grade 1 tograde 9). Other demographic characteristics are shown in Table 1.

HAT-QOL Scores

The highest mean score was in medication worries (91.65),followed by provider trust (91.63) and life satisfactions (91.94).Disclosure worries had the least mean score (27.50) (Fig. 1).

Likert-Rated QOL Scores

Medication worries had the highest Likert mean score (4.67 �

0.894; range 1–5), followed by provider trust (4.67 � 0.980).Disclosure worries had the least Likert score (2.10 � 1.152). TheCronbach’s alpha for the domains ranged from 0.580 to 0.954(Table 2).

Impact of Sociodemographic Characteristics on HAT-QOLDomains

Age and gender of the patients were associated with the sexualfunction domain. Furthermore, level of education and

employment status were associated with the provider trustdomain (P o 0.05). Details of the statistically significant associ-ations between the sociodemographic factors and HAT-QOLdomains are presented in Table 3.

Discussion

The validity of HAT-QOL has been demonstrated [18–20]. Nostudies, however, have been conducted in this southern part ofNigeria. The instrument produced high internal consistency inthe various domains, indicating the validity of the HAT-QOLinstrument in a Nigerian setting, bearing the cultural sensitivityof QOL instruments.

Patients in this cross-sectional study reported a high QOL inall the domains except the domain of disclosure worries. This isconsistent with the study carried out in Kogi State, Nigeria, wheredisclosure worries was also found to be low [2]. This could be aresult of fear of stigmatization and discrimination amongpatients with HIV/AIDS. Stigmatization is an inhibiting factor tothe uptake of HIV counseling and testing and other preventive,care, and support services. Stigmatization and discriminationcontribute inadvertently to the transmission of HIV/AIDS amongunsuspecting sexually active people. The fear of violence as aconsequence of HIV/AIDS status disclosure also contributes toHIV transmission. HIV stigma has a significantly negative andconstant impact on life satisfaction QOL for people with HIVinfection. In the absence of any intervention to address andreduce stigmatization, individuals will continue to report poorerlife satisfaction, evidenced by reduced living enjoyment, loss ofcontrol in life, decreased social interactivity, and decreasedperceived health status [19].

Table 1 – Sociodemographic data of patients.

Variable Number reporting, n (%)

Age (y)Below 20 5 (1.2)20–29 73 (17.9)30–39 147 (36.1)40–49 98 (24.1)50 and above 84 (20.7)

SexMale 72 (17.7)Female 331 (81.3)Not indicated 4 (1.0)

Marital statusNot married 239 (58.7)Married 106 (26.0)Divorced 22 (5.4)Widowed 40 (9.9)

OccupationUnemployed 46 (11.3)Employed 361 (88.7)

Monthly income (NGN 150 to $1)Below 20,000.00 338 (83.0)20,000–49,000.00 51 (12.5)50,000–69,000.00 9 (2.2)70,000–99,000.00 5 (1.2)100,000.00 and above 4 (1.0)

Educational levelNo formal education 33 (8.1)Basic education (grades 1–9) 303 (74.4)Post–basic education (grades10–16)macmac 71 (17.4)

Fig. 1 – Mean scores across the HAT-QOL domains. HAT-QOL, HIV/AIDS Targeted Quality of Life.

Table 2 – Likert-rated scores and Cronbach’s alphafor the HAT-QOL domains.

Domain Likert score Cronbach’s α

Overall function 4.56 � 0.936 0.824Life satisfaction 4.68 � 0.905 0.816Health worries 4.48 � 1.070 0.633Financial worries 4.24 � 1.317 0.694Medication worries 4.67 � 0.894 0.704HIV mastery 3.84 � 1.555 0.871Disclosure worries 2.10 � 1.152 0.790Provider trust 4.67 � 0.980 0.580Sexual function 3.81 � 1.759 0.954

HAT-QOL, HIV/AIDS Targeted Quality of Life.


HAT-QOL scores were high in the following domains: healthworries, medication worries, HIV/AIDS mastery, provider trust,overall function, financial worries, life satisfaction, and sexualfunction. This is in line with the studies conducted in Philadel-phia, PA [18], although a similar study carried out in the NigerDelta region in Nigeria revealed that there was impairment in theQOL except in medication worries and sexual function domainsin PLWHA in that region [19].

Income assessment did not affect any of the HAT-QOLdomains. This also supports the finding in a study carried outin three hospitals in Thailand [21]. This is mainly due tosimilarities in some sociodemographic characteristics, free avail-ability of and adherence to ART, and accessibility to other HIV/AIDS care components.

Provider trust and sexual function were affected by gender.Gender refers to the expectations or norms within a society aboutthe roles and responsibilities that are appropriate for women andmen. Research has shown that the gender-based imbalance inpower found in the economic and social spheres of life isreflected in sexual relationships. Women often have less controlover the nature and timing of sex and the practice of protectivebehaviors. A woman’s ability to practice safer sex may beinfluenced by her ability to communicate openly about sex with

her partner, the power dynamic in their relationship, or howmuch the partner believes in traditional gender roles. Beliefs ornorms about masculinity and femininity often encourage men tohave multiple partners and women to be passive and ignorantabout matters of sexuality and reproduction. Gender, therefore,affects both women’s and men’s risk of HIV and other sexuallytransmitted infections [23].

Employment status affected health worries and provider trustdomains. Level of education affected financial worries andprovider trust, while life satisfaction, HIV/AIDS mastery, andsexual function were affected by age. Positive changes tendedto be associated with higher educational levels and greaterincome, while negative changes were inversely related to healthstatus and level of optimism [12,13,22,23].

Implications of the Findings

Patients reported high QOL in all domains, except disclosureworries, indicating their concern for stigmatization and discrim-ination in the society. There is need for governmental andnongovernmental agencies to heighten advocacy toward reduc-ing stigmatization and discrimination of PLWHA to encourage theuptake of HIV/AIDS care and reduce its spread.

Table 3 – Impact of sociodemographic characteristics on HAT-QOL domains.

HAT-QOL domain Variable Mean score � SD Calculated value P

Provider trust Sex○ Male○ Female

4.433 � 1.206 t ¼ 1.967 0.04994.690 � 0.956

Sexual functions ○ Male○ Female

4.450 � 1.375 t ¼ 3.388 0.00083.675 � 1.831

Health worries Employment status○ Employed○ Unemployed

4.525 � 1.0144.150 � 1.401 t ¼ 2.229 0.0264

Provider trust ○ Employed○ Unemployed

4.710 � 0.8894.146 � 1.541 t ¼ 3.628 0.0003

Financial worries Level of education○ Basic education○ Post–basic education

4.357 � 1.241 t ¼ 3.451 0.00063.793 � 1.146

Provider trust ○ Basic education○ Post–basic education

4.705 � 0.936 t ¼ 2.820 0.00504.320 � 1.360

Life satisfaction Age (y)○ 20–29○ 30–39○ 40–49○ 50 and above

4.382 � 1.2594.700 � 0.871 F ¼ 6.169 0.00044.692 � 0.8504.875 � 0.528

HIV/AIDS mastery ○ 20–29○ 30–39○ 40–49○ 50 and above

3.450 � 1.145 F ¼ 4.370 0.00483.685 � 1.6244.160 � 1.2684.025 � 1.508

Sexual function ○ 20–29○ 30–39○ 40–49○ 50 and above

2.130 � 1.490 F ¼ 26.767 0.00014.190 � 1.5183.705 � 1.8172.970 � 1.968

HAT-QOL, HIV/AIDS Targeted Quality of Life.


The QOL was found not to be associated with income. Thereason may probably be due to free care the patients receivedthrough donor agencies. That would remain so if free care issustained. Bearing donor fatigue, Nigerians need to build localcapacity for long-time sustainability of free HIV/AIDS care atcommunity levels.

Provider trust has implications for patient care. Patients witha lower level of trust in their care provider are more likely toreport that requested or needed services are not provided.Understanding this relationship may lead to better ways ofresponding to patient requests that preserve or enhance patienttrust, leading to better outcomes.

Study Limitations

This study used a cross-sectional design, which could notaccount for changes in the QOL of PLWA over time such as alongitudinal survey would provide. Some respondents were aidedto complete the survey, and the possibility of bias cannot becompletely eliminated, despite efforts. Patients were consecu-tively recruited during the study period; a randomized selectionprocess would yield a more generalizable result.

Conclusions

Patients reported satisfactory QOL in the domains of overallfunction, life satisfaction, health worries, financial worries, med-ication worries, HIV mastery, provider trust, and sexual function.The QOL was low in the domain of disclosure worries, indicatingconcerns for discrimination and stigmatization. Age, level ofeducation, and employment status had a strong impact on theQOL of patients with HIV/AIDS.


R E F E R E N C E S

[1] United Nations Joint Programme on HIV/AIDS (2008). Report on theglobal AIDS epidemic. Geneva, UNAIDS. Available at: www.unaids.org/en/media/unaids/contentassets/dataimport/pub/globalreport/2008/jc1510_2008globalreport_en.pdf. [Accessed August 30, 2013].

[2] Fatiregun AA, Mofolorunsho KC, Osagbemi KG. Quality of life of peopleliving with HIV/AIDS in Kogi State, Nigeria. Benin J Postgrad Med2009;2:2–10.

[3] Lesserman J, Perkins DO, Evans DL. Coping with the threat of AIDS: therole of social support. Am J Psychiatr 1992;149:1514–20.

[4] Namir S, Wolcott D, Fawzy F, Alumbaugh M. Implications of differentstrategies for coping with AIDS. In: Temoshack L, Baum A,eds.Psychological Perspectives on AIDS. Hillsdale, NJ: Erlbaum, 1990.

[5] Rabkin JG, Remien R, Kattoff L, Williams JB. Residence in adversityamong long time survivors of AIDS. Hosp Comm Psychiatr1993;44:162–7.

[6] Elisabete C, Morandi Dos Santos, Ivan FJ, Fernanda L. Quality of life ofpeople living with HIV/AIDS in São Paulo. Brazil: Rev Saúde Pública2007;741(supp.2):647.

[7] McDowell I, Newell C. Measuring Health: A Guide to Rating Scales andQuestionnaires. (2nd ed.). New York: Oxford University Press, 1996.

[8] Razera F, Ferreira I, Bonamigo R. Factors associated with health-relatedquality of life in HIV infected Brazilians. Int J STD 2008;19:519–23.

[9] Basavarj KH, Navya MA, Rashmi R. Quality of life in HIV/AIDS. Ind J SexTransmission 2010;31:75–80.

[10] Razera F, Ferreira I, Bonamigo R. Factors associated with health-relatedquality of life in HIV infected Brazilians. Int J STD 2008;19:519–23.

[11] Wig N, Lekshmi R, Pal H, et al. The impact of HIV/AIDS on the quality oflife. Ind J Med Sci 2006;60:3–12.

[12] Odili VU, Oparah AC, Usifoh SF, Ikhurionan IB. Determinants of qualityof life in HIV/AIDS patients. West Afr J Pharm 2011;22:42–8.

[13] Chandra PS, Gandhi C, Satischchandra P, et al. Quality of life of HIV subtype c infections among asyptomatic subjects and its association withCD4 counts and viral load—a study from India Qual Life Res2006;15:1597–605.

[14] Janeen H, Jennifer J, Emilou M, et al. The health-related quality of life ofpeople living with HIV/AIDS. Informa Healthcare 2004;26:371–6.

[15] Robberstad B, Olsen JA. The health related quality of life of peopleliving with HIV/AIDS in sub-Saharan Africa—a literature review andfocus group study. Cost-Effect Res Alloc 2010;8:5. Available from: http://www.resource-allocation.com/content/8/1/5). [Accessed February 28,2012].

[16] Cotton S, Puchalski CM, Sherman SN, et al. Spirituality and religion inpatients with HIV/AIDS. J Gen Intern Med 2006;21(Suppl):S5–13.

[17] Holmes WA, Shea JA. Performance of a new HIV/AIDS Targeted Qualityof Life (HAT-QOL) instrument in asymptomatic sero-positiveindividuals. Qual Life Res 1997;6:561–71.

[18] Holmes WC, Ruocco JE. Test-retest evaluation of HAT-QoL and SF-36 inan HIV-seropositive sample. AIDS Care 2008;20:1084–92.

[19] Greeff M, Uys LR, Wantland D, Makoae L. Perceived HIV stigma and lifesatisfaction among persons living with HIV infection in five Africancountries: a longitudinal study. Int J Nurs Stud 2009: (NS-1559:12).

[20] Ekott FA, Bassey JU, Etukumana AE. Quality of life in people living withHIV/AIDS in Niger Delta Region, Nigeria. J Ment Health 2010;19:211–8.

[21] The HIV/AIDS Survey Library HIV research domains: gender and sexualrelationships. Available from: www.popcouncil.org/Horizons/ORToolkit/AIDQuest/topics/gender.html. [Accessed February 29, 2012].

[22] Updegraff JA, Taylor SE, Kemeny ME, Wyatt GE. Positive and negativeeffectives of HIV infection in women with low socioeconomicresources. Pers Soc Psychol Bull 2002;28:382–94.

[23] Marie P, Fabienne M, Patrizia MC, France L. Health-related qualityof life in French people living with HIV in 2003: results from the nationalANRS-EN12-VESPA Study. The VESPA Study Group. 2007;21(Suppl.):S19–27.



www.unaids.org/en/media/unaids/contentassets/dataimport/pub/globalreport/2008/jc1510_2008globalreport_en.pdf





































http://www.resource-allocation.com/content/8/1/5

http://www.resource-allocation.com/content/8/1/5














www.popcouncil.org/Horizons/ORToolkit/AIDQuest/topics/gender.html.

www.popcouncil.org/Horizons/ORToolkit/AIDQuest/topics/gender.html.










CLINICAL OUTCOMES STUDIES

Clinical Burden of Invasive Pneumococcal Disease in Selected DevelopingCountriesNamaitijiang Maimaiti, PhD1,2,�, Zafar Ahmed, PhD1,3, Zaleha Md Isa, PhD2, Hasanain Faisal Ghazi, PhD1,2,Syed Aljunid, PhD1,3

1International Institute for Global Health, United Nations University, Kuala Lumpur, Malaysia; 2Department of Community Health, Faculty of Medicine, UKM;3International Centre for Case-Mix and Clinical Coding, UKM Medical Centre

A B S T R A C T

Objective: To measure the clinical burden of invasive pneumococcaldisease (IPD) in selected developing countries. Methods: This is anextensive literature review of published articles on IPD in selecteddeveloping countries from East Asia, South Asia, Middle East, sub-Saharan Africa, and Latin America. We reviewed all the articlesretrieved from the knowledge bases that were published betweenthe years 2000 and 2010. Results: After applying the inclusion,exclusion, and quality criteria, the comprehensive review of theliterature yielded 10 articles with data for pneumococcal meningitis,septicemia/bacteremia, and pneumonia. These selected articles werefrom 10 developing countries from five different regions. Out of the 10selected articles, 8 have a detailed discussion on IPD, one of them hass detailed discussion on bacteremia and meningitis, and another onehas discussed pneumococcal bacteremia. Out of these 10 articles, only

5 articles discussed the case-fatality ratio (CFR). In our article review,the incidence of IPD ranged from less than 5/100,000 to 416/100,000population and the CFR ranged from 12.2% to 80% in the developingcountries. Conclusions: The review demonstrated that the clinicalburden of IPD was high in the developing countries. The incidence ofIPD and CFR varies from region to region and from country to country.The IPD burden was highest in sub-Saharan African countries fol-lowed by South Asian countries. The CFR was low in high-incomecountries than in low-income countries.

Keywords: clinical burden, developing countries, East Asia, IPD, LatinAmerica, Middle East, South Asia, sub-Saharan Africa.


Introduction

Streptococcus pneumoniae is a major public health problem causingmeningitis, bacteremia, pneumonia, and acute otitis media [1,2].S. pneumoniae is the most common cause of pneumonia world-wide, causing approximately 36% of all childhood pneumonias.S. pneumoniae can cause potentially life-threatening lung infec-tions including severe pneumonia, which hinders the movementof oxygen into the bloodstream, potentially resulting in deathfrom respiratory failure [3]. Invasive pneumococcal infectionsoften prove rapidly fatal, even where good medical treatment isreadily available. In developed countries, up to 20% of people whocontract pneumococcal meningitis die; however, in the develop-ing world, mortality is closer to 50%, even among hospitalizedpatients [4]. In one study from Gambia, 48% of children whocontracted pneumococcal meningitis and reached hospital didnot survive [5]. The World Health Organization (WHO) estimatedthat 1.6 million people die from pneumococcal disease every year[6], with 0.7 million to 1 million being children younger than 5years [7]. The pneumococcal disease burden among adults is also

high, with an estimated 600,000 to 800,000 adult deaths each yearfrom pneumococcal pneumonia, meningitis, and sepsis [8]. Thevast majority of its victims come from the world’s poorestcountries, and half of them are children younger than 5 years.It is extraordinary, in view of these facts, that pneumococcaldisease remains a relatively unknown disease and does not havea higher place on the agenda of the international community [3].Compared with other diseases affecting the developing world,determining the incidence of pneumococcal disease is relativelydifficult [9]. This is due to a number of factors including thedifficulties involved in stringent laboratory testing and samplecollection and the unavailability of quality surveillance data indeveloping countries. This lack of epidemiological evidence haslikely contributed to a gross underappreciation of the economic,clinical, and human burdens imposed by pneumococcal diseaseand hindered public health planning and decision making indeveloping countries [3]. Most of the researchers have done thereview of invasive pneumococcal disease (IPD) for patientsbelonging to a particular age group, for example, pediatric agegroup, and limited to a single region. We extended our review of






* Address correspondence to: Namaitijiang Maimaiti, International Institute for Global Health, United Nations University, UKMMC, JalanYaacob Latif, Bandar tunrazak, Cheras 56000, Kuala Lumpur, Malaysia.


dx.doi.org/10.1016/j.vhri.2013.07.003

dx.doi.org/10.1016/j.vhri.2013.07.003

dx.doi.org/10.1016/j.vhri.2013.07.003


dx.doi.org/10.1016/j.vhri.2013.07.003

IPD to all age groups on the basis of availability of data andextended to five different regions in the world. The developingcountries are defined as a nation with a low living standard,underdeveloped industrial base, and low human developmentindex relative to other countries [10]. This will help us betterunderstand the disease burden in all the age groups in selecteddeveloping countries in different parts of the world. We meas-ured the clinical burden of IPD in selected developing countriesby geographical and economical point of view. Clinical burden isa burden that differs from genetic burden mainly in the addedcomponent of morbidity; a trait that is either clinically orgenetically lethal may be grossly disabling [11].

Methods

We conducted a detailed review of published articles on IPD.Literature searches were conducted by using the PubMed data-base, Google scholar, and The Lancet, and were limited to articleswritten in English. We selected 10 developing countries from fiveregions, namely, East Asia (Malaysia and Thailand), South Asia(Bangladesh and Nepal), Middle East (Saudi Arabia and Qatar),sub-Saharan Africa (The Gambia and Mozambique), and LatinAmerica (Cuba and Peru). The World Bank categorizes thecountries by income into five different categories: low-incomeeconomics, lower-middle-income economics, upper-middle-income economics, high-income economics, and high-incomeOrganisation for Economic Co-operation and Development mem-bers [12]. The main criterion for the selection of these countriesfrom the designated regions was that they should belong to sameeconomic status (as defined by the World Bank) and preferably beneighbors. When selected countries did not have an equivalentneighbor, we selected the next nearest country in the region withthe same economic status.

The search term combinations used to search the knowledgebase included pneumococcal pneumonia, epidemiology, inci-dence rate, clinical burden AND pneumococcal meningitis, epi-demiology, incidence rate, clinical burden AND pneumococcalsepticemia, epidemiology, incidence rate, and clinical burden. Westarted with reviewing the abstracts of these articles publishedbetween the year 2000 and 2010 to find out which of the studiesmet our inclusion criteria and then reviewed only those full-length articles that complied with our inclusion criteria. Theknowledge base searches were conducted from March 1, 2011, tillJuly 31, 2011. We included only published articles and articlesthat described the clinical burden of pneumococcal disease with

the quantitative data of the selected developing countries in thedifferent regions. If there was more than one article in eachcountry, the most comparable study or the study with a highquality of methodology was reviewed. We excluded case reports,reports, and special populations, such as reports for patients withHIV/AIDS. We also excluded articles published before 2000 andafter 2010, or articles from developed countries. We acceptedeach author’s definition and methods except for the exclusioncriteria mentioned.

Two individuals (Namaitijiang Maimaiti and Zafar Ahmed[NM and ZA]) independently screened the titles and abstracts ofeach citation retrieved from the search term and identified thearticles for full review. The decision to review an article in detailwas based on the content of the abstract and whether the articledescribed the clinical burden of pneumococcal disease. If thereviewed article included both clinical and economic burdens ofpneumococcal disease, we still included this article in our study,but reviewed only the clinical burden of disease for our study.

Initially, the search process using the search term yielded 51articles. Finally, 22 articles remained to be reviewed in detail. Ifmore than one article from the same country was identified thatdescribed the clinical burden of pneumococcal disease, we selectedthe article with the later publication date with same quality orbetter quality methods compared with previous studies. Thedetailed review of the selected 22 articles was carried out by MNand ZA in a team, and finally they selected 10 articles that fulfilledall the selection criteria. Out of these 10 selected articles, 8 werefull-text original publications and 2 were review articles, and the 10articles selected were then used for our analysis. We gatheredinformation such as study design (retrospective/prospective study),study period, study location (urban/rural), study population andage of sample, incidence rate, and case-fatality ratio (CFR) (Fig. 1).

Results

The list of the selected articles from 10 developing countriesbelonging to five different regions is presented in Table 1.

Among the 10 articles selected, 8 articles had a detaileddiscussion on IPD, 1 discussed meningitis, and 1 discussedpneumococcal bacteremia. Out of the 10 articles selected, only 5articles described the CFR (Table 2). The studies included in thisreview were carried out in a variety of settings. Seven studieswere carried out in an urban seating; two studies were carried outin rural areas, whereas only one study was carried out in bothurban and rural settings.

51 were identified

22 were read and abstracted 29 from full text did not meet inclusion criteria, published before 2000 or after 2010, duplicate reference, repeated data or poor quality.

10 Full text article, review

Fig. 1 – Flow diagram for the process of review of the literature. *HI, high income; LI, low income; UMI, upper-middle income.


Table 3 shows that the IPD incidence ranged from less than 5per 100,000 to 416 per 100,000. In general, the incidence rate ofIPD varies directly with the respective country’s income category,with higher income countries having a low incidence rate andlow-income countries having a high incidence rate. The onlycountry that shows an exception is the Saudi Arabia, where theoverall incidence rate for IPD (bacteremia and meningitis) is 17.4per 100,000 population; if we look at the incidence rate formeningitis alone, 4/100,000, it is higher than that in Qatar, 2.4/100,000.

Figure 2 also shows that the incidence rate in low-incomecountries is way higher than the incidence rate in countriesamong other income categories. We selected low-income coun-tries from two different geographic regions. The data show thatthe incidence rate of IPD in the low-income countries of sub-Saharan Africa (Gambia and Mozambique) is considerably higherthan the incidence rate of IPD in the low-income countries fromSouth Asia (Bangladesh and Nepal).

When we look at the incidence rate of IPD according to thegeographic region, we found that the incidence rate in East Asianand Latin American countries was not very different from eachother—less than 10 per 100,000 for both the regions. The inci-dence rate of IPD in middle eastern countries was slightly higherthan that in East Asian and Latin American countries, but it wasstill less than 20 per 100,000.

Among the five regions, the CFR was reported in four regions—East Asia, South Asia, Middle East, and Latin America. Based onavailable data, the highest CFR was recorded in South Asia (80%),followed by East Asia (33.3%), Latin America (27%), and MiddleEast (12.2%). Concluding from these published evidences, we cansay that the incidence rate is high in low-income developingcountries such as Mozambique, Gambia, Bangladesh, and Nepal,and this could be due to inequality, inequitable distribution ofwealth, inadequate provision of health care services, lack ofhuman resource and human capacity to manage IPD effectively,and also lack of latest technology in the health care system.

Table 1 – List of the 10 selected articles and data of publication.

No. Title of selected articles Year ofpublication

Reference

1 Epidemiology of invasive pneumococcal infection in children aged five years and under in SaudiArabia: a five-year retrospective surveillance study

2010 [13]

2 Acute bacterial meningitis in Qatar 2006 [14]3 Invasive pneumococcal disease among children in rural Bangladesh: results from a population-

based surveillance2009 [15]

4 Invasive pneumococcal disease in Kanti Children’s Hospital, Nepal, as observed by the SouthAsian Pneumococcal Alliance Network

2009 [16]

5 Invasive pneumococcal diseases among hospitalized children in Lima, Peru 2010 [17]6 Bacterial meningitis in children and adolescents: an observational study based on the national

surveillance system2005 [18]

7 Invasive pneumococcal disease in children o5 years of age in rural Mozambique 2006 [19]8 Efficacy of nine-valent pneumococcal conjugate vaccine against pneumonia and invasive

pneumococcal disease in The Gambia: randomised, double-blind, placebo-controlled trial2006 [20]

9 Overview of the disease burden of invasive pneumococcal disease in Asia 2010 [21]10 Incidence of pneumococcal bacteremia requiring hospitalization in rural Thailand 2009 [22]

Table 2 – Summary of epidemiology of invasive pneumococcal infections in selected developing countries bygeographic area and per-capital income.

Region Country Countrycategory

Study design Study period Studylocation

Studypopulation

Age ofsample

East Asia Malaysia Upper-middleincome

Retrospectivestudy

2004–2006 Urban 216 o5 y

Thailand Upper-middleincome

Population-basedsurveillance

2005–2009 Rural 23,853 All agegroups

South Asia Bangladesh Low-income Population-basedsurveillance

2004–2007 Rural 22,378 o5 y

Nepal Low- income Retrospective study 2004–2007 Urban 2,528 o5 yMiddle east Saudi Arabia High- income Retrospective study 1999–2003 Urban 82 o5 y

Qatar High- income Retrospective study 1998–2002 Urban 64 o12 ySub-Saharan

AfricaMozambique Low-income Prospective

surveillance2001–2003 Urban 10,702 o5 y

Gambia Low-income Clinics: controls in avaccine trial

2000–2003 Urban 17,437 o3 y

LatinAmerica

Peru Upper-middleincome

Retrospective study 2006–2008 Urban 101 o5 y

Cuba Upper-middleincome

Retrospective study 1998–2003 Both 1,023 o18 y

CFR, case-fatality ratio; IPD, invasive pneumococcal disease; IR, incidence rate; PP, pneumococcal pneumonia.


When we look at the CFR, it was high in low-income countries(80%) and upper-middle-income countries (33.3%, 31.7%, and 22%)than in high-income countries (12.2%). Even though the incidencerate was higher in high-income countries than in upper-middle-income countries, the CFR was low in these high-income coun-tries than in middle-income countries. This can be explainedwith the argument that generally high-income countries havegood capacity to provide excellent quality health care to theirpopulation by using latest technology and hire properly qualifiedprofessionals to operate health facilities.

Discussion

The clinical burden of the disease was slightly different amongthe neighboring countries within the same region in this review.These differences in the disease burden between neighboringcountries even though they belong to the same economic levelmay be due to differences in the health care system, humanresource, budget for health care expenditures, health insurancecoverage, climate, and so on. Collectively, these differences canalso be due to the objectives of the study, study design, differentsample demographic, location, and other factors. One of thelimitations of the study is that only two countries were selectedfrom each region.

We recommend that future studies should be carried out withlarge sample size, be conducted in both urban and rural areasamong all age groups, and should report the CFR.

This study included five different regions covering both urbanand rural settings and critical morbidity. In this study, most ofthe articles covered all age groups. A limitation of this study wasthat it was conducted only in 10 developing countries and theresult may not reflect the situation in other developing countries.

Conclusions

Our review confirmed that the clinical burden of IPD was high indeveloping countries. The incidence rate and the CFR of IPD weredifferent from region to region and country to country. Thedisease burden was high in sub-Saharan African countriesfollowed by South Asian countries. The CFR was low in high-income countries than in low-income countries. The diseaseburden was not significantly different among the countries thatbelong to the category of upper-middle-income countries. Thereview found that generally the disease burden was high in ruralareas than in urban areas, which can be explained by the inequityin health care facilities available in rural areas. Studies fromMiddle Eastern countries and Peru had a very small sample size;therefore, because of the small sample size, these studies may

Table 3 – The IR and CFR of pneumococcal disease by region, country, and economic level of countries

Region Country category Country IR of IPD/100,000 CFR

Middle eastHigh-income Saudi Arabia 17.4 12.20%High-income Qatar 2.4� NA

East AsiaUpper-middle income Malaysia 8.6� 33.30%Upper-middle income Thailand 5.7� NA

Latin AmericaUpper-middle income Peru 7.7 22%Upper-middle income Cuba 6 27%

South AsiaLow-income Bangladesh 86 NALow-income Nepal 52.4 80%

Sub-Saharan AfricaLow-income Mozambique 416 NALow-income Gambia 388 NA

CFR, case-fatality ratio; IPD, invasive pneumococcal disease; IR, incidence rate; NA, not applicable/available.� Single disease IR.

Fig. 2 – Incidence rate of IPD by income. IPD, invasive pneumococcal disease.


not reflect the true disease burden of these countries. Only threestudies were conducted among all age groups, and the otherstudies were carried out only among children.

The WHO reported that 1.6 million people die every yearfrom pneumococcal disease, with 0.7 million to 1 million ofthese being children younger than 5 years. The pneumococcaldisease burden among adults is also high, with an estimated600, 000 to 800, 000 adult deaths each year from pneumococcalpneumonia, meningitis, and sepsis [7]. The review done is inaccordance with the WHO report that says that the pneumo-coccal disease burden is high among children as well as amongadult populations.

Source of financial support: This research project was fundedby GlaxoSmithKline SdnBhd.

R E F E R E N C E S

[1] WHO position paper: pneumococcal vaccines. Wkly Epidemiol Rec1999;74(23):177–83.

[2] O’Brien KL, Wolfson LJ, Watt JP, et al. Burden of disease caused byStreptococcus pneumoniae in children younger than 5 years: globalestimates. Lancet 2009;374:893–902.

[3] Preventing pneumococcal disease: report from the All-Party ParliamentaryGroup on Pneumococcal Disease Prevention in the Developing World.2008. Available from: www.appg-preventpneumo.org.uk.

[4] World Health Organization. Pneumococcal vaccines. Wkly EpidemiolRec 2003;14:110–9.

[5] Goetghebuer T, West TE, Wermenbol V, et al. Outcome of meningitiscaused by Streptococcus pneumoniae and Haemophilus influenza type b inchildren in The Gambia. Trop Med Int Health 2000;5:207–13.

[6] World Health Organization. Pneumococcal conjugate vaccine forchildhood immunization: WHO position paper. Wkly Epidemiol Rec2007;82(12):93–104.

[7] World Health Organization. WHO Report: Bi-regional Meeting onPrevention of Childhood Pneumonia and Meningitis by Vaccination,Kuala Lumpur, Malaysia, 30–31 March 2006. Manila, Philippines:World Health Organization Regional Office for the Western Pacific,2006.

[8] Chun first initial, 2 more authors, et al. Estimation of otitis mediadisease and cost burden in Korea. Presented at 12th Western PacificCongress on Chemotherapy and Infectious Diseases [WPCCID], 2010.

[9] Madhi SA, Kuwanda L, Cutland C, et al. The impact of a 9-valentpneumococcal conjugate vaccine on the public health burden ofpneumonia in HIV-infected and -uninfected children. Clin Infect Dis2005;40:1511–8.

[10] Sullivan A, Sheffrin SM. Economics: Principles in Action. Upper SaddleRiver, NJ: Pearson Prentice-Hall, 2003.

[11] Drugs.com. Clinical burden. Available from: http://www.drugs.com/dict/clinical-burden.html.

[12] The World Bank. Country and landing groups. Available from: http://data.worldbank.org/about/country-classifications/country-and-lend-ing-groups#East_Asia_and_Pacific.

[13] Memish ZA, El-Saed A, Al-Otaibi B, et al. Epidemiology of invasivepneumococcal infection in children aged five years and under in SaudiArabia: a five-year retrospective surveillance study. Int J Infect Dis2010;14(8):e708–12.

[14] Elsaid MF, Flamerzi AA, Bessisso MS, Elshafie SS. Acute bacterialmeningitis in Qatar. Saudi Med J 2006;27:204–10.

[15] Arifeen SE, Saha SK, Rahman S, et al. Invasive pneumococcal diseaseamong children in rural Bangladesh: results from a population-basedsurveillance. Clin Infect Dis 2009;48(Suppl. 2):S103–13.

[16] Shah AS, Knoll MD, Sharma PR, et al. Invasive pneumococcal disease inKanti Children’s Hospital, Nepal, as observed by the South AsianPneumococcal Alliance Network. Clin Infect Dis 2009;48(Suppl. 2):S123–8.

[17] Ochoa TJ, Egoavil M, Castillo ME, et al. Invasive pneumococcal diseasesamong hospitalized children in Lima, Peru. Rev Panam Salud Publica2010;28:121–7.

[18] Dickinson FO, Pérez AE. Bacterial meningitis in children andadolescents: an observational study based on the national surveillancesystem. BMC Infect Dis 2005;5:103.

[19] Roca A, Sigaúque B, Quintó L, et al. Invasive pneumococcal disease inchildreno5 years of age in rural Mozambique. Trop Med Int Health2006;11:1422–31.

[20] Cutts FT, Zaman SM, Enwere G, et al. Efficacy of nine-valentpneumococcal conjugate vaccine against pneumonia and invasivepneumococcal disease in The Gambia: randomised, double-blind,placebo-controlled trial. Lancet 2005;365(9465):1139–46.

[21] Bravo LC. Overview of the disease burden of invasive pneumococcaldiseases in Asia. Vaccine 2009;27(52):7282–91.

[22] Baggett HC, Peruski LF, Olsen SJ, et al. Incidence of pneumococcalbacteremia requiring hospitalization in rural Thailand. Clin Infect Dis2009;48(Suppl. 2):S65–74.









www.appg-preventpneumo.org.uk























http://www.drugs.com/dict/clinical-burden.html

http://www.drugs.com/dict/clinical-burden.html

http://data.worldbank.org/about/country-classifications/country-and-lending-groups#East_Asia_and_Pacific



































HEALTH POLICY ANALYSIS

Capacity Building for HTA Implementation in Middle-Income Countries:The Case of HungaryZoltán Kaló, MSc, MD, PhD1,2,�, József Bodrogi, MSc3, Imre Boncz, MSc, MD, PhD4,5, Csaba Dózsa, MSc, PhD6,Gabriella Jóna, MD, MSc7, Rita Kövi, MD7, Zsolt Pásztélyi, MD, MSc8, Balázs Sinkovits, MSc9, on behalf of ISPOR Hungary Chapter1Health Economics Research Centre, Eötvös Loránd University (ELTE), Budapest, Hungary; 2Syreon Research Institute, Budapest, Hungary; 3Independent SeniorHealth Economist, Budapest, Hungary; 4Faculty of Health Sciences, Institute for Health Insurance, University of Pécs, Pécs, Hungary; 5Faculty of Economics,Health Economics and Health Technology Assessment Research Centre, Corvinus University of Budapest, Budapest, Hungary; 6Health Care Faculty, University ofMiskolc, Miskolc, Hungary; 7National Institute for Quality- and Organizational Development in Healthcare and Medicines, Budapest, Hungary; 8Railway HealthServices, Budapest, Hungary; 9AstraZeneca, Budapest, Hungary

A B S T R A C T

Objectives: Middle-income countries often have no clear roadmap forimplementation of health technology assessment (HTA) in policydecisions. Examples from high-income countries may not be relevant,as lower income countries cannot allocate so much financial andhuman resources for substantiating policy decisions with evidence.Therefore, HTA implementation roadmaps from other smaller-size,lower-income countries can be more relevant examples for countrieswith similar cultural environment and economic status. Methods: Wereviewed the capacity building process for HTA implementation inHungary with special focus on the role of ISPOR Hungary Chapter.Results: HTA implementation in Hungary started with capacity build-ing at universities with the support of the World Bank in the mid 90's,followed by the publication of methodological guidelines for conduct-ing health economic evaluations in 2002. The Hungarian HealthEconomics Association (META) - established in 2003 - has beenrecognized as a driving force of HTA implementation. META becamethe official regional ISPOR Chapter of Hungary in 2007. In 2004 the

National Health Insurance Fund Administration made the cost-effec-tiveness and budget impact criteria compulsory prior to grantingreimbursement to new pharmaceuticals. An Office of Health Tech-nology Assessment was established for the critical appraisal ofeconomic evaluations submitted by pharmaceutical manufacturers.In 2010 multicriteria decision analysis was introduced for newhospital technologies. Conclusion: The economic crisis may createan opportunity to further strengthen the evidence base of health caredecision-making in Hungary. In the forthcoming period ISPOR Hun-gary Chapter may play an even more crucial role in improving thestandards of HTA implementation and facilitating international col-laboration with other CEE countries.Keywords: capacity building, HTA implementation, ISPOR HungaryChapter, middle-income countries.


Introduction

The health status of the population in middle-income countriesis usually worse than in high-income countries. Because healthcare resources are scarcer in these countries, the societal cost ofinappropriate pricing and reimbursement decisions of newhealth care technologies is even higher. Implementation ofhealth technology assessment (HTA) in the decision-makingprocess may alleviate this problem. Key success factors for HTAimplementation, however, are building human resource andfinancial capacities, establishing a transparent decision-makingprocess, and implementing robust HTA methodology.

Middle-income countries often have no clear roadmap for HTAimplementation. Examples from high-income and resource-rich

countries such as the United Kingdom (i.e., the “National Institutefor Health and Clinical Excellence experience”) may not be relevantbecause lower income countries cannot allocate so many financialand human resources for substantiating health policy decisionswith evidence. In addition, the size of the country matters; thesmaller a country is the more limited facilities for preparing fullHTA reports it has. Therefore, HTA implementation roadmapsfrom other smaller size, middle-income countries can be morerelevant, especially if the country is from the same geographicaland cultural environment with similar economic status.

One of the most important questions of the HTA implemen-tation roadmap is whether capacity building should come firstor whether mandatory HTA requirement in the reimburse-ment process can induce the necessary background knowledge.




Conflict of interest: Authors have no conflicts of interest.


* Address correspondence to: Zoltán Kaló, Faculty of Social Sciences, Eötvös Loránd University, Pázmány Péter sétány 1/a, H-1117Budapest, Hungary.


dx.doi.org/10.1016/j.vhri.2013.06.002

dx.doi.org/10.1016/j.vhri.2013.06.002

dx.doi.org/10.1016/j.vhri.2013.06.002


dx.doi.org/10.1016/j.vhri.2013.06.002

Both options are feasible and have their advantages anddisadvantages.

Previous publications have already addressed major steps ofHTA implementation in Hungary [1–3]. Because HTA implemen-tation in Hungary started with capacity building, our objectivewas to summarize the role of International Society of Pharma-coeconomics and Outcomes Research (ISPOR) Hungary Chapter inhuman resource development. The Hungarian example for HTAcapacity building can be especially relevant for other CentralEastern European countries.

Hungary has 9.94 million inhabitants, and its gross domesticproduct per capita was €10,016 in 2011 [4]; therefore, the countryhas passed theWorld Bank’s threshold of high-income countries. Amajor part of the HTA implementation, however, happened in theyears when Hungary was classified as a middle-income country.

Step 1: HTA Capacity Building

In the mid-1990s, two new academic centers, the School of PublicHealth at the University of Debrecen and the Health CareManagement Training Center at Semmelweis University, wereestablished. Tutors received World Bank scholarships to studyhealth economics, health care management, epidemiology, bio-statistics, public health, and other related sciences in distin-guished international academic centers. Many graduates of theWorld Bank program left the country, but many of them stayed oreven returned after some years of international experience.

Smaller scale educational projects also contributed to HTAcapacity building. The TUDOR program at the University of Szegedwas established to facilitate the application of evidence-basedmedicine in Hungary. The program was sponsored by the BritishDepartment for International Development, Know How Fund [5].

By 2000, the number of trained professionals with a thoroughunderstanding of HTA reached 50.

Several Hungarian universities (Corvinus University of Buda-pest, University of Debrecen, University of Pécs, and University ofWest Hungary) introduced training programs in economic eval-uation of medical technologies for undergraduate students. Thefirst postgraduate course with a major focus on economicevaluation and economic modeling was introduced by the Fac-ulty of Social Sciences at the Eötvös Loránd University in 2007.

By 2010, the number of trained professionals with personalexperience in HTA research or appraisal exceeded 200.

Step 2: Methodological Guidelines

Standardization of economic evaluations is a necessity prior tomandating the use of economic evaluation in policy decisions.The Hungarian methodological guidelines for conducting eco-nomic evaluation of health care interventions were published in2002 [6]. These guidelines covered all health care interventions;therefore, they were not specific for pharmaceuticals and notlimited to reimbursement questions. The intention was to updatethe guidelines every 2 years; however, the Ministry of Health didnot implement any revision before 2013.

Step 3: Scientific Organization

The Hungarian Health Economics Association (Magyar Egészség-gazdaságtani Társaság [META]) was established in 2003. Thefounders aimed at establishing an independent organization todiscuss major health economic and health policy issues atmonthly meetings. Since 2003 META has been organizing 8 to10 meetings a year. Each 2-hour monthly meeting is dedicated toa particular research, policy, or methodological topic, with anadvocate (or researcher), an opponent, and a moderator. Pricingand reimbursement policy of new health technologies and

methodological standards of HTA research have been discussedat several meetings over the years.

In 2006, META was one of the main organizers of the highlysuccessful 6th European Congress of Health Economics in Buda-pest [2]. In 2007, META became the official Hungarian Chapter ofthe ISPOR. ISPOR Hungary Chapter has a strong commitment tofacilitate international collaboration with other Central EasternEuropean countries.

In 2007, ISPOR Hungary Chapter organized its first 1-day annualnational congress in health economics. Since 2010, the 2-day con-gress with more than 200 participants has an international plenarysession with invited speakers from other ISPOR regional chapters.

The achievements of the ISPOR Hungary Chapter have beenrecognized by policymakers. Several former ministers and statesecretaries gave lectures at the annual health economics congress.Moreover, in 2010, the State Secretary of Health invited META toestablish the Management and Health Economics Section of theProfessional Health Care College (the Advisory Board of the HealthCare Secretariat of the Ministry of Human Resources). META alsogained official recognition on behalf of other professional medicalsocieties by joining the Association of Hungarian Medical Society(Magyar Orvostársaságok és Egyesületek Szövetsége [MOTESZ]).

In 2012, ISPOR Hungary Chapter and Eötvös Loránd Universitylaunched a 1-week summer university course with the title of“Implementation of HTA in CEE countries,” attended by partic-ipants from 10 countries.

In 2013, ISPOR Hungary Chapter has 114 members, with 350professionals visiting local meetings in 2012.

Step 4: Compulsory HTA in Policy Decisions

In 2004, the Hungarian National Health Insurance Fund Admin-istration made the cost-effectiveness and budget impact criteriacompulsory prior to granting reimbursement to new pharmaceut-icals. The Ministry of Health established a Department of HealthTechnology Assessment (HTA Department) at one of its back-ground institutes for the critical appraisal of economic evaluationssubmitted by pharmaceutical manufacturers. The summaryappraisal prepared by the HTA Department is taken into accountby the National Health Insurance Fund Administration in reim-bursement decisions. The single HTA process for patented out-patient pharmaceutical products is described in Fig. 1.

Experience of the first 6 years of the Hungarian fourth hurdleindicated that the quality of economic evaluations submitted inreimbursement dossiers was rather heterogeneous. In 2009, aworking group was set up to develop a policy-relevant, detailedHungarian critical appraisal checklist to improve the quality ofpharmacoeconomic evaluations submitted for single HTA inpharmaceutical reimbursement applications. The criticalappraisal checklist has been published recently [7].

In 2010, new multicriteria decision analysis was introducedfor new hospital technologies, mainly for medical devices.

Recent Activities and Further Steps

ISPOR Hungary Chapter has been growing continuously. Thesociety has a balanced membership structure from the academia,public, and private sectors. It has a Young Professional Unit withan age limit of 35 years. Young professionals are especially activein driving changes by establishing working groups in severalpolicy and research areas. The chapter established a student unitfor graduate, postgraduate, and PhD students in 2012.

ISPOR Hungary Chapter has coordinated the revision ofmethodological guidelines for conducting economic evaluationof health care interventions. The revised guidelines were pub-lished in March 2013. The new guidelines indicate explicit cost-effectiveness thresholds for Hungary (two to three times of gross


domestic product per capita). According to the guidelines, thesethresholds are considered as a tool (reference point), and so theyare not used as mandatory criterion for reimbursement. In thenew guidelines, the discount rate for both costs and health gainshas been reduced from 5% to 3.7%.

Yet, health policy and major reimbursement decisions are stillnot fully transparent in Hungary. The National Health InsuranceFund Administration has implemented serious cost-containmentmeasures for pharmaceuticals in recent years; therefore, budgetimpact has become the most important element for reimburse-ment decisions with mandatory financial risk-sharing agreements.

The Young Professional Unit of ISPOR Chapter Hungary hasestablished a working group to develop proposals for improvingthe transparency and evidence base of pharmaceutical pricingand reimbursement decisions.

Conclusions

The global economic crisis significantly influenced the Hungarianeconomy. Public health care resources are highly limited, andthey are not sufficient to maintain the current health careinfrastructure and the publicly funded benefit package; therefore,health care financing and provision have to be restructured.There is growing pressure on policymakers to justify their majorpolicy decisions. The economic crisis may create an opportunityto strengthen the evidence base of health care decision making inHungary [8].

ISPOR Hungary Chapter has played a crucial role in HTAimplementation. In the forthcoming period, the chapter will be most

likely to play an evenmore crucial role in improving the standards ofHTA implementation and facilitating international collaboration withother countries in the Central Eastern European region.

Source of funding: The authors have no other financialrelationships to disclose.

R E F E R E N C E S

[1] Gulácsi L, Boncz I, Drummond M. Issues for countries consideringintroducing the “fourth hurdle”: the case of Hungary. Int J TechnolAssess Health Care 2004;20:337–41.

[2] Boncz I, Dózsa C, Kaló Z, et al. Development of health economics inHungary between 1990-2006. Eur J Health Econ 2006;7(Suppl.1):4–6.

[3] Gulácsi L, Orlewska E, Péntek M. Health economics and healthtechnology assessment in Central and Eastern Europe: a dose of reality.Eur J Health Econ 2012;13:525–31.

[4] Central Statistical Office of Hungary. Per capita gross domestic product(GDP) (1995-), 2012. Available from: http://www.ksh.hu/docs/eng/xstadat/xstadat_annual/i_qpt016.html. [Accessed November 18, 2012].

[5] EBM TUDOR 2012. Available from: http://tudor.szote.u-szeged.hu/webeng/what/whatbe.php. [Accessed November 18, 2012].

[6] Szende Á, Zs Mogyorosy, Muszbek N, et al. Methodological guidelines forconducting economic evaluation of healthcare interventions inHungary: a Hungarian proposal for methodology standards. Eur J HealthEcon 2002;3:196–206.

[7] Inotai A, Pékli M, Jóna G, et al. Attempt to increase the transparency offourth hurdle implementation in Central-Eastern European middleincome countries: publication of the critical appraisal methodology.BMC Health Serv Res 2012;21:332.

[8] Kaló Z, Boncz I, Inotai A. Implications of economic crisis on health caredecision-making in Hungary: an opportunity to change? J Health PolOutcomes Res 2012;1:20–6.

Fig. 1 – Pricing and reimbursement process with single HTA for patented outpatient pharmaceutical products. HTA, healthtechnology assessment.











http://www.ksh.hu/docs/eng/xstadat/xstadat_annual/i_qpt016.html

http://www.ksh.hu/docs/eng/xstadat/xstadat_annual/i_qpt016.html

http://tudor.szote.u-szeged.hu/webeng/what/whatbe.php

http://tudor.szote.u-szeged.hu/webeng/what/whatbe.php














What Influences Recommendations Issued by the Agency for HealthTechnology Assessment in Poland? A Glimpse Into DecisionMakers’ PreferencesMaciej Niewada, MD, PhD1,2,�, Małgorzata Polkowska, MS2, Michał Jakubczyk, PhD2,3, Dominik Golicki, MD, PhD1,2

1Department of Experimental and Clinical Pharmacology, Medical University of Warsaw, Warsaw, Poland; 2HealthQuest Consulting Company, Warsaw, Poland;3Institute of Econometrics, Warsaw School of Economics, Warsaw, Poland

A B S T R A C T

Objective: This study aimed to evaluate the factors that are associ-ated with positive (supporting public funding) and negative recom-mendations of the Agency for Health Technology Assessment inPoland. Methods: Two independent analysts reviewed all the recom-mendations publicly available online before October 7, 2011. For eachrecommendation, predefined decision rationales, that is, clinicalefficacy, safety, cost-effectiveness, and formal aspects, were sought,either advocating or discouraging the public financing. In the analysis,we used descriptive statistics and a logistic regression model so as toidentify the association between predefined criteria and the recom-mendation being positive. Results: We identified 344 recommenda-tions—218 positive (62.8%) and 126 negative (37.2%). Negativerecommendations were better justified and also the comments wereless ambiguous in accordance with the recommendation (except forclinical efficacy). In general, the specified criteria supported thedecision (either positive or negative) in 209 (60.8%), 107 (31.1%), 124(36.0%), 96 (27.9%), and 61 (17.7%) recommendations, respectively, andran contrary to the actual decision in the remaining ones. Threshold

values for either cost-effectiveness or budget impact distinguishingpositive from negative recommendations could not be specified. Thefollowing parameters reached statistical significance in logistic regres-sion: clinical efficacy (both explicitly positive and explicitly negativeevaluations impacted in opposite directions), lack of impact on hardend points, unfavorable safety profile, cost-effectiveness results, andformal shortcomings (all reduced the probability of a positive recom-mendation). Conclusions: Decision making of the Agency for HealthTechnology Assessment in Poland is multicriterial, and its resultscannot be easily decomposed into simple associations or easilypredicted. Still, efficacy and safety seem to contribute most to finalrecommendations.

Keywords: decision making, health technology assessment,incremental cost-effectiveness ratio, reimbursement.


Introduction

Health technology assessment (HTA) agencies play a vital role inthe decision-making process, whether or not to reimburse givenhealth technologies. These agencies are expected to be guided bymedical, economic, and ethical criteria and to account for limitedresources and sometimes limited evidence regarding the profileof assessed technologies. Therefore, there are many possibledrivers for the final decision.

The aim of the current scientific project was in general todetect the criteria that can be considered important for theAgency for Health Technology Assessment in Poland (AHTAPol),and in particular to try to find the characteristics of HTA reportsthat are associated with positive and negative recommendations.

The AHTAPol was established in 2005 by the Ministry ofHealth as a first of its kind of institution in Central and EasternEurope. Since 2009, the AHTAPol is defined as a legal and

independent entity playing a key role in reimbursement decisionmaking. The most important role of the AHTAPol is to preparerecommendations for and support decision making by the Min-istry of Health on financing health care services from the publicbudget. The AHTAPol assesses and appraises all medical tech-nologies, drugs, devices, and other services that are claimingpublic funding. The role of the AHTAPol covers the assessmentand appraisal of the HTA reports including systematic review ofclinical findings, economic evaluation, and budget impact anal-ysis, majority of which are submitted by the pharmaceuticalindustry. Assessment is provided by a team of analysts and basedon the Polish HTA guidelines (first issued in 2007 and reviewed in2009) [1]. Appraisal is completed by the Consultative Council(transformed into the Transparency Council with the beginningof 2012), a team of highly qualified and experienced specialists,and the president of the AHTAPol. Final judgment is made in thespecific context of the alternative options available, social






� Address correspondence to: Maciej Niewada, Department of Experimental and Clinical Pharmacology, Medical University of Warsaw,Krakowskie Przedmieście 26/28, 00-927 Warsaw, Poland


dx.doi.org/10.1016/j.vhri.2013.05.002

dx.doi.org/10.1016/j.vhri.2013.05.002

dx.doi.org/10.1016/j.vhri.2013.05.002


dx.doi.org/10.1016/j.vhri.2013.05.002

consequences, health care delivery organization implications,national health priorities, and social and ethical aspects. Basedon AHTAPol recommendations, reimbursement decisions aremade by the Ministry of Health following negotiations withpharmaceutical industry representatives.

Recommendations issued by the AHTAPol have evolved overtime. The new types of recommendations (i.e., conditional,temporal, combined, and others [2]) were introduced. Legal back-ground of recommendations has also changed and currentlystatements by the Consultative Council and final recommenda-tions by the President of the AHTAPol are issued [3].

The current article can be located in the line of researchestablished by a classical article of Devlin and Parkin [4] and astudy by Towse [5]. Devlin and Parker analyzed past decisionsmade by the National Institute for Health and Clinical Excellence(NICE) in the United Kingdom to determine factors that wereassociated with positive decisions, and in particular the thresh-old level for the incremental cost-effectiveness ratio (ICER). Intheir study, they managed to detect that the threshold levelprobably lies above approximately 35,000 GBP. The study ofDevlin and Parker motivated subsequent articles. And so,Tappenden et al. [6] tried to identify the preferences of themembers of NICE Appraisal Committees by using aquestionnaire-based study. They concentrated more on theethical issues, that is, on the impact of such variables as baselinequality of life or age of the beneficiaries. Dakin et al. [7]introduced multinomial approach to these kinds of studies,accounting for conditional approval by NICE.

Methods

Material

The analysis covered all recommendations and statements of theConsultative Council of the AHTAPol issued following two sepa-rate regulations (the Ordinance of the Minister of Health datedSeptember 10, 2009, and the Act on Healthcare Services FinancingFrom Public Funds) and available on the official Web site of theagency (http://www.aotm.gov.pl) before October 7, 2011. It may besomewhat misleading that we call “a recommendation” both thetext published by the AHTAPol and the final conclusion thereof.We do not call the latter “a decision” because this is made only bythe Ministry of Health and need not agree with the AHTAPolrecommendation. At the same time, we decided to analyzerecommendations, not decisions, because the decisions are notaccompanied by any justifications and thus would be difficult tospot any regularities.

Inclusion and exclusion criteriaAll recommendations and statements were included with theexception of collective recommendations for dental interventions(covering not a single technology but a group of technologies).Some recommendations were explained either poorly or not at all—then the recommendation was excluded altogether.

Analysis—Data Extraction and Interpretation

Only recommendations’ texts were analyzed, neither HTA reportsnor critical appraisal, which in most cases were not available onthe official Web site of the AHTAPol. For every recommendation,the following data were extracted: medical technology beingevaluated, medical therapeutic area in which the technologyreimbursement was appealed, and the year of issuing therecommendation. Different types of AHTAPol recommendations(e.g., supporting or rejecting funding, conditional, temporary,and combined) were redefined into statements of limited orno financing technology (negative recommendations) or onessupporting financing or increase in funding (positive recom-mendations).

Each recommendation was evaluated independently by tworesearchers by using predefined criteria listed below (languagespecialist with experience in auditing of HTA reports and HTAspecialist). Disagreements were resolved by discussion. For everyrecommendation (positive or negative), it was classified whetherthe final recommendation was supported or discouraged by eachcriterion. Table 1 presents the data interpretation. Consistencywas found if for a particular criterion positive and negativefindings were reported and explicitly referred to support positiveand negative recommendations, respectively. For other situa-tions, we interpreted the criterion as not reflected in the finaljudgment. Following pilot analysis of the AHTAPol recommen-dations, clinical, economic, and formal criteria used to judge finalstatements were distinguished. In a few recommendations,rationales used to judge final statement could not be classifiedas the above-listed criteria and were not defined separately.

Clinical criteriaThe importance of general relative efficacy and safety overcomparators in decisions’ reasoning by the Consultative Councilwas recognized in pilot analysis. Thus, the clinical criteria werefurther split into three subcriteria: the efficacy (benefit over thecomparator used in the analysis—an active treatment or pla-cebo), safety, and the impact of the technology on clinical hardend points (which were treated separately as anticipated signifi-cant driver of clinical decision making). Hard end points weredefined following reviewed Polish guidelines for HTA as clinicallysignificant end points, playing an important role in a givendisease, that is, deaths, cases or recoveries, quality of life, adverseeffects (divided into serious and nonserious), or medical events[1]. The issue of difference between the efficacy, studied inclinical trials, and effectiveness, observed in real life settings,was not taken up explicitly in any recommendation; thus, it wasnot addressed in our analysis.

Economic criteriaEconomic criteria were also further split into two subcriteria:cost-effectiveness and the impact on the payer’s budget. Theevaluation of the technology’s cost-effectiveness significance wasbased on values for cost per quality-adjusted life-year (QALY)or life-year gained (LYG) reported in the recommendationsand assumed cost-effectiveness threshold of three timesthe gross domestic product per capita (∼83,239 PLN) [7]. The

Table 1 – Data interpretation for predefined criteria determining the AHTAPol recommendations.

Positive recommendations Negative recommendations

Criterion Positive data (consistentimpact on final judgment)

Negative data (notdriving final judgment)

Negative data (consistentimpact on final judgment)

Positive data (notdriving final judgment)

AHTAPol, Agency for Health Technology Assessment in Poland.


http://www.aotm.gov.pl

cost-effectiveness interpretation of technology profitability by theConsultative Council was analyzed independently of the statis-tical analysis of cost per QALY or LYG values reported inrecommendations.

The extent of the medical technology reimbursement impacton the payer’s budget and its influence on the recommendationswere also analyzed. Because there was no clear definition of alarge or small impact on the budget, we could extract only theConsultative Council perception and interpretation of budgetimpact data and analyze independently estimated annual meanpayer’s expenses reported in recommendations.

Formal criteriaAnother group of rationales recognized in pilot analysis, whichinfluenced the issuing of the positive or negative recommenda-tion by the AHTAPol Consultative Council, was of legal orprocedural nature and encompassed failure to meet require-ments of the AHTAPol for HTA reports or deficiencies of thereports submitted to the agency, submission of reports after thedeadline, the discrepancy between expert opinions and analysessubmitted by pharmaceutical industry, submission for unauthor-ized (off-label) use, and so forth.


We compared frequencies of values for binary criteria and meansfor continuous ones between positive and negative recommen-dations. Mean values were compared with unpaired t test withseparate variance estimation. We performed multivariate analy-sis by using logistic regression to determine the associationbetween criteria values and final recommendation. In thisapproach, we redefined the binary criteria to denote that a givencriterion is favorable or unfavorable for technology (rather thanagrees or not with the final recommendation). Because for manyrecommendations, there was neither a favorable nor an unfav-orable comment, for each criterion we introduced two binaryvariables that were used in econometric modeling—denoting,respectively, that a favorable, unfavorable comment is presentedin the recommendation text. Obviously, both these variablescould be equal to zero, but not to one for any single recommen-dation. While building the logistic regression model, we removedindividually variables that were not statistically significant(α ¼ 0.05).

Results

Three hundred forty-four recommendations were analyzed. Weidentified 218 positive (62.8%) and 126 negative (37.2%) recom-mendations. Eighty-nine recommendations addressed oncologi-cal technologies, mostly drugs. Other most commonrecommendations were identified for psychiatry, cardiology,neurology, rheumatology, and diabetes (32, 24, 20, 19, and 17recommendations, respectively). Only 14 recommendations werededicated to technologies authorized for rare diseases.

Clinical efficacy, impact of hard end points, safety, cost-effectiveness, and formal issues were explicitly discussed bythe Consultative Council in 238 (69.2%), 169 (49.1%), 155 (45.1%),140 (40.7%), and 47 (13.7%) recommendations, respectively. Fromthis perspective, clinical issues seem to be more important thaneconomic ones. Exact ICER values could be found for 106 (30.8%)recommendations, while exact budget impact estimates could befound for 193 (56.1%) recommendations.

In general, we observed that in the recommendations textsthe elements that support the final recommendation wereexplicitly stated more often. The clinical efficacy, impact of hardend points, safety, cost-effectiveness, and formal issues havebeen explicitly pointed by the Consultative Council to justify 209(60.8%), 107 (31.1%), 124 (36.0%), 96 (27.9%), and 61 (17.7%)recommendations, respectively. However, 29 (8.4%), 62 (18.0%),31 (9.0%), 44 (12.8%), and 36 (10.5%) recommendations were madeagainst the results reported on the clinical efficacy, impact ofhard end points, safety, cost-effectiveness, and formal issues,respectively. This data are presented in Table 2.

In addition, these numbers are presented separately forpositive and negative recommendations in Figures. 1 and 2. Thefollowing interpretations can be made. First, Consultative Counciltries explicitly to give rationale especially for negative recom-mendations—the percentage of recommendations accompaniedwith a comment on criterion that backs this recommendation upis higher for negative recommendations (stacked columns higherin Fig. 2), except for clinical efficacy. This type of reasoning ismore often presented for positive recommendations. Second, theConsultative Council is more reluctant to present argumentsopposing the final opinion in case of negative recommendations(stacked columns consisting mostly of the “agree” part). In otherwords, in the case of positive recommendations, the Counciltones down the message more often. Again, the clinical efficacy is

Table 2 – Summary of the impact of different criteria on positive and negative AHTAPol recommendations.

Criteria Positive recommendations (n ¼ 218) Negative recommendations (n ¼ 126)

n (%)* n (%)† n (%)‡ n (%)§

ClinicalEfficacy 141 (64.7) 9 (4.1) 68 (54) 20 (15.9)Impact on hard end points 59 (27.1) 43 (19.7) 48 (38.1) 19 (15.1)Safety 67 (30.7) 21 (9.6) 57 (45.2) 10 (7.9)

EconomicCost-effectiveness 40 (18.3) 41 (18.8) 56 (44.4) 3 (2.4)Budget impact 18 (8.3) 8 (3.7) 17 (13.5) 4 (3.2)

OtherFormal issues 0 35 (10.2) 61 (17.7) 0

AHTAPol, Agency for Health Technology Assessment in Poland.� Number (%) of positive recommendations that referred directly to the favorable profile of technologies in the analyzed criterion.† Number (%) of positive recommendations issued despite the explicitly referred lack of benefit (advantage) of the technology for the analyzedcriteria.

‡ Number (%) of negative recommendations referred to unfavorable profile of the technology for the analyzed criteria.§ Number (%) of negative recommendations issued despite the explicitly referred benefit (advantage) of the technology for the analyzedcriteria.


an exception— here, it is more uniformly in concordance with thefinal recommendation for positive ones than for negative ones.

Third, the criteria that are most ambiguous for positiverecommendations (dashed and solid parts almost equal instacked columns) are the impact on hard end points and cost-effectiveness—among all the positive recommendations thatmention these criteria in only 57.8% and 49.4%, respectively,these criteria support the final recommendation. For example, forclinical efficacy in 94% of the cases it is presented for a positiverecommendation, it is also favorable for the technology. What isquite intuitive is that formal issues are almost exclusivelybrought forward in a fashion unfavorable for a technology, andmuch more frequently in negative recommendations. Except forthat, it is cost-effectiveness that is usually casting an unfavorablelight on the technology.

We also calculated how much the relative frequencies offavorable or unfavorable comments differ for positive or negativerecommendations. We calculated the odds ratio (OR) for a criterionbeing favorable between positive and negative recommendations(by definition identical to the OR of a recommendation being

positive between favorable and unfavorable criterion). Clinicalefficacy changes most between two types of recommendations(OR ¼ 53.3) followed by safety and cost-effectiveness (OR ¼ 18.2both), and then impact on hard end points (OR ¼ 3.5). Thecomposition for the formal issues criterion barely differs.

More in-depths analysis of economic findings, either cost-effectiveness (mean cost per QALY or LYG) or budget impact(mean annual payer’s spending) values, showed no difference inmean values reported for positive and negative recommenda-tions (P ¼ 0.9045 for ICER and ¼ 0.3868 for budget impact). Noclear relation between cost-effectiveness (Fig. 3) and budgetimpact (Fig. 4) values and positive and negative recommenda-tions rates was observed. This is the first signal, before thelogistic regression, that threshold values either for cost-effectiveness or for budget impact distinguishing positive andnegative recommendations cannot be specified.

Finally, a logistic regression model showed joint associationsbetween analyzed criteria and final recommendation (being pos-itive; Table 3). Clinical efficacy seems to be the most importantvariable—impacting both when explicitly stated to be favorable or

Fig. 1 – Number of references to single criteria for positive recommendations. BI, budget impact; CE, cost-effectiveness.

Fig. 2 – Number of references to single criteria for negative recommendations. BI, budget impact; CE, cost-effectiveness.


unfavorable. The remaining variables have predictive power onlywhen presented in an unfavorable fashion.

Discussion

In deciding whether or not to recommend the reimbursement oftechnology from the public budget, the AHTAPol ConsultativeCouncil seems to be considering many aspects of the problem.Although the AHTAPol has adopted a standardized approach andcontinuously publishes key documents online, in case of individ-ual decisions, various aspects of the assessment were weighteddifferently. It is still therefore a challenge to define a clear patternof decision making by the AHTAPol.

The most important criterion found to be used as a rationaleto back up the decisions by the AHTAPol Consultative Councilwas clinical efficacy. It is called for most often in the recom-mendations texts (69.4%). It is illustrated by the number ofpositive recommendations guided by the proved efficacy and

negative ones resulting from the lack of it, which altogetheraccounted for 61% of all recommendations. Safety and cost-effectiveness contributed less often to justification of the finaldecisions. To be granted with positive recommendation, technol-ogy should prove unequivocal efficacy. Almost twice less often,the impact on hard end points was supportive for the recom-mendations. Hard end points are hierarchical, which wasreflected in the content of the recommendations. It is likely thatthe Consultative Council of the ATHAPol appreciated informationon mortality more than on the quality-of-life benefit, but it needsfurther studies.

Another important criterion for reasoning the recommenda-tions was safety for either negative or positive ones. The eco-nomic profile of the technology, that is, cost-effectiveness andbudget impact, was less often used to justify the final decisions.The cost-effectiveness seems not to be a sine qua non criterionwhen issuing positive recommendations, because in 41 cases,even though the council referred to the reports of an unfavorablecost-effectiveness profile, the recommendation was positive.Such a nonstrict use of cost-effectiveness data was also reportedfor NICE health technology coverage decisions [8]. NICE seems toperceive cost-effectiveness as secondary to clinical efficacy, andprofitability is considered only if the technology has passed aclinical effectiveness hurdle.

We failed to detect any empirical threshold value for cost-effectiveness and budget impact analyses that would separatepositive and negative recommendations. Similar approach forNICE decisions was based on modeling and produces inconsis-tent results [9,10]. Studies indicated that NICE makes use of someform of cost-effectiveness threshold but expressed concern aboutits basis and its use in decision making [4]. In addition, sensitivityanalysis around cost-effectiveness represents a challenge inmaking it accessible to those making decisions [11].

In our study, we follow most closely the approach by Devlinand Parker, trying to detect factors associated with positiverecommendation, and among others, to find the impact of ICERon AHTAPol’s decisions. Both the explanatory variables andinterpretation, however, differ from above-mentioned studies.And so we dispose over not the original reports but the recom-mendations issued by the AHTAPol. These recommendationsusually contain information on the interpretation by theAHTAPol whether a technology was shown in the original HTAreport to offer benefits in each of the analyzed criteria. Moreimportantly, these descriptors are available only when AHTAPoldecides to include it in the recommendation text. Therefore, thecharacteristics of the HTA report are often missing, and mostprobably not missing in random. Therefore, we would hardlyinterpret our analysis as finding predictors of AHTAPol’s deci-sions, but rather finding regularities in the way its decisions arerationalized.

Fig. 3 – Distribution of ICER values for positive and negativerecommendations. ICER, incremental cost-effectivenessratio.

Fig. 4 – Distribution of budget impact values for positive andnegative recommendations.

Table 3 – Logistic regression modeling results:Variables associated with positiverecommendation.

Variable; only statisticallysignificant predictors are listed(all P values o0.005)

Odds ratio for positiverecommendation

Clinical efficacy established 12.29No evidence for impact on hard end

points0.251579

Lack of clinical efficacy 0.213248Safety ambiguous or unfavorable 0.170623No cost-effectiveness 0.187739Formal shortcomings 0.21062


Further efforts on promoting systematic consideration of alldecision criteria and the underlying evidence are necessary. Adecision support framework can allow a consistent approach toappraising health care interventions and evidence-basedresource allocation as shown by incorporating multiple-criteriadecision analysis in HTA to support transparent and systematicappraisal of health care interventions [12–15].

Limitations

Our analysis was based only on recommendation texts (predom-inantly on recommendation justification section); thus, indirectconclusions on recommendation determinants could not be fullyobjective. It should be emphasized that the recommendationsavailable on the Web site of the AHTAPol do not have a unifiedstructure that would allow a thoroughly consistent analysis.Although the Consultative Council certainly evaluated all theaspects of submissions, in the recommendations’ texts, presum-ably selected and the most important elements determining thedecision were reported. For this reason, our evaluation of recom-mendations is subject to some subjectivity, which we tried tolimit as far as possible by developing a common interpretationdriven by both language and technical interpretation of theavailable documents. Our analysis, because of formal grounds,defined by available information (the text of individual recommen-dations issued by the Consultative Council or the opinion of thepresident), could not focus on all aspects that influenced issuing aparticular opinion, but only on those that appeared explicitly in therecommendations’ text. We analyzed not only justification sec-tions, but all the recommendations’ document content; therefore,some arguments not raised in the justification but emphasized asimportant could be also identified and analyzed.

In assessing the technology efficacy it was taken into accountwhether the Consultative Council found it satisfactory, regardlessof whether the effectiveness of the technology was comparedwith placebo or an active comparator. Comparable efficacy profileof the assessed intervention over its comparators, depending onindividual circumstances, might have been assessed by theCouncil separately, that is, negatively, as an intervention thatdid not add anything and represented only a burden on thebudget of the payer, or positively, as an intervention of similarefficacy to others used for the same therapeutic indication.

Difficulties in determining the significance of a technology’scost-effectiveness resulted from alternative ways of expressingthe ICER, other than the cost of the QALY or the LYG, which madeinterpretation practically impossible.

Conclusions

The decision-making process of the AHTAPol Consultative Coun-cil is a multifaceted and a multicriterial one, and its resultscannot be easily decomposed into simple associations nor can be

easily predicted. Still, efficacy and safety profile seem to contrib-ute most to final recommendations.

Source of financial support: These findings are the result ofwork supported by the Medical University of Warsaw, WarsawSchool of Economics, and HealthQuest Consulting Company.The views expressed in this article are those of the authors,and no official endorsement by the Medical University ofWarsaw or Warsaw School of Economics is intended or shouldbe inferred.

R E F E R E N C E S

[1] Guidelines for conducting Health Technology Assessment (HTA),version 2.1, April 2009. Available from: http://www.aotm.gov.pl/assets/files/wytyczne_hta/2009/09.06.29_wytyczne_HTA_eng_MS.pdf.[Accessed December 15, 2012].

[2] Agency for Health Technology Assessment in Poland (AHTAPol). http://www.archiwum.aotm.gov.pl/pliki/edu/AOTM%20zarz%2020.pdf.[Accessed December 7, 2012].

[3] Załącznik nr 2 do Zarządzenia nr 20 Dyrektora Agencji OcenyTechnologii Medycznych z dnia 27 marca 2007 r.

[4] Devlin N, Parkin D. Does NICE have a cost-effectiveness threshold andwhat other factors influence its decisions? A binary choice analysis.Health Econ 2004;13:437–52.

[5] Towse A. What is NICE’s threshold? An external view. In: Devlin N,Towse A,eds., Cost Effectiveness Thresholds: Economic and EthicalIssues. London: King’s Fund/Office for Health Economics, 2002:chapter 2.

[6] Tappenden P, Brazier J, Ratcliffe J, et al. A stated preference binarychoice experiment to explore NICE decision making.Pharmacoeconomics 2007;25:685–93.

[7] Dakin HA, Devlin NJ, Odeyemi IA. “Yes”, “no” or “yes, but”? Multinomialmodelling of NICE decision-making. Health Policy 2006;77:352–67.

[8] Agency for Health Technology Assessment in Poland (AHTAPol). http://www.aotm.gov.pl/assets/files/rada/uchwala_rk_aotm_56_16_2008_sunitynib_Sutent.pdf. [Accessed December 7, 2012].

[9] Williams I, Bryan S, McIver S. How should cost-effectiveness analysisbe used in health technology coverage decisions? Evidence from theNational Institute for Health and Clinical Excellence approach. J HealthServ Res Policy 2007;12:73–9.

[10] Schlander M. The use of cost-effectiveness by the National Institute forHealth and Clinical Excellence (NICE): no(t yet an) exemplar of adeliberative process. J Med Ethics 2008;34:534–9.

[11] Williams I, McIver S, Moore D, et al. The use of economic evaluations inNHS decision-making: a review and empirical investigation. HealthTechnol Assess 2008;12:1–175.

[12] Andronis L, Barton P, Bryan S. Sensitivity analysis in economicevaluation: an audit of NICE current practice and a review of its use andvalue in decision-making. Health Technol Assess 2009;13:1–61.

[13] Tony M, Wagner M, Khoury H, et al. Bridging health technologyassessment (HTA) with multicriteria decision analyses (MCDA): fieldtesting of the EVIDEM framework for coverage decisions by a publicpayer in Canada. BMC Health Serv Res 2011;11:329.

[14] Miot J, Wagner M, Khoury H, et al. Field testing of a multicriteriadecision analysis (MCDA) framework for coverage of a screening testfor cervical cancer in South Africa. Cost Eff Resour Alloc 2012;10:2.

[15] Goetghebeur MM, Wagner M, Khoury H, et al. Bridging healthtechnology assessment (HTA) and efficient health care decision makingwith multicriteria decision analysis (MCDA): applying the EVIDEMframework to medicines appraisal. Med Decis Making 2012;32:376–88.



http://www.aotm.gov.pl/assets/files/wytyczne_hta/2009/09.06.29_wytyczne_HTA_eng_MS.pdf

http://www.aotm.gov.pl/assets/files/wytyczne_hta/2009/09.06.29_wytyczne_HTA_eng_MS.pdf

http://www.archiwum.aotm.gov.pl/pliki/edu/AOTM%20zarz%2020.pdf

http://www.archiwum.aotm.gov.pl/pliki/edu/AOTM%20zarz%2020.pdf















http://www.aotm.gov.pl/assets/files/rada/uchwala_rk_aotm_56_16_2008_sunitynib_Sutent.pdf





























A Framework for Applying Health Technology Assessment in Cyprus:Thoughts, Success Stories, and RecommendationsPanagiotis Petrou, MBA, PhD, PhDc�, Michalis A. Talias, PhD

Healthcare Management Programme, Open University of Cyprus, Nicosia, Cyprus

A B S T R A C T

Objectives: Health care decision making, assessment, and procure-ment of medicines is a complex, human resource–demanding, andtime-consuming process. A thorough evaluation of all factors involvedis necessary to optimize the process. The objective of this study wasto describe and analyze the current stage of health technologyassessment (HTA) in Cyprus. Methods: Literature research and pri-vate communication with all involved parties and competent author-ity. Moreover, data, decisions, and recommendations of the Drug’sCommittee were used. Results: Cyprus is a latecomer in this field.HTA has entered a growing phase after the 2007 reform. It has notreached its full potential, and the current state is applicable only tothe public sector, because of the nonexistence of a national healthsystem. Therefore, this poses both a great challenge and a great

barrier considering maximization of the value of money spent andhealth access equity. Conclusions: There is definitely enough spaceand clear necessity for further dissemination, and early successesindicate that steps should be taken toward the introduction of an HTAprocedure that will cover both private and public sectors. Theintroduction of a national health system will further enhance theuptake of HTA, optimize the process, and use the common knowledgestrategy for evidence-based decision making.

Keywords: Drug’s Committee, evaluation, HTA, pharmaceuticals,private sector, public sector.


Health Care Sector in Cyprus

Currently in Cyprus, two fragmented systems run in a parallel,overlapping, and competitive manner with clear disparitiesamong them: public sector and private sector. This situation iscaused by the absence of a national health system. The Ministryof Health (MOH) is the provider, regulator, and payer of publicsector beneficiaries. Public health care is highly centralized, andthe policymaking process takes place at the macro (ministerial)level. There are five major categories of beneficiaries [1] accordingto income and employment status. It is essential to underlinethat 85% of the total population is entitled to free public medicalcare, without any direct or indirect contribution. As a result,moral hazard [2] has been prominent and was expressed byoveruse and misuse of medicines in the pharmaceutical sector. Incontrast, private sector’s patients pay the full amount out ofpocket, unless they are covered by an optional private insurance.

Health care costs in Cyprus account for 6% of gross domesticproduct [3], which pushes Cyprus to the European low segment.The rate of increase in costs in the health care sector outpacesalmost all other European Union (EU) countries [4] primarilybecause of the following reasons [5]:

1. An aging population that has an increasing life expectancy,with concurrent increased morbidity.

2. Lack of prescribing control due to the nonexistence of aninterface management system. The system was launched in2010, but it is still not fully operational.

3. No direct contribution of beneficiaries—Exploitation of moralhazard.

4. Policy susceptible to colloquial evidence especially regardingnew expensive products.

5. Pharmaceuticals in the private sector are regulated only at theprice level.

6. There is a duplication of high-cost hospital services in Cyprus,which have high running cost but are not fully utilized.

7. The above remark is augmented by the low value of the publicsector perceived by beneficiaries. This was an undisputablefinding of a recent study [6] that examined the value formoney regarding beneficiaries of the public sector. Under thehypothesis that all health care systems want to gain morehealth for the same amount of money, the perceived value ofthe health system was assessed. The most important findingis presented in Fig. 1.

8. Preventive programs are underfunded. Preventive programsapply usually to beneficiaries, while the financial burden ofmany diseases is entirely shifted to the MOH.

9. There are no quality indicators. As a result, the MOH cannotassess any health policy, and consequently arbitrary decisionsare taken regarding abortion or carryover of them.






* Address correspondence to: Panagiotis Petrou, Healthcare Management Programme, Open University of Cyprus, P.O. BOX 12794, 2252.Latsia, Nicosia Cyprus.


dx.doi.org/10.1016/j.vhri.2013.06.016

dx.doi.org/10.1016/j.vhri.2013.06.016

dx.doi.org/10.1016/j.vhri.2013.06.016


dx.doi.org/10.1016/j.vhri.2013.06.016

The pharmaceutical market has some unique and inherenttraits that make it quite hard to interpret and definitely tell itapart from other products with regard to their market analysis.The pharmaceutical market posseses an unrivalled demand andsupply feature. There is a three-tier demand structure in whichthe recipient (patients) of the products consumes but has little, ifany at all, involvement in the decision-making process. More-over, the prescriber of the product is perceived as the customer,but does not consume the product. Another feature is that thecost does not represent the production cost. This is quiteillustrative in generic products that have, in certain cases suchas in Italy, half of the original product’s price, and they are stillprofitable. The price of the product is set to offset the researchand development expenses and in certain cases, such as inFrance, is set at a premium to reward innovation.

Governments worldwide and health agencies have appliedspecific and strict legislation to the pharmaceutical market toensure that

1. Life-saving products are available; health systems should notbe exploited by industry.

2. Good manufacturing processes are safeguarded alongthe way.

3. The unique demand and supply does not hinder the controlrole of health agencies regarding product availability [7].

HTA in Cyprus

Many authors have described HTA in a detailed manner [8]. InCyprus, HTA appeared as a term of reference of the Drug’sCommittee in early 2000 as a tool to address uncontrolledincrease in expenditure through rationalization of the decision-making process [9]. Terms of reference were updated andenriched in 2007, allowing further flexibility and introduction ofmore complex and legally demanding schemes. HTA is per-formed through the Drug’s Committee, which falls under theMOH (Pharmaceutical Services). We must highlight the partic-ipation of Health Insurance Organization in the Stakeholder

Forum and the participation of Pharmaceuticals Services at theJoint Action 2 of European Network of Health TechnologyAssessment .

The successful use of tendering, however, led to significantlylow prices for the public sector, which distorted the need for asustaining and rational decision-making process (Fig. 2).

Goals of HTA in Cyprus

According to the terms of reference, HTA should reach thefollowing goals [9,10]:

1. Constantly upgrade, change, and improve clinical guidelines.Currently, guidelines exist in the majority oftherapeutic areas.

2. Define performance indicators and assess effectiveness ofmedicines.

3. Limit the use of newly launched technologies to therapeuticareas for which there is sufficient documentation of efficacyand safety.

4. Reevaluate high expenditure monopoly medicines that con-tribute disproportionately to the overall cost.

5. Categorize evidence deficit in areas in which certain technol-ogies are destined and ways to fill this.

6. Disinvestment.

Criteria for Inclusion of a Medicine in the Formulary

The Drug’s Committee decides on the reimbursement (or not) of aproduct. It assesses drug request on the basis of five main pillars:

1. Prevalence and epidemiology of the disease (prioritization ofresource allocation).

2. Comparative effectiveness according to common practice.3. Economic evaluation, primarily budget impact analysis and to

a lesser degree substantial cost-effectiveness studies (noinclusion of indirect data).

Fig. 1 – Estimated savings from free public health care by age per household member (as percentage of household income). Forbeneficiaries to free public health care of the age group 30 to 50 years, “no perceptible benefit is realized from access to free ofcharge public medical care.” This partly explains the fact that although 85% of the total population is a beneficiary of freemedical care, Cyprus has one of the highest out-of-pocket contributions in the European Union, along with the higher pricesof the private sector.


4. Appraisal of medicine by other HTA agencies such as theNational Institute of Clinical Excellence.

5. Existing competitive medicines in the formulary.

The breadth and quality of data are assessed. As in other smallcountries, the goal is to foster best practices instead of developingones. Because assessment is a context-free and context-sensitiveissue [11], transferability of data may be flawed because of

1. Demographic heterogeneity.2. Costs. Difference in pricing, reimbursement rates, and

between the negotiating power of health prices will lead tocost divergence between countries.

3. Health care practices/different efficiency factor betweenhealth systems.

4. Cultural differences and social values between different pop-ulations [12].

The Committee assesses medicines on the basis of severalcriteria (Table 1) [13]:

Health outcomes measures include cost per quality-adjustedlife-years, life-years gained, progression-free and overall survival,and disease-specific measures such as Psoriasis Approach SeverityIndex and American College of Rheumatology Index. The numberneeded to treat approach was implemented in the assessment ofsmoking cessation products. This was also implemented forcompetitive medicines that have a significant price difference (e.g., different Anatomic Therapeutic Chemical 3 categories) in orderto enhance competition in the tendering process.

In 2007, the ministerial decision [13] enabled the formation of atherapeutic algorithm based on the outcome of the tender. Ther-apeutically equal products competed and instead of eliminatingthe losers, these were designated as second- and third-line therapy,respectively. This occurred in certain therapeutic categories for

which there is enough documentation that tolerance to relevantmedicines is limited and therefore, there is strong possibility thatpatients may need to switch treatment. The case of the anti–tumornecrosis factor agents was a landmark because the contribution ofall stakeholders (MOH, physicians, and patients) to the HTA processhas led to a mutually beneficial outcome, and as a result every yearthe list for anti–tumor necrosis factor agents has still available slots.

Major therapeutic categories that got into this scheme includearomatase inhibitors, adjuvant immunosuppressive treatment,such as mycophenolic acid, antidepressants, antiepileptic agents,and erythropoetins. Before the assessment, companies areallowed to provide further supporting materials regarding theefficacy and estimated cost of their products, which adds to thetransparency of the process.

Another concern of the Drug’s Committee is the possible off-label use of expensive products, due to the overuse of medicinesin Cyprus as addressed earlier. This may lead to a reduction inhealth benefits associated with the use of a product, due to theuncertainty associated with its off-label use. Therefore, the risk ofoff-label use is counterbalanced by the requirement of preap-proval. This was the primary reason for the rejection of ranibi-zumab, despite recommendations by the National Institute forHealth and Clinical Excellence [14].

HTA implementation has not always been very successful inbroadening the scope and in certain cases overlooked oneintrinsic factor, interrelation to health policy [15]. The MOHimplemented a public campaign to create awareness among thepublic for the prevention of cervix cancer; however, the Drug’sCommittee did not approve the only available vaccine. Moreover,as the complexity factor of the therapeutic regimen increases,such as in immunosuppressive ones, assessment can be compli-cated and lengthy. A prominent example was assessment of themammalian target of rapamycin inhibitors whose dosage ishighly personalized and budget impact of each product is

Fig. 2 – SWOT analysis of health technology assessment in Cyprus. Source: Authors approach. HTA, health technologyassessment; MOH, Ministry of Health; NHS, National Health System; SWOT, strength, weakness, opportunities, and threat;WTP, willingness to pay.


determined by other factors, such as adjuvant immunosuppres-sive treatment.

Prescription Guidelines and Preapprovals

The Drug’s Committee has successfully implemented controlledprescription of certain medicines. The majority of guidelines specifytreatment line, exceptions, patient profile, further requirements suchas serum marker levels, expected duration of treatment, and indica-tion of failure. Statins were one of the most successful examples. Theintroduction of prescription guidelines for statins (including preap-proval for high-potency statins) concomitantly with the introductionof generic simvastatin was successful in avoiding the “reallocation ofdemand,” as observed in Belgium (2007–2011: Daily defined doseconsumption increased by 52%, cost decreased by 48%).

Similar guidelines were elaborated for the prescription of alloncology medicines, insulin glargine, rosiglitazone, cinacalcet,and darbeopetin alpha. For the majority of these products, apreapproval is also necessary, usually with the obligation for thesubmission of relevant laboratory documentation. The details ofthe patients are filed.

Indication-based guidelines were elaborated for angiotensin IIreceptor blockers. Different protocols were compiled for hyper-tension, congestive heart failure, and diabetic nephropathy.

In oncology medicines with significant uncertainty and highcost, due to the lack of effectiveness data, the Drug’s Committeehas made exceptions for the compassionate use of cancer drugsin a small target group population in which benefits may not besufficiently captured. Criteria are as follows:

1. Patient’s life expectancy is less than 24 months.2. There is sufficient data that the treatment will extend life for

at least an additional 3 months compared with currenttreatment.

3. There is no alternative treatment with equal effectivenessavailable.

4. The target group is a small patient population [16].

Managed Entry Agreements in Cyprus

Risk sharing [17] has not been implemented and price volumeagreement has been applied only scarcely, mainly due becauseof human resources required for monitoring. Pemetrexed gained

another indication of malignant pleural mesothelioma,in addition to the one existing for non–small cell lung cancer.Because of comparative effectiveness among all availabletreatments for non–small cell lung cancer, an approach wasset up that consisted of three scenarios. The prices incorpo-rated expected efficacy of the product and net benefit for thecompany.

The addition of a new indication of deferasirox and theincrease in the daily dose to 40 mg/kg led to the managed entryagreement of dose capping agreement between Novartis andthe MOH. On the basis of this agreement, the MOH reimbursesdaily doses up to 30 mg/kg (average 2160 mg daily per patient)while additional dosage burdens the company. The company isobliged to provide free goods to the MOH, based on dose agree-ments. Currently, 38 patients are registered in this scheme, whichwill last for 3 years, and data will be revised every 6 months tocheck for deviations.

Further Potential for HTA in the New Financial Era

Cyprus has recently signed a memorandum of understanding withTroica (term used to define the committee consisting of the Interna-tional Monetary Fund, the European Union, and the European CentralBank) to secure a life-sustaining bailout of 10 billion euros. As Troica’sprimary target is to enhance the efficiency of governance, oneimportant prerequisite for the health sector is the implementationof HTA for the 10 costlier pharmaceutical and medical equipment.This provision will further leverage shift toward an integrated HTAsystem, and several approaches are currently being considered, suchas the introduction of two HTA formats, according to the estimatedbudget impact (light and full version), an approach currently imple-mented in many countries such as The Netherlands [18].

Challenges for HTA in Cyprus

The current product mix of these two fragmented markets doesnot allow dissemination of policies in the private sector, neg-atively affecting health equity. Moreover, the Drug’s Committeeshould be authorized to act proactively regarding the assessmentof new medicines.

An important limitation is the lack of an official willingness-to-pay threshold in Cyprus [19] regarding economic evaluation.

Finally, the conflicting role of the MOH, which in the caseof HTA, it assesses, appraises, and procures as well, is a

Table 1 – Criteria for assessment.

Criteria Importance Comments

Disease prevalence Major Easy to assessGuidelines of the National Institute

for Health and Clinical ExcellenceMajor Transferability of data has to be checked for major

divergencesEfficacy data Major Clinical effectiveness must be assessedBudget impact analysis Dominant May conflict with cost-effectiveness approach

Off-label use Medium (unless specific trenddocumented)

Difficult to assess, may compromise actual medical need.Interface management will address this issue

Cost-effectiveness Major (difficult to apply for eachmedicine a country-specificstudy)

A basic economic analysis is performed

Impact on spending for othermedical interventions

Medium Incorporation of nonpharmaceutical interventions. Interfacemanagement may enable control. Difficult to assess

Source: Terms of Reference, Drug’s Committee.


major negative aspect of the system, compromising objectiveappraisal.

Discussion

HTA should be incorporated in the context of a general healthpolicy because health policy defines the optimum use of amedicine to deliver superior results. Health policy should coverboth sectors because ultimately all chronic patients will bebeneficiary of the public sector and prevention of discriminationmust be pursued.

Incorporation of other stakeholders (university, patient asso-ciations) may enrich the decision-making process and lead to acontext for more informative outcomes, diverting the conflictingrole of the MOH.

A detailed cost-effectiveness analysis must be implemented,at least for the cost driver categories, which also needs to beregulated with regard to technical parameters, such as timehorizon, modeling methods, discounting, and type of economicevaluation (budget impact, cost-effectiveness analysis). This wasalso recommended by the Internal Audit department of Cyprus[20].

Uncertainty needs to be addressed in the context of HTA in allits expressions because it affects the impact of the decision andsensitivity analysis must cover a range of assumptions.

Because of the variety of statistical approaches such asmetanalysis or mixed treatment comparison, an accepted formatshould be elaborated taking into consideration particularities ofcertain therapeutic categories. Moreover, the rating of evidence iscrucial and we suggest that GRADE classification should beadapted because it provides an excellent grading of evidence [21].

The scope of the assessment should be broad and must beable to compare interdisciplinary interventions (i.e., medical vs.pharmaceutical interventions) [22,23].

The existing coverage system does not make provision fordifferent levels of reimbursement, because all diseases areconsidered to have the same utility. Consequently, the reim-bursement levels are the same for all diseases. A classification ofstrength of evidence will enable better forecasting, resourceallocation, and demand control.

We observed that reduction in the pharmaceutical growthrate seems to have coincided with the introduction of HTAmethods; however, this has to be verified. The normalization ofthe annual pharmaceutical growth rate to less than 3% is a goodfirst sign and at least partially is attributed to HTA implementa-tion, which is gaining ground in Cyprus [24], as seen in Table 2.

We must however interpret this reduction in the increase ratewith caution because Cyprus’s economy entered a prolongedrecession period in 2008 [25]; therefore, this reduction can beattributed to a tighter control of expenditures, rather than to theability to control expenses as such.

The Drug’s committee evaluates each product only at request.We believe that this leads to lack of symmetry of the system.More importantly, this does not allow comparison among treat-ments and does not elaborate an overall approach to a disease,which could be achieved by assessing other intervention meth-ods. Several therapeutic interventions run in parallel and com-parative effectiveness has not been documented. Periodicalassessment of guidelines and disease management should beestablished and also assessment of medicines by the Drug’sCommittee would minimize reactions from the industry andsubsequent exerted pressures.

Finally, newer approaches must be incorporated taking intoaccount clinical uncertainty. The introduction of risk-sharing meth-ods, price volume agreements, and managed entry agreement willfurther optimize the HTA context in Cyprus. HTA is emerging inCyprus’s health care sector. There is an imperative need, and thecurrent financial era does support further dissemination. The smallsize of the country andmarket fragmentation hinders its full uptake.

Sources of financial support: The authors have no otherfinancial relationships to disclose.

R E F E R E N C E S

[1] Golna C, Pashardes P, Allin S, et al. Health Care Systems in Transition:Cyprus. Copenhagen: WHO Regional Office for Europe on behalf of theEuropean Observatory on Health Systems and Policies, 2004.

[2] Sachs JD. Macroeconomics and Health: Investing in Health forEconomic Development. World Health Organization WHO LibraryCatalogue-in-Publication Data World Health Organization. Geneva,Switzerland: World Health Organization, 2001.

[3] Eurostat [online]. Health care statistics, 2009. Available from: http://epp.eurostat.ec.europa.eu/statistics_explained/index.php/Healthcare_statistics. [Accessed September 17, 2011].

[4] Sammoutis G, Paschalides C. Will the sun shine over Cyprus’s nationalhealth system? Lancet 2011;377:29.

[5] Antoniadou M. Can Cyprus overcome its health care challenges? Lancet2005;365:1017–20.

[6] Andreou M, Pashardes P, Pashourtidou N. Cost and value of health carein Cyprus. Policy paper. Nicosia: University of Cyprus, 2010. Availablefrom: http://www.ucy.ac.cy/data/ecorece/DOP02-10. [Accessed August18, 2011].

[7] Kanavos P, Vandegrift M. Health policy versus industrial policy in thepharmaceutical sector: the case of Canada. Health Policy1997;41:241–60.

[8] Battista RN, Hodge NJ. The development of health care technologyassessment: an international perspective. Int J Technnol Assess HealthCare 1995;11:287–300.

[9] Petrou P. PHIS Hospital Pharma Report; 2009. Available from: http://phis.goeg.at/downloads/hospitalPharma/PHIS%20Hospital%20Pharma%20Report%202009%20Cyprus.pdf. [Accessed September 6, 2011].

[10] Ministry of Health. Pharmaceutical reform in Cyprus. 2006. Availablefrom: http://www.cyprus.gov.cy/moi/pio/pio.nsf/All/BEFF3C4943A9C696C2257127003D66A8?OpenDocument&print. [Accessed September 11,2011].

[11] Velasco Garrido M, Gerhardus A, Røttingen JA, Busse R. Developinghealth technology assessment to address health care system needs.Health Policy 2010;94:196–202.

[12] Goeree R, He J, O’Reilly D, et al. Transferability of health technologyassessments and economic evaluations: a systematic review ofapproaches for assessment and application. Clinicoecon Outcomes Res2011;3:89–104.

[13] Drug’s Committee, Ministry of Health. Terms of reference drug’scommittee. Available from: http://www.moh.gov.cy/moh/phs/phs.nsf/dmlphcomm_gr/dmlphcomm_gr?OpenDocument. [Accessed June 10,2013].

[14] Appraisal of macular degeneration (age-related)—ranibizumab andpegaptanib (TA155). 2008. Available from: http://www.nice.org.uk/TA155. [Accessed December 17, 2011].

[15] Drummond M, Kanavos P, Sorenson C. Ensuring Value for Money inHealth Care: The Role of Health Technology Assessment in theEuropean Union. Denmark: World Health Organization on behalf of theEuropean Observatory on Health Systems and Policies, 2008.

[16] National Institute for Health and Clinical Excellence. Appraising end oflife medicines. 2008. Available from: http://www.nice.org.uk/media/26E/43/Endoflifemedicines.pdf. [Accessed December 17, 2011].

Table 2 – Output of Drug’s Committee.

Year Number ofHTAsperformed

Total number of appraisals(including resubmissions)

2010 11 1162009 24 1162008 8 165

HTAs, health technology assessments.Source: Annual Report of Ministry Of Health, 2010.









http://epp.eurostat.ec.europa.eu/statistics_explained/index.php/Healthcare_statistics









http://www.ucy.ac.cy/data/ecorece/DOP02-10







http://phis.goeg.at/downloads/hospitalPharma/PHIS%20Hospital%20Pharma%20Report%202009%20Cyprus.pdf




http://www.cyprus.gov.cy/moi/pio/pio.nsf/All/BEFF3C4943A9C696C2257127003D66A8?OpenDocument&print











http://www.moh.gov.cy/moh/phs/phs.nsf/dmlphcomm_gr/dmlphcomm_gr?OpenDocument

http://www.moh.gov.cy/moh/phs/phs.nsf/dmlphcomm_gr/dmlphcomm_gr?OpenDocument


http://www.nice.org.uk/TA155

http://www.nice.org.uk/TA155






http://www.nice.org.uk/media/26E/43/Endoflifemedicines.pdf

http://www.nice.org.uk/media/26E/43/Endoflifemedicines.pdf

[17] Espín J, Rovira J, García L. Experiences and Impact of European RiskSharing Schemes Focusing on Oncology Medicines. Grenada, Spain:Andalusian School of Public Health, 2011.

[18] Stolk AE, de Bont A, van Halteren AR, et al. Role of health technologyassessment in shaping the benefits package in The Netherlands. ExpertRev Pharmacoecon Outcomes Res 2009;9:85–94.

[19] Dorte Gyrd-Hansen. Willingness to pay for a QALY. Health Econ2003;12:1049–60.

[20] General audit recommendations for rational medicines procure-ment. 2011. Available from: http://pdf.politis-news.com/pdf/pdf?-A=257761,pdfview.html&-V=pdfarchivefullpaper. [AccessedNovember 17, 2011].

[21] Guyatt GH, Oxman AD, Vist GE, et al. GRADE: an emerging consensuson rating quality of evidence and strength of recommendations. BMJ2008;336:924–6.

[22] Culyer AJ, Lomas J. Deliberative processes and evidence-informed

decision making in healthcare: do they work and how might we know?

Evidence Policy 2006;2:357–71.[23] Garrido VM, Kristensen FB, Nielsen CP, et al. Health Technology

Assessment and Health Policy Making in Europe: Current Status,

Challenges and Potentials. Denmark: World Health Organization, 2008.[24] Cyprus Ministry of Health. Annual health report, Nicosia; 2010.

Available from: http://www.moh.gov.cy/moh/moh.nsf/page09_gr/

page09_gr?OpenDocument. [Accessed June 10, 2013].[25] Statistical Service of the Republic of Cyprus. Latest figures: GDP growth

rate, 1st quarter 2010. Available from: http://www.cystat.gov.cy/mof/

cystat/statistics.nsf/All/ADAE80767589A7B9C22579AE0029E929.

[Accessed June 8, 2012].











http://pdf.politis-news.com/pdf/pdf?-A=257761,pdfview.html&-V=pdfarchivefullpaper













http://www.moh.gov.cy/moh/moh.nsf/page09_gr/page09_gr?OpenDocument

http://www.moh.gov.cy/moh/moh.nsf/page09_gr/page09_gr?OpenDocument


http://www.cystat.gov.cy/mof/cystat/statistics.nsf/All/ADAE80767589A7B9C22579AE0029E929

http://www.cystat.gov.cy/mof/cystat/statistics.nsf/All/ADAE80767589A7B9C22579AE0029E929



Systematic Review of Economic Evaluation Literature in Ghana: Is HealthTechnology Assessment the Future?Emmanuel Ankrah Odame, MBCHB, MPH�

Ghana College of Physicians and Surgeons, Accra, Ghana

A B S T R A C T

Objectives: In many countries, such as Ghana, there is an increasingimpetus to use economic evaluation to allow more explicit and trans-parent health care priority setting. An important question for policymakersin low-income countries, however, is whether it is possible to introduceeconomic evaluation data into health care priority-setting decisions.Methods: This article systematically reviewed the literature on economicevaluation on medical devices and pharmaceuticals in Ghana publishedbetween 1997 and 2012. Its aim was to analyze the quantity, quality, andtargeting of economic evaluation studies that relate tomedical devices andpharmaceuticals and provide a framework for those conducting similarhealth technology assessment reviews in similar contexts. Results: Thereview revealed that the number of publications reporting economic eval-uations was minimal with regard to medical devices and pharmaceuticals.Conclusions: With the introduction of the National Health Insurance

Scheme since 2004 policymakers are confronted with the challenge ofallocating scarce resources rationally. Priority setting therefore has to beguided by a sound knowledge of the costs of providing health services. Theneed for economic evaluation is thus important. More costing studies werefound; there were very few cost-effectiveness analysis studies. If economicevaluation is useful for policymakers only when performed correctly andreported accurately, these findings depict barriers to using economicevaluation to assist decision-making processes in Ghana; hence, there isa need for an independent health technology assessment unit.

Keywords: economic evaluation, Ghana, medical devices, pharma-ceuticals.


Background

Modern health care is cost and technology intensive and expectsvalue for money, creating demand for evidence-based practiceand health technology assessment (HTA). In high-income coun-tries, HTA is often done in specialized HTA institutions. Indeveloping countries, however, HTA is often lacking, despitethe apparent need. Therefore, health care decisions are oftensubjective. Improved understanding of the practice of evidence-based medicine (EBM) in many developing countries, and organ-izations such as the Cochrane Collaboration, now facilitateevidence-informed decisions [5].

HTA is the scientific process of evaluating health technologies(pharmaceuticals, vaccines, surgical procedures, medical equip-ment and devices, etc.) to facilitate informed decisions by stake-holders: health care providers, payers, consumers, regulators,policymakers, and so on [1]. In high-income countries, HTA is aformal discipline undertaken by trained professionals to guidestakeholders, including governments, to make decisions on thebasis of sound scientific principles. Most resource-poor settings lackformal HTAmechanisms; in such settings, health care decisions areoften based on no evidence and are more subjective. A recentsurvey [9] evaluating the use of key HTA principles [2] reported thateven in the few developing countries in which HTA is being used,

although the principles were considered relevant by HTA producersand users, the level of application was uniformly low.

Although resource allocation for health and demand for newhealth technologies have increased in many low-income coun-tries, robust decision-making mechanisms have not developed inparallel. Decisions are often driven by experience, thrust of donoragencies, and lobbying pressure [6]. For example, a report fromPeru noted that decisions on the human papilloma virus vaccineat the local level were mainly driven by local political pressurerather than scientific evidence [11]. In Rwanda, the governmenthad allocated a disproportionately large amount of funds for HIV/AIDS than for malaria and other greater perceived needs, becausedonor grants were specifically allocated for HIV/AIDS [22]. Like-wise in India, sustained single-point focus on poliomyelitiseradication using supplementary immunization (owing to WorldHealth Organization and global pressure) has critically weakenedthe routine immunization program with other childhood vac-cines [20].

Commercial pressure is also a major force skewing thedecision-making process in developing countries; this is especiallyrelevant for newer vaccines, expensive drugs, devices, and equip-ment [24]. For example, current immunization recommendationsof the Indian Academy of Pediatrics were produced by expertconsensus at a meeting sponsored by a multinational company.




Conflicts of interest: The author has indicated that he has no conflicts of interest with regard to the content of this article.


� Address correspondence to: Emmanuel Ankrah Odame, Ghana College of Physicians and Surgeons, 54 Independence Avenue, Ridge,Accra, Ghana.


dx.doi.org/10.1016/j.vhri.2013.07.006

dx.doi.org/10.1016/j.vhri.2013.07.006

dx.doi.org/10.1016/j.vhri.2013.07.006


dx.doi.org/10.1016/j.vhri.2013.07.006

Not surprisingly, the stated objective of the recommendations wasto produce guidance for three products recently marketed by thecompany in India [15].

Although poorer countries should be more careful in spendingmoney, the opposite often happens in most instances. Health caresystems sometimes successfully negotiate lower pricing for phar-maceuticals, but public health programs end up paying more thanthe negotiated prices [16]. Such observations corroborate theargument of Chalkidou et al. [6] that in many developing countries,“health services and technologies purchased with public funds areselected through idiosyncratic processes that often have little todo with systematic analysis of their potential health benefit orvalue for money.” In many developing countries, “expert-based”guidance is used as a surrogate for robust methods, perhaps ingood faith [12]. A group of “experts” prepares a “consensus” state-ment on a given health technology. The procedure for selectingthe experts and the processes used to reach consensus are seldomdescribed [15]. In developing countries, physicians often base their“advice” on nonscientific considerations, particularly the influenceof the pharmaceutical industry [21]. Material provided by pharma-ceutical manufacturers is reported as the most frequently usedresource by many physicians, with prescribing decisions influ-enced by training activities sponsored by pharmaceutical compa-nies and visits by sales representatives [21].

In Ghana, a National Health Insurance Scheme (NHIS) wasestablished in 2004, and the Ghana Diagnostic Related Groupprovider payment mechanism has been fully implemented, andalthough there are a number of challenges, the payment systemis functioning well and generally accepted by providers. It hasnot, however, succeeded in containing costs, particularly foroutpatient services, with outpatient claims now accounting for70% of total NHIS claims and 30% of total costs of the NHIS.Furthermore, between 2007 and 2009, the average outpatient costper claim increased by nearly 50% from US$3.47 to US$5.06.Without a control of the rapid rise in service delivery costs ofthe NHIS, in addition to mobilization of more revenue, thescheme will not be sustainable [29].

Meanwhile, to date there is no institution that does cost-effectiveness analysis of the pharmaceuticals that are part ofthe benefit package. There is no evidence base guiding the drugsthat are part of the benefit package.

Clearly, health care decisions by all stakeholders in Ghana areoften highly subjective. There is an urgent need to bring inobjectivity, reproducibility, and transparency. HTA as a scientificprocess of evaluating health technologies (pharmaceuticals, vac-cines, surgical procedures, medical equipment and devices, etc.)to facilitate informed decisions by stakeholders—health careproviders, payers, consumers, regulators, policymakers, and soon—can address this need. Hence, the need of this systematicreview to critically evaluate the evidence base of cost-effectiveness analysis of medical devices, vaccines, pharmaceut-icals, and surgical procedures.

Methods

Literature searches were carried out in November 2012 by usingthe following keywords: “Ghana” and “economic evaluation” or“cost-minimisation” or “cost-effectiveness” or “cost-utility” or“cost-benefit.” The search was performed by using the followingdatabases: PubMed, EMBASE (Ovid), and Academic Search Elite(EbscoH). It included all published and unpublished literatureavailable between January 1, 1997, and October 31, 2012. Inclusioncriteria were all economic evaluations on medical devices andpharmaceuticals including vaccines.

All identified abstracts were reviewed by the first author.Studies were excluded if they did not present both the costs

and the outcomes of a study, or if they were an editorial ormethodological article. Studies were also rejected if they were notapplied to a Ghanaian context and all other economic evaluationsapart from medical devices, vaccines, and pharmaceuticals. Allremaining articles were reviewed by using their full-text formatsand classified according to: 1) the type of evaluation, 2) the type ofintervention, and (3) the body system affected by the particularhealth problem.

Published articles were grouped by type of evaluation, andwere considered to be: 1) a partial economic evaluation if onlyeither the costs or the outcomes of a single intervention werecompared; 2) a cost-minimization analysis if costs of differentinterventions were compared with evidence of equal ease burdenin terms of disability-adjusted life-years; 3) a cost-effectivenessanalysis if health outcomes were presented in intermediateterms, for example, disease prevented; (4) a cost-utility analysisif health outcomes were expressed in terms of quality-adjustedlife-years or disability-adjusted life-years; and 5) a cost-benefitanalysis if health outcomes were measured in monetary units.Only those studies that did economic evaluation in relation tomedical devices, pharmaceuticals, and vaccines were considered.

The quality of studies was measured in two different ways.First, studies were appraised on their adherence to specificmethodological and reporting practices based on published rec-ommendations [7,8]. These included: 1) clearly indicating thestudy perspective; 2) description of comparator(s); 3) use ofdiscounting methods if the costs and/or outcomes were from astudy period of more than 1 year; 4) reporting the incrementalcost-effectiveness ratio (ICER) rather than the average cost-effectiveness ratio; 5) performing uncertainty analysis on theresults; and 6) disclosing funding sources.

Results

A total of 50 abstracts were identified from the search of bothpublished and unpublished material (Fig. 1).

Of these, 45 abstracts were initially excluded because costsand outcomes were not mentioned simultaneously and becausethey disclosed funding sources. Seven articles were reviewed infull-text format. From the review of seven full-text articles, fourarticles were found not to be relevant because they were noteconomic evaluations of medical devices and pharmaceuticalsincluding vaccines. The culmination of this was three economicevaluations, one looked at vaccines and two on malariamanagement.

In terms of where they were published, international peerreview was the source and an international person was theprincipal author for two of them, with a Ghanaian author as theprincipal author for one of them. All three economic evaluationswere cost- effectiveness analysis. Two of them used an ICER.

International standards recommend that economic evalua-tion studies should extend (through modeling) over a time periodthat is long enough to capture the full costs and consequences ofan intervention. The funding sources were all from internationalnonprofit organizations.

There was no significant relationship between the source offunding and the quality of the report (using chi-square andFisher’s exact tests, where appropriate). See Table 1.

Discussion

HTA is the scientific process of evaluating health technologies(pharmaceuticals, vaccines, surgical procedures, medical equip-ment and devices, etc.) to facilitate informed decisions by stake-holders: health care providers, payers, consumers, regulators,


policymakers, and so on [1]. A medical device is a product usedfor medical purposes in patients, diagnosis, therapy, or surgery.Pharmaceutical products achieve their principal action by phar-macological, metabolic, or immunological means. However, med-ical devices act by other means such as physical, mechanical,thermal, physicochemical, or chemical [28]. A survey of experts,conducted by Daar et al. [29], clearly showed the greatest need ofnew technology for affordable, simple diagnosis of infectiousdiseases in developing countries.

From the systematic review it is evidently clear that veryscanty technology assessment has been done for medical devicesand pharmaceuticals. In Ghana, currently with the NHIS the costof medicines is a major cost driver, but evaluation of medicines aspart of the benefit package is done without a robust evidencebase. In this age in which there are multiple solutions andgenerally a wide spectrum of possibilities and strategies for mosthealth care problems, the need for the evaluation of the technol-ogy and of the relevant alternative technologies available hasnever been greater. Second, as the health sector has limitedbudgets, HTA becomes one of the most important tools used tocontain the increasing cost of health care without compromisingsafety. HTA should form the basis for health technology policies[2]. Prioritizing resource allocation and the need for new medicaltechnologies are also increasing in developing countries [3].Health insurance programs are emerging and expanding in sub-Saharan Africa [6]. Besides, health insurance programs are emerg-ing and expanding more and more in this region [27]. But decis-ions can easily be influenced by experience, thrust of donoragencies, and lobbying pressure for new technologies, for exam-ple, from commercial organizations. This can lead to inappropri-ate use of technologies, which do not address health needs, andinefficient use of resources. Pichon-Riviere et al.’s [9] survey aboutthe usage of HTA methods in a resource-poor setting suggestedthat principles of HTA were seen as relevant, but there was a lackof application. South Africa, as a middle-income coun-try, however, has planned strategies for HTA. Yet, the implemen-tation of such a national HTA framework has been slow. The lackof skills related to HTA is also a critical problem. Although there isa growing base of skills in the running of clinical trials, there arevery few health economists trained in applying HTA method-ologies. In Ghana, there was some training on pharmacoeconom-ics organized by the Ministry of Health, but the follow-upactivities have not been sustained.

The emergence and spread of EBM was expected to addressrelevant needs, through building local capacity for using system-atic reviews and influencing policymakers to make evidence-based decisions. However, EBM has major limitations in that itfocuses on generating evidence of efficacy alone. In addition, it

fails to factor in local needs and contexts for transferability ofevidence generated in different health care settings [5]. Thedevelopment process of HTA is usually expensive and timeconsuming [6], which poses problems in many resource-poorsettings. In Ghana, as in other low- and middle-income Africancountries, policymakers have in recent years come under increas-ing pressure to justify resource allocation decisions in the healthsector [26]. The number of economic evaluation studies in Ghanais quite low, however, especially in the area of medical devices,vaccines, and pharmaceuticals, in contrast to Canada, the UnitedKingdom, or The Netherlands [17] where economic evaluation hasbeen formally accepted for use in policy decision making. Inaddition, this review found that the majority of economic evalua-tion studies performed in Ghana between January 1, 1997, andOctober 31, 2012, were vulnerable to bias because of the quality ofthe evidence used. Poor reporting quality limits the usefulness ofeconomic assessment in the making of policy decisions.

Evidence-based policy according to scientific methods canreduce costs and improve the outcome for patients. Such analysisthus provides an ethical way of evaluating new health technol-ogies. Therefore, the participation of health care authorities inGhana is crucial. There is an urgent need to bring in objectivity,reproducibility, and transparency for local health policy makersin Ghana.

The review indicates that cost-effectiveness analysis was theonly study type for economic evaluations performed in a Gha-naian setting for the study period. This is comparable to findingsin other settings [13,14], probably because the approach isrelatively straightforward. It compares costs with outcomesmeasured in natural units, such as per life saved, per casedetected, and per pain- or symptom-free day.

This is in contrast to cost-utility analyses (which require moreassumptions on health-state preferences [18]) and cost-benefitapproaches (which face difficulties and controversy in applying amonetary value to human life [19]).

Cost-effectiveness analysis can be very useful when differenthealth care interventions are not expected to produce the sameoutcomes. This type of study alone, however, cannot handlequestions of efficiency of resource allocation when such deci-sions have to be made across different health problems [23].

In both technical and resource-allocation terms, there is aneed to encourage and support the undertaking of cost-utility orcost-benefit analyses by academics and health researchersbecause these evaluation types are better able to assist decisionmakers in judgments about the allocation of resources acrosshealth care programs.

In addition, this review highlights that serious attention needsto be given to the quality of the reporting and information used in

50 abstracts were identified

45 abstracts were excluded

2 articles were identified from references of the initiallysearched articles

5 full articles were reviewed

3 economic evaluation publications werefinally reviewed

4 articles excluded after reviewing their fulltexts

Fig. 1 – Flow diagram of systematic review.


economic analyses performed in Ghanaian settings. The advan-tages of using systematic reviews of clinical effects are twofold[4]. First, a more precise estimate can be obtained from combin-ing outcome data from a number of studies. Second, by usingresults from studies carried out in a range of settings (assumingthat these studies are sufficiently homogenous to be compara-ble), the estimate can be applied to a more general patientpopulation with different baseline risks, rather than specificallyto the narrow population selected for individual economic eval-uations as was observed in the studies trial.

It is noteworthy that there were two serious methodologicalpitfalls that were commonly found in the Ghanaian economicevaluations. The first was the lack of calculation of an ICER in oneof the studies. This study reported an average cost-effectivenessratio, that is, total cost divided by total effect for the interventionsbeing compared. The report of an average ratio may lead todangerously flawed conclusions and may limit attempts to makedirect comparisons between interventions because an averageratio implies the comparison of each alternative with an inter-vention that incurs no costs and no effects [10]. Without calcu-lating and presenting ICERs, it is possible for readers to bemisguided by the results and to conclude that the new techniquewas simply fourfold more expensive than the standard test.

Combined with the use of poor-quality evidence for estimat-ing clinical effects, this could seriously undermine confidence inthe findings of these economic evaluations and their ability toinform health care resource-allocation decisions.

Among the studies that applied discounting, the review alsofound that the discount rate used was 3%. There is still nointernational agreement, however, on how to deal with futurecosts and benefits.

Major debate continues about whether it is justified todiscount health benefits, and if so, what discount rate shouldbe used and whether it should be different from that used formonetary costs [3]. Nevertheless, this review of the literature oneconomic evaluations performed in Ghanaian settings has shownlimited evidence of economic evaluation publications with regardto medical devices and pharmaceuticals.

It is important to point out the limitations of this study. Thereis no national database for health care publications in Ghana.This study searched the literature in international databases andonly included literature published in English and also unpub-lished literature in some of the university campuses. The use ofthese databases also meant that abstracts, conference proceed-ings, Ghanaian publications, masters and doctorate theses, andarticles presented at symposia or seminars were not included inthe search results. The number of publications available, how-ever, can be used as a proxy to reflect the research capacity inthis field in Ghana, and this review has shown that the quantity,quality, and targeting of economic evaluation studies is not yetadequate to meet the needs and concerns of decision makers inGhana. Current studies support the establishment of HTA insti-tutions in low- and middle-income countries [25].

Conclusions

This review demonstrates an urgent need for a comprehensiveand systematic method for conducting economic evaluations formedical devices and pharmaceuticals, especially in an era ofincreasing cost in the Ghanaian health care system and alsodwindling donor support to the health sector. HTA of medicaldevices and pharmaceuticals seems the most viable option toensure cost-containment, which is the main agenda of the NHIS,and ultimately save the health care system.

Acknowledgment

I thank Amanda Odame for her secretarial support.Source of financial support: The author has no other financial

relationships to disclose.

R E F E R E N C E S

[1] Anonymous. Health technology assessment. Available from: http://www.inahta.org/HTA/. [Accessed November 28, 2012].

[2] Drummond MF, Schwartz JS, Jonsson B, et al. Key principles for theimproved conduct of health technology assessments for resourceallocation decisions. Int J Technol Assess Health Care 2008;24:244–58.

[3] Cairns J. Discounting in economic evaluation. In: Drummond M,McGuire A, eds. Economic Evaluation in Health Care: Merging Theorywith Practice. Oxford: Oxford University Press, 2001.

[4] Drummond M, Pang F. Transferability of economic evaluation results.In: Drummond M, McGuire A, eds. Economic Evaluation in Health Care.Oxford: Oxford University Press, 2001.

[5] Mathew JL. KNOW ESSENTIALS: a tool for informed decisions in theabsence of formal HTA systems. Int J Technol Assess Health Care2011;27:139–50.

[6] Chalkidou K, Levine R, Dillon A. Helping poorer countries make locallyinformed health decisions. BMJ 2010;341:c3651.

[7] Drummond MF, Jefferson TO. Guidelines for authors and peer reviewersof economic submissions to the BMJ. BMJ 1996;313:275–83.

[8] Drummond M, Sculpher M, Torrance G, et al. Method for the EconomicEvaluation of Health Care Programmes. (3rd ed.). Oxford: OxfordUniversity Press, 2005.

[9] Pichon-Riviere A, Augustovski F, Rubinstein A, et al. Health technologyassessment for resource allocation decisions: are key principles relevantfor Latin America? Int J Technol Assess Health Care 2010;26:421–7.

[10] Drummond M, Sculpher M. Common methodological flaws ineconomic evaluations. Med Care 2005;43(7, Suppl.):5–14.

[11] Piñeros M, Wiesner C, Cortés C, Trujillo LM. HPV vaccine introductionat the local level in a developing country: attitudes and criteria amongkey actors. Cad Saude Publica 2010;26:900–8.

[12] Kahveci R, Meads C. Analysis of strengths, weaknesses, opportunities,and threats in the development of a health technology assessmentprogram in Turkey. Int J Technol Assess Health Care 2008;24:235–40.

[13] Lee KS, Brouwer WB, Lee SI, et al. Introducing economic evaluation as apolicy tool in Korea: will decision makers get quality information? Acritical review of published Korean economic evaluations.Pharmacoeconomics 2005;23:709–21.

[14] Neumann P. Using Cost-Effectiveness Analysis to Improve Health Care:Opportunities and Barriers. Oxford: Oxford University Press, 2005.

Table 1 – Type of economic evaluation and their characteristics.

Type of evaluation by year ofpublication

Type of uncertainity analysis Use ofICER

Discountrate used

Funding source

Cost-effectiveness analysis publishedin 2010

Probabilistic sensitivity analysis ICER used 3% International fundingagency


One-way and multiway sensitivityanalysis

ICER used 3% International fundingagency


One-way and multiway sensitivityanalysis

ICER notused

3% International fundingagency

ICER, incremental cost-effectiveness ratio.


http://www.inahta.org/HTA/

http://www.inahta.org/HTA/































http://refhub.elsevier.com/S2212-1099(13)00085-X/otheref0005






[15] Indian Academy of Pediatrics Committee on Immunization.Recommendations on immunization, 2008. Ind Pediatr 2008;45:635–48.

[16] Seoane-Vazquez E, Rodriguez-Monguio R. Negotiating antiretroviraldrug prices: the experience of the Andean countries. Health Policy Plan2007;22:63–72.

[17] Oliver A. Health economic evaluation in Japan: a case study of: oneaspect of health technology assessment. Health Policy 2003;63:197–204.

[18] Coast J. Is economic evaluation in touch with society’s health values?BMJ 2004;329:1233–6.

[19] Tan-Torres E, Baltussen R, Adum T, et al. Making CHOICES in Health:WHO Guide to Cost-Effectiveness Analysis. Geneva: World HealthOrganization, 2003.

[20] Mittal SK, Mathew JL. Polio eradication in India: the way forward. Ind JPediatr 2007;74:153–60.

[21] Vancelik S, Beyhun NE, Acemoglu H, Calikoglu O. Impact ofpharmaceutical promotion on prescribing decisions of generalpractitioners in Eastern Turkey. BMC Public Health 2007;7:122.

[22] Republic of Rwanda, Ministry of Financial Planning, Ministry of Health.Scaling up to achieve the health MDGs in Rwanda. 2006. Available from:www.oecd.org/dataoecd/34/39/37759625.pdf. [Accessed May 16, 2009].

[23] Teerawattananon Y, Russell S, Mugford M. Systematic reviewof economic evaluation literature in Thailand—are the data goodenough to be used by policy makers? Pharmacoeconomics2007;25:467–79.

[24] Mathew JL. Pneumococcal vaccination in developing countries:where does science end and commerce begin? Vaccine 2009;27:4247–51.

[25] Howitt P, Darzi A, Yang G-Z, et al. Technologies for global health.Lancet 2012;380:507–35.

[26] Doherty J, Kamae I, Lee KKC, et al. What is next for pharmaco-economics and outcomes research in Asia? Value Health 2004;7:118–32.

[27] Carapinha JL, Ross-Degnan D, Wagner AK. Health insurance systems infive sub-Saharan African countries: medicine benefits and data fordecision making. Health Policy 2011;99:193–202.

[28] European Commission. COUNCIL DIRECTIVE 93/42/EEC of 14 June 1993concerning medical devices.

[29] Daar AS, Thorsteinsdóttir H, Martin DK, et al. Top ten biotechnologiesfor improving health in developing countries. Nat Genet2002;32:229–32.





















www.oecd.org/dataoecd/34/39/37759625.pdf























Dossier System as a Practical Tool for Compiling Reimbursement ListsMaria V. Sura, MD, PhD*, Vitaly V. Omelyanovskiy, MD, PhD

National Center for Health Technology Assessment, Moscow, Russia

A B S T R A C T

The article describes the procedure for preparing reimbursement listswith the “Dossier” automated system at the regional level. Basicadvantages and characteristics of the system, procedures for filling outthe application form (dossier) for the inclusion of the drug into reim-bursement lists, and the algorithm of expert evaluation are presented.

Keywords: “Dossier” automated system, health technology assessment(HTA), reimbursement lists.


Introduction

The project of preparing formulary lists of pharmaceutical drugsat various levels, from formularies for particular medical organ-izations to federal lists, has been going on in Russia for nearly 20years. These lists are made by different expert bodies and serve anumber of functions: provide the basis for state regulation ofprices, determination and adjustment of wholesale and retailsurcharges, drug reimbursement, and state supply orders inhealth care.

At present, the system of drug selection at the federal level isthe most transparent and well regulated [1–3], while that at thelevel of medical organizations is the least developed. As for therules and procedures for making reimbursement lists at theregional level, in practice there are no standardized requirementsor criteria for drug evaluation: the principles and rules of drugevaluation, decision-making criteria, and requirements concern-ing the information to be submitted vary to a considerable extent(if such mechanisms exist at all) [4]. The need to standardize therequirements for expert evaluation and submission of informa-tion about the drug, the decision-making criteria, the implemen-tation of the principles of evidence-based medicine, and clinicaland economic analysis (pharmacoeconomics), as well as the needto make the process of decision making for the inclusion ofparticular drugs in the regional reimbursement lists more trans-parent, has led to the creation of an automated system called“Dossier” for the preparation of formulary lists.

This system was originally created as a potential healthtechnology assessment (HTA) tool for compiling reimbursementdrug lists on the regional level. If and when necessary, however,such a system can be easily adapted for the purpose of compilingformulary lists on the federal and hospital/medical organization’slevels.

Currently, there is no accredited official HTA body in Russia.Formulary commissions working on the different levels of the

health care system can serve as prototypes of such HTA bodies.Unified algorithms, criteria, and decision-making rules as well asevidence-based medicine and clinical and economic analysisprinciples that were developed within the framework of theDossier automated system represent an important platform forincorporating the HTA methodology in the activities of theaforementioned expert organizations.

Dossier: An Automated System for the Preparation ofReimbursement Lists

Dossier, an automated system for preparing reimbursement lists(hereafter referred to as “the system”), was developed in theResearch Center for Clinical and Economic Evaluation and Phar-macoeconomics of the N.I. Pirogov Russian National ResearchMedical University in 2010. It is designed to help formularycommissions of the ministries/departments of health in federalsubjects of Russia in their task of preparing reimbursement listsof pharmaceuticals (formulary lists) on the regional level. Theonline system Dossier allows the user to enter, update, store, andevaluate information about the medicines when they are sub-mitted for inclusion into reimbursement lists. The entire cycle ofevaluation is thus supported by the system, from the initialapplication to the final decision of the formulary commission togrant or refuse inclusion of the drug in the reimbursement list.

Three basic principles were taken into account during thedevelopment of the Dossier system: the principle of objectiveevaluation, the principle of intellectual support, and the principleof informatization, or the information technology principle.

The principle of objective evaluation presumes a multilevelassessment of pharmaceutical drugs that includes three mainstages: technical evaluation, scientific evaluation (by nonstaffchief specialist of the regional Department/Ministry of Healthin different fields of medicine), and the final decision to include(or not to include) the drug into the reimbursement list.





E-mail: [email protected].* Address correspondence to: Maria V. Sura, National Center for Health Technology Assessment, POB 88, Moscow 117335, Russia.


dx.doi.org/10.1016/j.vhri.2013.06.012

dx.doi.org/10.1016/j.vhri.2013.06.012

dx.doi.org/10.1016/j.vhri.2013.06.012


dx.doi.org/10.1016/j.vhri.2013.06.012

Technical evaluation is a preliminary stage of processing theapplication (drug dossier), which includes verifying whetherthe data submitted by the applicant meet all relevant require-ments. Scientific evaluation is the stage of assessment by achief specialist. This task calls for expertise in a particularmedical field and a good working knowledge of modern thera-pies and management of the disease in question in real-lifeclinical practice and its current financial burden. The evalua-tion includes an analysis of data on clinical efficacy and cost-effectiveness of the drug, real-life management of patientscovered by the regional reimbursement scheme, the cost ofthe new medication, and so on.

The first two stages of the evaluation take place on the Webpage of the automated system. At the third stage, members of theformulary commission meet and discuss the issue with chiefspecialists, and after their meeting, the decision of the commis-sion is posted on the Internet page.

The principle of intellectual support refers both to the applicant(usually members of the pharmaceutical industry) and to healthcare authorities (regional ministries or departments of health,formulary commissions). To implement this principle, the appli-cation form contains sections that encourage the applicant totake a more rational approach to the task of determiningprioritized indications, target patient groups, and the requestedreimbursement. For example, the system asks to specify thepatient group, the number of patients covered by the regionalreimbursement scheme, the cost of medical treatment of thedisease in question, and the estimated cost to the budget ofpurchasing the drug if it is included in the list for the prioritizedgroup. Thus, the applicant (when requesting that the drug beincluded in the list), the chief specialist of the Department ofHealth (when providing recommendations regarding the expe-diency of adding the new medication to the reimbursement list),and the representative of the Ministry/Department of Health(when deciding whether to include the drug in the list) focus onthe actual costs and the actual patients who are alreadyreimbursed for the treatment of this particular medical condi-tion. Thanks to this approach, it becomes possible to provide aclear justification of the need to include the new drug and theadditional financial costs that it entails, or, alternatively, of theneed to reject the application. The essential considerations thatdetermine whether the drug will be included in the list withinthe framework of the Dossier system are the following: areasonable (scientifically proven) reduction in costs for thebudget, an adequate (effective and economically expedient)deployment of the available resources, and a clear recognitionof the existing situation with regards to the provision ofmedications to patients covered by the regional reimbursementscheme.

The principle of informatization presumes that the submission,evaluation, and scientific analysis of the dossier (application forinclusion in the list) take place online, while the wizard presentsprompts (requirements) that guide the user in the process offilling in the blanks. Besides, informatization of the entire processof adding a new drug to the reimbursement list makes the finaldecision more transparent because the system allows users totrack and time each stage of evaluation and decision making andto view comments by experts. Working online makes it possibleto ensure interregional cooperation of formulary commissions,and the involvement of experts from various fields (more than 25)enables an interdisciplinary approach to selecting the drugs to beincluded in the reimbursement lists.

At the moment of publication, the automated system con-sisted of four independent modules, one for each of the followingregions: Moscow region, Sverdlovsk region, Khanty-Mansi auton-omous area, and Samara region. More regional modules can beadded to the system as required.

Registration and User Levels in the Dossier AutomatedSystem

To start using the system, it is first necessary to choose the rightregion and section and to register. Depending on the accessibilityof particular databases and the procedures used for evaluatingthe submitted applications and for their revision/modification(changing application status), there are five possible user levels:applicant, technical expert, scientific expert, member of theformulary commission, and regional Ministry/Department ofHealth. The lowest level of access to databases of the system isthe applicant level, and the level of regional Ministry/Departmentof Health is the highest.

Access to various databases of the automated system, author-ization to edit/moderate applications (dossiers), menu types, andother options that depend on the user level can be changed(adjusted) depending on the needs of particular regions andmedical organizations.

Status of the Applications Stored in the Dossier AutomatedSystem

Depending on the stage of evaluation, the application (dossier)stored in the automated system is assigned a particular status(moderation). The system specifies by whom (the name ofapplicant and expert), when (time), and how (comments on theevaluation and its result) the application was submitted andevaluated and at which time it was assigned a particular status.The status of an application can be modified in the “Moderation”section. The system supports eight different statuses, for exam-ple, “under evaluation,” “passed” or failed” technical or scientificevaluation, “approved” or “rejected” by the formulary commis-sion, and others. The entire cycle of expert evaluation is thusfully transparent, from the submission of the initial applicationuntil the final decision of the formulary commission.

Application Form for Inclusion in the Regional List and ItsCompletion

The application (dossier) for inclusion of the drug in the regionalreimbursement list contains more than 30 sections that may bedivided into three main blocks: general information about themedication, or the passport block (name and pharmaceuticalgroup of the drug, etc.); study results, or the evidence block(results of clinical and pharmacoeconomic studies); and theregional specification block (the number of patients entitled toreimbursement in the region, local cost of pharmacotherapy forthis disease, as determined by the system based on the submittedInternational Statistical Classification of Diseases, 10th Revision [ICD-10] code, high-priority groups of patients who will be receivingthis medication, including the number of patients, the cost oftherapy, etc.) (Table 1). The application thus contains as muchinformation as possible not only about the drug but also aboutthe current situation with supply of reimbursed medications inthe region. Submitting an application for inclusion in the list andevaluating submitted applications, applicants, chief specialists,and members of the formulary commission together determinethe strategy of drug promotion and its potential/optimal place(“niche”) in the system of drug reimbursement.

Now we describe in greater detail how to fill out the applica-tion and some sections that are particularly problematic in termsof the frequency of mistakes made by the applicants. Thesections that seldom give rise to questions and mistakes arementioned only briefly.

Sections 1 to 8, 10 to 13, and 30 may be called the “passport”part of the application that provides a description of the drug: itsname, pharmaceutical form, pharmacotherapeutic group, groupof Anatomical Therapeutic Chemical classification, grounds for


inclusion in the list, and so on, as well as information about theapplicant. In section 6 (“About the applicant”), it is important toinclude a personal e-mail address (of the person in charge ofcompleting the application) without any spelling mistakes in theaddress. Failure to do so will result in the applicant not being ableto stay in touch with the database and be informed about theprogress of evaluation and its results.

Section 9 asks to specify the disease (or the disease within aparticular target social category) that is the indication for pre-scribing the drug, in accordance with Decree № 890 of the Russiangovernment of July 30, 1994.

In this section, the applicant has to specify whether the drugcan be included in the regional reimbursement list for a partic-ular social or nosological category (a particular medical condition,independently of a social category), or both.

Section 14 of the application asks the applicant to describe theadvantages of the new drug that justify its inclusion in thereimbursement list. The applicant should list the main featuresof the medication that make it preferable to, and different from,other medications with the same indications that are already onthe reimbursement list. Apart from clinical advantages, it isimportant to specify the economic advantages of the drug in thissection (if any), as well as convenience of use (a more convenientregimen, tablets vs. injections, etc.), good compliance with treat-ment, inclusion in international and Russian treatment guide-lines or standards of medical care, and so on. Any advantages of

the drug have to be very specific (preferably numbered) andsupported by evidence from well-designed clinical and pharma-coeconomic studies, approved standards, guidelines, andinstructions.

Section 15, “Clinical efficacy and safety,” requires filling out aseparate form (abstract) describing the clinical studies that arereferred to in this section. Priority is given to systematic reviews,meta-analyses, and randomized controlled double-blind clinicaltrials. Abstracts should contain a description of study design,duration of follow-up, number of patients included in the study,comparators, study results, level of evidence for the efficacy ofthe drug, and so on.

At present, many applicants consciously (exaggerating thequality of the study on purpose) or unconsciously (because oftechnical errors or poor knowledge) make mistakes in thissection. The problem is that incorrect data that contradict theclinical trials submitted for evaluation are entered in the form; forexample, randomization is claimed when it was not performed,and a control group is mentioned where there was none, or thelevel of evidence for drug efficacy is exaggerated. The datapresented in the abstract are always compared with the actualresults of the clinical trial in the course of evaluating theapplication; therefore, the expert will not be led astray by suchincorrect statements.

In section 15A, the applicant has to specify the level ofevidence proving the clinical efficacy of the drug, that is, the

Table 1 – “Dossier” automated system application form: three main blocks and their 30 sections.

Passport block Evidence block Regional specification block

1. International NonproprietaryName (INN)

14. Advantages of a new drug 9. Should the drug be covered fromthe regional budget under thegovernment resolution regulatingreimbursement at the regionallevel?

2. Trade name 15. Clinical efficacy and safety(clinical trials)

17. Clinical trials at the regional level orreal-practice data

3. Original/generic 15А. Level of evidence 18. Priority patients’ groups (age,severity and duration of disease,complications)

4. Included in other reimbursementlists or not

16. Pharmacoeconomics (cost-effectiveness of the drug)

19. Other indications (not forreimbursement list)

5. Manufacturer 20. Cost of 1-y treatment6. About the applicant 21. The number of patients entitled for

reimbursed drugs in the region7. Pharmacotherapeutic group 22. Share and number of patients

eligible for prescribing of a newdrug

8. Anatomical Therapeutic Chemical(ATC) code

23. Share and number of prioritypatients’ groups eligible for anew drug

10. Indications for which the drug isreimbursed

24. Current standard treatmentpractice

11. Dosage 25. Presence of drug in the Standards ofCare, Clinical Guidelines

12. Dosage forms 26. Current cost of medications fortreating of patients for whom theproposed drug is prescribed

13. Course of treatment/lifelongprescription

27. Exclusion of other drugs

30. References 28. Cost for inclusion of a drug into thereimbursement list for prioritypatients’ groups

29. Training of physicians


highest level achieved in clinical trials submitted for evaluation.A list of possible levels of evidence is given in the application.

Section 16 contains data on the cost-effectiveness of the drug.As with clinical efficacy and safety, the applicant has to fill out aseparate form—an abstract of the pharmacoeconomic study. Theabstract should include a description of study design, the type ofpharmacoeconomic analysis, effectiveness criteria, costs, effec-tiveness, cost/effectiveness ratio, etc. As in the clinical efficacyand safety section, applicants often make mistakes here ormisinterpret the results of pharmacoeconomic studies. Often,the results of these studies do not correspond to the dataincluded in the abstract. Thus, costs may be incorrectly describedas direct and indirect, whereas only direct costs were consideredin the study, effectiveness criteria may be described incompletelyor incorrectly, incremental cost-effectiveness ratio may be men-tioned while it was never calculated in the study, and so on.

In section 17, experience in the region, the applicant has to listall clinical trials that were conducted in medical organizations ofthe region for which inclusion in the reimbursement list isconsidered. In practice, however, such studies are very uncom-mon. Accordingly, the applicant will normally simply list allmedical organizations that have already used this drug. Acomplete lack of experience using the drug in this particularregion or by the chief specialist may become the ground forrejection of the application for inclusion into reimbursementlists, even if the chief specialist provides a formal approval. Therehave already been such cases.

In section 18, the applicant is asked to define prioritizedgroups of patients who will receive the drug for a particularindication(s) as far as this is possible (these are the patients whowill be particularly likely to need this medication if it is includedin the reimbursement list). Prioritized patient groups, or seg-ments, may be defined by age, severity and duration of thedisease, complications, comorbidities, and so on. This section iscompleted together with a chief specialist who is a member ofthe formulary commission. In practice, it may be hard for theapplicant (who is usually a representative of the pharmaceuticalcompany that manufactures the drug) to define the prioritizedgroup(s) of patients. First, the applicant is not well informedabout all the characteristics of regional reimbursed patients.Second, the applicant’s a priori position is that all patients towhom such treatment is indicated and who are entitled toreimbursement should be given this particular drug. This reason-ing, however, fails to take into account the types of therapycurrently in use and the results of comparative analysis of theefficacy, safety, and cost-effectiveness of alternative therapiesalready on the list.

In section 19, the applicant has to indicate whether the drugmay be used for other indications than the one chosen as themain justification for inclusion in the list (section 9 based on ICD-10 and section 10 based on instructions for use). In other words, isthere a possibility that there will be a “leakage” of the drug for thetreatment of other medical conditions or forms of disease if it isincluded in the list?

Section 20 asks the applicant to estimate the cost of treatmentwith the new medication per patient per year. The cost should becalculated for the categories listed in section 9 in accordance withICD-10. The source of price information, the day when calcula-tions were performed, and the steps of cost estimation should beprovided in the field for text comments.

Sections 21 to 23 are devoted to quantitative analysis ofpatients on the regional reimbursement list for the given ICD-10 code.

The number of patients on the regional reimbursement list forthe given ICD-10 code (section 21) is determined automatically,thanks to an inbuilt database of regional reimbursement (region-specific), which contains information about the number of

prescriptions for the previous year according to the statistics ofregional ministries/departments of health. Sections 22 and 23 areconcerned with “quantitative prioritization” of patients entitledto reimbursement. The percentage and number of patients whomay be prescribed the reimbursed drug are estimated in theformer and the percentage and number of patients who shouldhave priority in drug supply are estimated in the latter. Once thepercentages are entered, the number of such patients is calcu-lated automatically: in the former case (section 22) as a share ofthe total number of patients entitled to reimbursement in theregion (section 21), and in the latter case (section 23) as a share ofthe total number of patients entitled to reimbursement in theregion who may be prescribed the evaluated drug (section 22).High-priority patient groups are always determined by the chiefspecialist or members of the formulary commission. The appli-cant, if this is a representative of a pharmaceutical company,does not have enough competence or authority to make suchdecisions.

In section 24, the applicant has to describe the existingapproaches to the management of patients with this disease inthe region. Two approaches are possible when filling out thissection. Either the chief specialist may provide this description or(the more formal approach) a query may be made for theappropriate ICD-10 code to collect data about the prescription ofdrugs to patients entitled to reimbursement over the last year.This information can be found in the automated system itself asa separate database. If the second approach is used, it isimportant to filter the results obtained from the database,selecting from the list only those drugs that are directly relevantto the therapy of this particular disease (have the same indica-tions as the evaluated drug). The reason for this is that thedatabase search yields information about all the medicationsdispensed to reimbursed patients under a particular ICD-10 code(which often refers to symptomatic treatment, therapy for exac-erbations or comorbidities, etc.).

In section 25, the applicant submits information about theinclusion of the evaluated drug in Russian and foreign regula-tions governing drug acquisition and prescription, such as the listof vital and essential drugs, standards of medical care, nationaland foreign treatment guidelines, and so on.

Section 26 of the application describes the cost of managingpatients with the medical condition for which the evaluated drugis indicated. This section also contains a database listing the costof prescribed medications for different ICD-10 codes over thelast year.

Section 27 of the application discusses the possible need toexclude or limit the use of other medications previously includedin the regional list in case the new drug is added. This section isparticularly relevant in case of regional budget deficit and shouldbe completed by a chief specialist on the basis of a detailedanalysis of the use of medications in the framework of theregional reimbursement scheme.

In section 28, the applicant has to estimate the costof including the evaluated drug in the list for prioritized reim-bursement groups. The following formula should be used forthis calculation: the cost of treating one patient with thismedication for 1 year (s. 20) � the number of patients entitledto reimbursement in the region who have a high priority for theprovision of this medication (s. 23) � savings due to exclusion/limited use of other medications (s. 27). This calculation shouldbe repeated for all conditions (ICD-10 codes) mentioned in theapplication.

Section 29 describes the need for special training of physiciansconcerning the prescription of the new drug in their practice.Overall, because the new medication will be prescribed and usedin ambulatory practice, any possible difficulties with dosing,administration, special training, and so on have to be minimal.


The applicant has to list all the sources of information used asreferences in the completed application in the bibliography(section 30). Bibliography can be arranged alphabetically orin the order of citation and should include the author(s) ofeach study (surname and initials), study/regulation title, andpublisher.

The application can be completed in several sessions. Incom-plete applications can be saved as drafts and continued later at aconvenient time.

Before previewing and submitting the application, the systemperforms a preliminary check. If the checkup is not successful,the user is prompted to fill in the incorrect or missing fields.

According to experts and representatives of the regionalministries of health, the Dossier automated system is conven-ient for use and allows one to make evidence-based decisionstaking fully into account all the preparation’s characteristics,subgroup analysis results, and capacities of the local budgets. Onaverage, the full expert cycle of one preparation (from themoment of the application’s submission in the system to theformulary commission’s final decision) takes 1.5 months. It cantake longer than that if there are mistakes in the Dossierapplication form (which may lead to the return of the applicationfor updating) and depending on the frequency of the formularycommissions’ meetings, which make a final decision on listing(in some regions, formulary commissions’ meetings take placenot more than once a quarter). Special trainings are arranged forexperts and potential applicants (trainings for experts are con-ducted separately) before the implementation of the Dossiersystem in the practical activities of the regional ministries ofhealth. Following completion of the theoretical section of thetraining, the participants are offered to work in the systemonline independently, for example, to enter application andperform evaluation of the drug.

The number of new drugs that receive positive assessment byexperts of formulary commissions and are listed may varysignificantly depending on the region, number, and compositionof people entitled to the regional pharmaceutical benefits pro-grams, financial situation in the region as well as the backlog ofapplications submitted for listing. For example, in the Moscowregion, the number of drugs that were included in reimburse-ment lists during 1 year did not exceed 15 names. The totalnumber of submissions through the Dossier system, however,was approximately twice as large.

Algorithm of Decision Making Based on an Analysis of theApplication (Dossier) for Inclusion in the Reimbursement List

The decision to include or not to include the evaluated drug inthe list has to be reached by the chief specialist and members ofthe formulary commission on the basis of a multicriteria analysisof the data pertaining to the evaluated drug, the number ofreimbursed patients in the region, the high-priority groups towhom the drug is indicated, the cost of therapy, and so on. Thus,an expert making a decision has to have an intimate knowledgeof both WHAT is recommended for inclusion in the list and FORWHOM this drug is added to the list (as regards the number ofpatients and the cost). The stages of expert evaluation ofsubmitted data leading to the final decision are presented inFig. 1.

Adherence to this algorithm, a careful consideration of allpros and cons of adding the drug to the list, and a stepwiseevaluation will allow us to optimize the purchase of medications,choosing from a large number of drugs suggested for inclusion inlimited lists the most expedient in terms of both clinical andeconomic characteristics and the real demands of pharmacolog-ical therapy and the budget resources.

Conclusions

The creation of the Dossier automated system is an initiative ofthe Research Center for Clinical and Economic Evaluation andPharmacoeconomics of the N.I. Pirogov Russian State MedicalUniversity with the support of several regional ministries anddepartments of health. Today, the actual implementation of thisautomated system is a unique experience of only a handful ofregions. The system, however, is ready for implementation and, ifnecessary, can be adjusted for use in every federal subject ofRussia.

The advantages of setting up an automated system for thecompilation of regional lists are self-evident. For instance, we canmention the following: a formal procedure for the completion ofthe application form (dossier) and expert evaluation; a stepwiseevaluation and multidisciplinary approach; decision makingbased on a comprehensive, multicriteria analysis of data on thenew drug; transparency of all decisions (every stage of theevaluation is reflected in the system, including the grounds forevery decision and the time of decision); a shift to electronicdocumentation, which can save time and money (the applica-tions can be completed, evaluated, and edited online); and anopportunity to process data on the new medication analytically,considering pooled data on pharmacotherapeutic and Anatomi-cal Therapeutic Chemical groups, nosological categories, and soon. Furthermore, the system has a significant educational com-ponent that helps the user build a logical chain to promote thenew drug on the market and helps the chief specialist todetermine whether there is a real need to add this drug to thelist, as well as to estimate the requisite purchase size and budgetexpenses.

Fig. 1 – Algorithm for evaluation of the application (dossier)for inclusion in the regional reimbursement list.


If such a system was implemented in federal subjects,this could provide standard algorithms of expert evaluation,decision-making criteria, and requirements concerning the datato be submitted with the application for inclusion in the lists. Asmore data on different medications are accumulated in regionalautomated systems, a single “databank” can be created with thehelp of both experts and applicants that will include comprehen-sive information about the efficacy, safety, and economic accept-ability of reimbursed drugs, as well as the cost and quantitativecharacteristics of pharmacotherapies of various medical conditions.


R E F E R E N C E S

[1] Sura M V, Omelyanovsky V V. Evolutsiya sistemy expertizy priformirovanii Perechnya zhiznenno neobhodimyh i vazhneyshih

lekarstvennyh preparatov. Meditsinskiye tehnologii. Otsenka ivybor 2011;3(5):30–3. In Russian [Sura MV, Omelyanovsky VV. Theevolving assessment system for preparing the list of vital andessential pharmaceuticals. Medical Technologies. Assessment andChoice.]

[2] Prikaz Ministerstva zdravoohraneniya i sotsialnogo razvitiya Rossi-yskoy Federatsii ot 27.05.2009 g. No276n “O poryadke formirovaniyaproekta Perechnya zhiznenno neobhodimyh i vazheyshih lekarstven-nyh sredstv”. In Russian [Order 276n of the Ministry of Health and SocialDevelopment of the Russian Federation of 27.05.2009, “On preparing alist of vital and essential pharmaceuticals”.]

[3] Proekt prikaza Ministerstva zdravoohraneniya i sotsialnogo razvi-tiya Rossiyskoy Federatsii ot 28.04.2011 g. “O poryadke formirova-niya proekta Perechnya zhiznenno neobhodimyh i vazheyshihlekarstven- nyh sredstv”. In Russian. [Draft order of the Ministry ofHealth and Social Development of the Russian Federation of28.04.2011, “On preparing a list of vital and essential pharma-ceuticals”.]

[4] Omelyanovsny V V. Avxentyeva M V. Soldatova I G, et al. Kliniches-kaya i ekonomicheskaya exper- tiza pri formirovanii perechneylekarstvannyh sredstv. Meditsinskiye tehnologii. Otsenka i vybor2010;1:28–31. In Russian. [Pharmaco-economic evaluation for preparingformulary lists of pharmaceuticals. Medical Technologies. Assessmentand Choice.]






























Impact of the Pharma Economic Act on Diffusion of Innovation andReduction of Costs in the Hungarian Prescription DrugMarket (2007–2010)Rok Hren, PhD, MSc, IHP (HE)�

University of Ljubljana, Ljubljana, Slovenia

A B S T R A C T

Objective: In this study, we examined the impact of the PharmaEconomic Act, which was introduced in Hungary in 2007. Methods:We used detailed data on the Hungarian prescription drug market,which had been made publicly available by the authorities. Weevaluated the effect of the Pharma Economic Act on both dynamicand static efficiencies and also on equity, which has been historicallya controversial issue in Hungary. We analyzed the overall prescriptiondrug market and statin and atorvastatin markets; as a proxy fordetermining dynamic efficiency, we examined the oncology drugmarket for some specific products (e.g., bortezomib) and the long-acting atypical antipsychotic drugs market. Results: There is nodenying that the authorities managed to control the overall prescrip-tion drug costs; however, they were still paying excessive rents for off-patent drugs. Examples of oncology and long-acting atypical anti-psychotic drugs showed that the diffusion of innovation was on per-capita basis at least comparable to G-5 countries. While the share of

out-of-pocket co-payments markedly increased and the reimburse-ment was lowered, the concurrent price decreases often meant thatthe co-payment per milligram of a given dispensed drug was actuallylower than that before the Act, thereby benefiting the patient.Conclusions: It appears that strong mechanisms to control volumerather than price on the supply side (marketing authorization holders)contained the drug expenditure, while offering enough room to strivefor innovation. Making data on prescription drug expenditures andassociated co-payments publicly available is an item that should bedefinitely followed by the surrounding jurisdictions.

Keywords: austerity measures, cost-containment, prescription drugspending.


Introduction

In this study, we examined the impact of the Pharma EconomicAct (PEA) [1], which was introduced in Hungary in 2007. Themotivation to analyze this particular legislative measure within asingle midsize country is in its comprehensiveness and uniqueapproach toward marketing authorization holders (MAHs). More-over, the recent economic crisis is making such a “laboratory ofcost-containment tools” attractive for authorities and payers notonly among the Central and Eastern European (CEE) countries butalso among Western (e.g., European Union [EU]-15) jurisdictions.The final reason for the analysis is the availability of detailed dataon the Hungarian prescription drug market, which are providedto the public by the authorities [2].

Controlling prescription drugs spending is fraught with diffi-culties, and particularly notable are examples of the GP Fund-holding scheme in the United Kingdom [3–5] and prescribing drugbudgets [5–7]. The difficulties arise because of multifactorialreasons for pharmaceutical expenditure growth, which are usu-ally addressed only partially with national prescription drugpolicies. Typically, the authorities/payers introduce two groupsof policies: one group on the so-called demand side (e.g.,

physicians, pharmacists, and patients) and one group on thesupply side (e.g., MAHs). If we use a simple equation thatexpenditure for prescription drugs E is

E¼∑pi � Vi

where pi is the price of a given drug and Vi is its correspondingvolume, then supply-side policies target mostly the price side ofthe equation, while demand-side policies work mostly on thevolume side of the equation. For example, international referencepricing [8] has been widely embraced by authorities/payers in CEEjurisdictions as a particular supply-side tool geared towardregulating prices of primarily branded drugs; while such anapproach may indeed result in a short-term reduction in pre-scription drug expenditure, the long-term effects may be ambig-uous because the volume of prescribed drugs may readilyexpand.

Hungarian authorities/payers with PEA took a somewhatdifferent approach and included—among a plethora of othermeasures—strong mechanisms to control volume on the supplyside, with MAHs (i.e., branded and generic firms) being respon-sible to cover any overshoot of the preagreed volume. In thatsense, PEA is unique among CEE jurisdictions; however, at its




Conflicts of interest: The author has indicated that he has no conflicts of interest with regard to the content of this article.


� Address correspondence to: Rok Hren, University of Ljubljana, Jadranska 19, 1000 Ljubljana, Slovenia.


dx.doi.org/10.1016/j.vhri.2013.06.013

dx.doi.org/10.1016/j.vhri.2013.06.013

dx.doi.org/10.1016/j.vhri.2013.06.013


dx.doi.org/10.1016/j.vhri.2013.06.013

conception there were serious warnings that the austerity actmay jeopardize innovation and could even lead to withdrawals ofbreakthrough branded drugs [9]. Accordingly, we are here specif-ically evaluating the effect of the PEA on both dynamic and staticefficiencies [10–12], which present well-known and fundamentalmicroeconomic framework, and also on equity, which has beenhistorically a controversial issue in Hungary (see, e.g., [13–15]).Throughout this article, we will present data in local currency—Hungarian forint (HUF), with the conversion rate (as of March 28,2013) of 1 euro (EUR) for 304.24 HUF, which means that 1 billionHUF equals approximately 3.3 million EUR.

Policy Drivers of the PEA

It is impossible to consider the PEA outside of the wider packageof the Hungarian governmental austerity measures aimed atreducing the then (in 2006) whopping budget deficit of 10.1% ofthe gross domestic product (GDP) to the Maastricht level of 3% ofthe GDP by 2010, with special focus on reducing the health carespending by 0.9% of the GDP by 2009. While Hungary spent 8.3%of the GDP on health care in 2006 (vs. 8.9% of the Organisation forEconomic Co-operation and Development [OECD] members' aver-age), the prescription drug spending appeared a convenient cost-containment target because of (1) its high proportion in the totalhealth care expenditure (in 2006, 26.7% vs. 15.5% of the OECDaverage) [16] and (2) the fifth highest growth rate in the 1998 to2003 period among the OECD members, following Ireland, Korea,the United States, and Australia, which all are jurisdictions withsubstantially higher GDP per capita than Hungary [17].

National Health Insurance Fund-Országos EgészségbiztosításiPénztár (NHIF-OEP) data [2] reveal that total public health

expenditure stood at 6% in 2006, with public pharmaceuticalexpenditure of 25%; the compound annual growth rate (CAGR) inthe period 1998 to 2006 was 13%.

Although these data appear intuitively supportive of austeritymeasures, it is noteworthy that population health of Hungary hasbeen severely lagging behind the EU-15 jurisdictions; for example, lifeexpectancy at birth is 73.3 years versus 80.9 years in Sweden, infantmortality rate is 5.82 versus 2.56 in Sweden, and coronary heartdisease death rate is 169 versus 30 in France [18]. Hungary’s populationhealth indicators are among the worst even among CEE countries.

Cost-Containment Tools of the PEA

Cost-containment measures had not started in 2007: The NHIF-OEP already in 2003 mutually agreed with the pharmaceuticalindustry to introduce the so-called claw-back system (firmsrequired to cover overspending the budget for reimbursabledrugs) and later negotiated with the industry the price freeze ofdrugs in exchange for allowing new products to enter thereimbursement list. The industry in fact under the claw-backregime paid back to the NHIF-OEP the amount of 20 billion HUF(66 million EUR) and 23 billion HUF (76 million EUR) in 2005 and2006, respectively. The PEA thus built on this general claw-backsystem by bringing additional broad cost-containment policiesdescribed below to the prescription drug market.

Supply-side control—MAHsIn this review, we will focus on cost-containment policies appliedto the MAHs. The first bold measure was an introduction of 12%statutory rebate/payback on reimbursed expenditure for bothbranded and generic firms, which substantially reduced the profit

0

100,000

200,000

300,000

400,000

500,000

600,000

700,000

800,000

2003 2004 2005 2006 2007 2008 2009 2010

Actual

Actual minus 12% rebate

Projected /GR 6%/

Projected /CAGR 03-06/

mio HUF 2003 2004 2005 2006 2007 2008 2009 2010Actual 228,050 257,370 316,196 364,250 295,845 310,973 332,929 350,416Actual minus 12% rebate 364,250 260,344 273,656 292,978 308,366Projected /GR 6%/ 364,250 386,105 409,271 433,828 459,857Projected /CAGR 03-06/ 364,250 425,784 497,713 581,794 680,078

year-on-year Actual change, including rebate 13% 23% 15% -29% 5% 7% 5%

Savings 07-10 = Actual minus Projected /GR 6%/ 90,260 98,298 100,899 109,441 398,89812% rebate 07-10 35,501 37,317 39,951 42,050 154,820

Fig. 1 – Public prescription drug expenditure in Hungary between 2003 and 2010 [2], shown in million HUF. Two counterfactualcurves are displayed: one for 6% annual growth rate and one for the CAGR of 16.9% for the period 2003 to 2006. CAGR,compound annual growth rate; GR, growth rate; HUF, Hungarian forint.


of the firms. The 12% rebate could be proportionately decreasedby the firm’s voluntary price decrease (e.g., 10% price decreasewould lead to 10.8% rebate) or could be partially or from 2010onwards fully waived because of the firm’s local R&D investment.As of July 1, 2011, this rebate was due to financial crisis increasedto 20% for the duration of 3 years [19].

Second, the PEA required firms to pay sales representativeregistration fee in the amount of 5 million HUF (about 16,500 EUR)per pharmaceutical sales representative per annum, thereby increas-ing firms’ operational costs and consequently decreasing theirprofits; the registration fees remained in power in terms of monthlyfees (416,000 HUF) despite the opposition of the pharmaceuticalindustry; more so, as of July 1, 2011, the fee has been increased to anannual level of 10 million HUF (33,000 EUR). Particularly domesticgeneric firms reacted to these fees by employing sales representa-tives as market researchers, while the branded firms reduced thesize of their sales force; the overall number of sales representativesin Hungary fell from 2700 to 2300 because of the PEA [20].

Third, price-volume agreements based on ad-hoc negotiationsbetween the branded firm and the NHIF-OEP were put in placealready in 2003; on the request of the pharmaceutical industry,these agreements have stayed confidential despite the fact thatthe NHIF-OEP in principle preferred transparency. Any overspend-ing beyond the agreed-upon expenditure (which in most casesalso includes wholesalers’ and pharmacists’ margins, and valueadded tax [VAT]) would need to be paid by the firm to the NHIF-OEP. Price-volume agreements were implemented on three levels:

1. a single product per 1 year;2. a single product per 3 years (e.g., 20%, 35%, and 45% of the

3-year expenditure within the first, second, and the third year,

respectively, although the overall expenditure remainedcrucial);

3. multiple products, sometimes even with multiple indications(e.g., biologics), where the payback is calculated by the marketshare of an individual drug within the group.

Price-volume agreements have been applied by the NHIF-OEPto new entries since 2003 and essentially constituted the micro-control of the overall prescription drug budget.

In addition to these policies controlling the expenditure, thatis, the “volume,” well-known price controls were strengthened bythe PEA, such as internal (“therapeutic”) reference pricing, quar-terly revised by the two-step “iterative” bidding process of thefirms. As of July 1, 2011, the bidding has been blind, precludingfirms to resubmit their prices.

Demand-side controlAmong other demand-side policies, the PEA decreased reim-bursement levels from 30% to 25%, 60% to 55%, and 90% to 85%,while to 100% reimbursement level the mandatory prescriptionfee in the amount of 300 HUF (1 EUR) was attached, with themaximum total prescription fee per patient per annum of 16,667HUF (55 EUR). Socially disadvantaged patients (approximately 5%of the population) had prescription fees and co-payments cov-ered by the special welfare fund set up by the government.

Assessment of the PEA Cost-Containment Tools in Practice

Overall prescription drug marketWe first examined the overall prescription drug market by simplyassessing the change in the public expenditure and co-payments

0

50,000

100,000

150,000

200,000

250,000

300,000

350,000

400,000

2003 2004 2005 2006 2007 2008 2009 2010

Public expenditure

Public expenditure minus 12% rebate

Social welfare "co -pay"

Co-payment

mio HUF 2003 2004 2005 2006 2007 2008 2009 2010Public expenditure 228,050 257,370 316,196 364,250 295,845 310,973 332,929 350,416

Public expenditure minus 12% rebate 364,250 260,344 273,656 292,978 308,366

Social welfare "co-pay" 15,789 15,893 18,320 19,062 18,943 17,835 17,986 18,326

Co-payment 73,695 82,412 95,956 104,634 119,205 113,137 115,782 116,201

Share of co-payment as of total

prescription drug market, rebate

included 23% 23% 22% 21% 30% 28% 27% 26%

year-on-year change of co-payment 12% 16% 9% 14% -5% 2% 0%

Fig. 2 – Public prescription drug expenditure along with out-of-pocket co-payment and “co-payment” covered by the socialwelfare for disadvantaged population in Hungary between 2003 and 2010 [2], shown in million HUF. HUF, Hungarian forint.


made by patients and to this end, we perused NHIF-OEP data,which are publicly available [2]. Co-payments are directly linkedto equity because of their regressive nature. Figure 1 shows publicprescription drug expenditure in Hungary between 2003 and 2010[2]; for the sake of assessing the impact of the PEA, two counter-factual curves were derived: one for 6% annual growth rate andone for the CAGR of 16.9% for the period 2003 to 2006, whichclearly indicates that the public expenditure on prescriptiondrugs was unsustainable and that reforms were indeed manda-tory. The effect of the PEA cost-containment policies due to pricereductions is evident, and the cummulative amount “saved” inthe period 2007 to 2010 was 399 billion HUF (1.3 billion EUR) if wetake conservative 6% annual growth curve as the comparator. Wemodeled additional savings of the NHIF-OEP by the lump sum12% rebate because the data on specific measures had not beenavailable before 2010: our estimated cummulative amount invarious forms of paybacks/rebates was for 2007 to 2010 155 billion

HUF (510 million EUR) (Fig. 1), thus bringing the grand totalsavings to 554 billion HUF (1.81 billion EUR). In fact, our “12%-rebate” estimate turns out to be conservative if we compare for2010 our lump sum 12% rebate of 42 billion HUF with publiclyavailable 2010 figures [2], which indicate that the overall OEP-NHIF revenue amounted to 50.4 billion HUF (165 million EUR): The12% rebate was actually 30.5 billion HUF (due to reductionsarising from R&D investment discussed above), the total salesrepresentative fees were 11.8 billion HUF, the wholesaler feeswere 0.55 billion HUF, the price-volume agreements for productstotaled 8.8 billion HUF, and there was no overall claw-back.

The public expenditure for prescription drugs decreased by18.8% in 2007 versus 2006 or by 29% (Fig. 1) if we take into accountour lump sum 12%-rebate estimate; our calculation is close to theBusiness Monitor International report [21], which appraised thatin 2007 versus 2006 the NHIF-OEP prescription drug expenditurecontracted by 30.5%. As shown in Figure 1, the prescription

0

5,000

10,000

15,000

20,000

25,000

30,000

2006 2007 2008 2009 2010

Atorvastatin

Other statins

Simvastatin

Total statin market (C10AA)

Public expenditure for statins minus 12% rebatemio HUF 2006 2007 2008 2009 2010Atorvastatin 9,682 8,236 10,943 12,698 13,247Other statins 6,440 3,992 3,018 3,719 4,766Simvastatin 9,311 6,662 5,162 4,110 2,849

Total statin market (C10AA) 25,433 18,890 19,123 20,528 20,862year-on-year change -26% 1% 7% 2%Simvastatin + ezetimibe 0 183 1,096 1,585 1,861Atorvastatin + amlodipine 0 0 305 546 627

Fig. 3 – Public expenditure, including 12% rebate, for statins (ATC4 group C10AA) between 2006 and 2010 [2], shown in millionHUF. HUF, Hungarian forint.

Table 1 – Public expenditure along with out-of-pocket copayment for statins (ATC4 group C10AA) between 2006and 2010 [2].

Total statin market

2006 2007 2008 2009 2010

Public expenditure 25,433 21,466 21,731 23,327 23,707Social welfare “co-pay” 13 354 464 551 647Co-payment 3,789 4,879 4,639 5,560 6,373Total market, no rebate 29,235 26,699 26,834 29,438 30,727Public expenditure minus 12% rebate 25,433 18,890 19,123 20,528 20,862Share of co-payment as of total statin drug market, rebate included (%) 13 20 19 21 23Year-on-year change of co-payment (%) 29 �5 20 15

Note. Values are in million HUF unless otherwise indicated.HUF, Hungarian forint.


expenditure grew moderately in the period between 2007 and2010, with a CAGR of 5.8%. The major reason for the 18.8%decrease was extensive price cuts of both branded and genericdrugs and an increase in co-payment of patients (see below). Theprice cuts throughout 2007 were truly extensive; in April 2007alone, the prices of 1000 drugs were reduced on average by 16%[2]. While the overall claw-back on total public expenditureremained in place even after 2007, it has never been realized: itseems that all other measures in the view of the NHIF-OEPsufficiently contained the costs.

Figure 2 shows the public prescription drug expenditure alongwith out-of-pocket co-payment and “co-payment” covered by thesocial welfare for disadvantaged population in Hungary between2003 and 2010. The out-of-pocket co-payment increased by 14%in 2007 versus 2006, and with public expenditure falling, its shareof the total prescription market surged sharply (30% in 2007 vs.21% in 2006). As shown in Figure 2, the absolute increase in co-payments in 2007 was around 15 billion HUF (50 million EUR),which is less than that estimated by the Business MonitorInternational report [21], putting the range at 20 to 50 billionHUF. An increase in co-payment was due to two factors: (1)decreasing the reimbursement level for a number of drugs and (2)a 300-HUF prescription fee. Co-payment was relatively stablefrom 2008 onwards and even slightly declined both in absolutevalue and in its relative share; in 2010, co-payment amounted to26% of the total prescription drug expenditure, which is still asubstantial figure [2].

Social welfare “co-payments” remained stable since 2005,being around 4% of the total prescription drug expenditure [2].Opposition against co-payments is considerable in Hungary, andin 2008 public with large majority rejected on the referendum co-payments for physicians’ visits and in-patient stays although co-payments for prescription drugs were not voted on. In the surveyof 2007, 11% of the people stated that prescription drug fees of 300HUF, which are capped at an annual maximum of 16,667 HUF (55EUR), would make otherwise fully reimbursed drugs unaffordable

[15]; recent local data (in the period from April 2010 throughMarch 2011) however showed that 97% and 75% of the populationpay less than 6,667 HUF (22 EUR) and 1,667 HUF (5.5 EUR) perannum, respectively.

Specific prescription drug marketsIn the previous section, the NHIF-OEP data provide strong evi-dence that the PEA resulted in cost-containment of the publicexpenditure for prescription drugs in Hungary. Cost-containmentitself, however, may not lead to efficiencies either in the prescrip-tion drug market, let alone in the overall health care market. As iswell known, cost-containment may stifle innovation, therebyreducing the dynamic efficiency; however, price reductions ofoff-patent drugs and their generic versions may be insufficient,allowing their prices to remain way above the marginal cost ofproduction and thus promoting static inefficiencies [11,22,23].

Statins (Anatomical Therapeutic Chemical [ATC4] groupC10AA)To assess static efficiency, we examined a group of statins (ATC4group C10AA), and within this group in more detail atorvastatins(ATC5 group C10AA05). Statins are suitable proxies for highlygenericized therapeutic group at the time of introduction ofthe PEA.

Figure 3 shows the public expenditure for statins, taking intoaccount the 12% rebate. Similarly as the overall market, the statinmarket contracted by 26% in 2007 (vs. 2006) and grew modestly atthe CAGR of 2.9% in the period 2008 to 2010. Statins accounted for8.0% of the total public expenditure in 2006 and 6.3% in 2010. In2006, simvastatin and atorvastatin were the dominant molecules,about equal in public expenditure and jointly accounting forabout 75% of the total statin market. From 2007 onwards, we havewitnessed a clear shift in prescribing habits of physicians asatorvastatin became the leading molecule, reaching a marketshare of 63% of the entire C10AA market by 2010. The decline in

0

5,000

10,000

15,000

20,000

25,000

2003 2004 2005 2006 2007 2008 2009 2010

Public expenditure

Projection /trend 04-06/

Social welfare "co-pay"

Co-payment

Public expenditure minus 12% rebate

atorvastatin

mio HUF 2003 2004 2005 2006 2007 2008 2009 2010Public expenditure 2,267 3,685 6,496 9,682 9,359 12,435 14,430 15,053Projection /trend 04-06/ 9,682 12,681 15,679 18,678 21,677Social welfare "co-pay" 0 0 0 6 126 217 317 389

Co-payment 908 612 845 1,492 1,836 2,362 3,402 3,935

Public expenditure minus 12% rebate 9,682 8,236 10,943 12,698 13,247

Share of co-payment as of total atorvastatin market, rebate included 29% 14% 12% 13% 18% 17% 21% 22%year-on-year change of co-payment -33% 38% 77% 23% 29% 44% 16%

Fig. 4 – Public expenditure along with out-of-pocket and social welfare co-payments for atorvastatins (ATC5 groupC10AA05) between 2006 and 2010 [2], shown in million HUF. A counterfactual curve based on the trend in the period 2004 to2006 is displayed. HUF, Hungarian forint.


simvastatin is notable even if we add the uptake of combinationof simvastatin and ezetimibe, which lies outside the C10AAATC group.

Table 1 outlines the public expenditure for statins along without-of-pocket and social welfare co-payments for the period 2006to 2010. As in the total prescription drug market, the share of out-of-pocket co-payment increased from 13% in 2006 to 20% in 2007,while in the absolute amount, it increased by 29%. The hike wasmostly due to higher co-payments and the unwillingness ofpatients to switch to cheaper substitution: study of GfK Hungarianoted that “only one tenth of population chooses their medica-tion based on price” and while general practitioners offeredcheaper alternatives to 61% of their patients, only 30% of thepatients actually agreed to the substitution [21].

Atorvastatin (ATC5 group C10AA05)Among statins, the most important single molecule is atorvasta-tin, which has been in the Hungarian prescription market for

years among the top 10 leading molecules. In 2010, atorvastatin’sshare of the total prescription market was 4.3%. Figure 4 showsthe public expenditure for atorvastatin along with out-of-pocketand social welfare co-payments for the period 2003 to 2010. Acounterfactual curve was constructed on the basis of averageyear-on-year added expenditure between 2004 and 2006: weestimate that in 2007 and 2008 alone, the PEA measures saved9.2 billion HUF (30.6 million EUR). Out-of-pocket co-paymentincreased steadily since 2004, with the share of co-paymentrepresenting 12% to 14% of the total atorvastatin market. Sim-ilarly as in the total and statin market, the co-payment shareincreased to 18% in 2007 from 13% in 2006.

So far, we have not related contracting of public expenditureto the trend in the number of atorvastatin units and one couldintuitively argue that the lower expenditure simply resulted frompatients failing to fulfill their prescriptions or patients coveringthe difference by out-of-pocket co-payment. Yet, Figure 5 showsthat the total milligrams of atorvastatin issued at pharmaciesgrew faster (þ32%) in 2007 versus 2006 than either public

0

500,000

1,000,000

1,500,000

2,000,000

2,500,000

2003 2004 2005 2006 2007 2008 2009 2010

10 mg

20 mg

40 mg

80 mg

atorvastatin - packs-per-30-tablets equivalents

2003 2004 2005 2006 2007 2008 2009 201010 mg 316,913 53,171 102,943 217,556 275,965 393,193 456,090 460,14620 mg 41,473 22,926 165,500 499,472 730,063 1,331,798 1,773,735 1,935,30440 mg 52,610 284,359 449,340 637,291 773,674 1,194,641 1,588,795 1,755,123

80 mg 0 0 13,167 32,139 62,320 80,226 97,807 97,099

Total mg 183,089,700 370,937,700 700,991,600 1,206,832,700 1,598,804,100 2,543,148,300 3,342,358,800 3,638,411,400

year-on-year change - total mg 103% 89% 72% 32% 59% 31% 9%

Fig. 5 – Number of units (in per-30-tablets equivalents) and total milligrams of atorvastatin dispensed in Hungarianpharmacies between 2003 and 2010 [2].

Table 2 – Average out-of-pocket co-payment per milligram and public expenditure per milligram (with 12%rebate included) in Hungary for all strengths (10 mg, 20 mg, 40 mg, and 80 mg) of atorvastatin between 2003 and2010 [2].

2003 2004 2005 2006 2007 2008 2009 2010

Co-payment per milligram (HUF) 5.0 1.6 1.2 1.2 1.1 0.9 1.0 1.1Year-on-year change (%) �67 �27 3 �7 �19 10 6Public expenditure per milligram (HUF), 12% rebate included 12.4 9.9 10.3 8.5 4.1 3.3 2.9 2.8Year-on-year change (%) �20 �7 �13 �36 �16 �12 �4

HUF, Hungarian forint.


expenditure (�15%) or out-of-pocket co-payment (þ23%). Thetrend continued in 2008, with total milligrams increasing by 59%,public expenditure by 33%, and co-payment by 29%. In other words,this means that both the NHIF-OEP and patients were paying lessper milligram than before the implementation of the PEA, as iscorroborated by Table 2. Table 2 reveals two important additionaltrends, though: since 2009, out-of-pocket co-payment per milli-gram started increasing again, while the decrease in public expen-diture per milligram (with included 12% rebate) markedly sloweddown. Both trends clearly show the unwillingness of the off-patentbranded firm and generic firms to further reduce prices.

Figure 6 substantiates this claim by explicitly showing anexample of the price dynamics for atorvastatin 40 mg during theperiod following the implementation of the PEA. The genericparadox [24,25] could not apply here because the off-patentbranded drug would be delisted if its price was increased,although the branded drug does retain some premium over

generics as has been noted elsewhere [11,26]. The premium ofoff-patent originator appears a logical choice in the Hungarianmarket because of patients’ relative price insensitivity. Lingeringpresence of the off-patent originator and price convergence ofgenerics, however, strongly suggest that both the NHIF-OEP andpatients may be financing rents to which neither the branded firmnor generic firms are entitled. But can we estimate the price thatwould be close to the marginal cost of production of atorvastatinwithout having direct access to this proprietary information?

Table 3 shows the ATC5 market share of the off-patentatorvastatin originator, which encountered the first genericentries toward the end of 2004 and rapidly lost the share in 10-mg and 20-mg strengths. As indicated by Table 4, the number ofentries grew larger with each year and at least formally formedthe conditions for the perfectly competitive market, whichshould drive the prices toward the marginal cost of produc-tion [27,28]. In Table 2, we can deduce that the average price

0

1000

2000

3000

4000

5000

6000

7000

8000

Jan-07 Jan-08 Jan-09 Jan-10 Jan-11

Pri

ces

(in

HU

N)

- at

orva

stat

in 4

0 m

g x

30

Off-patent originator

Generic #1

Generic #2

Generic #3

Generic #4

Generic #5

Generic #6

Fig. 6 – Prices (in HUF) of atorvastatin 40 mg (30 tablets) for an off-patent originator and various generic bioequivalentalternatives between January 2007 and January 2011 [2]. HUF, Hungarian forint.

Table 3 – ATC5 market share of off-patent atorvastatin originator in units per strength and in publicexpenditure [2].

2003 2004 2005 2006 2007 2008 2009 2010

Off-patent originator ATC5 market share (%) in units10 mg 100 100 41 21 0.0120 mg 100 99 24 16 0.002 0.00140 mg 100 100 92 64 27 14 9 780 mg 81 69 50 50 42 37

Off-patent originator ATC5 market share (%) in public expenditureAll strengths 100 100 88 53 16 9 6 5

Table 4 – Number of drug firms marketing various atorvastatin strengths in the Hungarian prescription drugmarket between 2003 and 2010 [2].

2003 2004 2005 2006 2007 2008 2009 2010

10 mg 1 2 4 7 8 9 10 1320 mg 1 2 4 8 8 10 11 1340 mg 1 2 5 6 9 11 12 1680 mg 0 0 3 3 4 4 4 4


(equaling public price plus co-payment) per milligram of the thenbranded atorvastatin with monopoly position in the market wasbetween 11.5 HUF (in 2004) and 17.4 HUF (in 2003). Smallmolecules launched in 1990s had—as a general rule of thumb—cost of goods sold (COGS) typically at 10% (or even lower) oftheir monopoly price. From this rather crude, but unlikely inac-curate, estimate, the price between 1.15 HUF per mg and 1.74 HUFper mg should be comfortable enough and should even retainsome considerable rent for generic firms. Table 2 reveals that in2010 the average price per mg was 4.7 HUF, about three timeshigher than the estimate above. On the basis of this estimate, wemay conclude that the NHIF-OEP could easily save two thirds of13.2 billion HUF (i.e., 8.8 billion HUF—29 million EUR) in its annualexpenditure for atorvastatin and patients could save two thirds oftheir annual out-of-pocket co-payments of 2.9 billion HUF (i.e., 1.9billion HUF—6.3 million EUR).

Long-acting injectable risperidone (ATC5 group N05AX08)To examine dynamic efficiency, we first analyzed the long-actinginjectable risperidone (administered once every 2 weeks), whichrepresents a line extension of the oral risperidone, thus address-ing the issue of adherence of patients suffering from schizophre-nia [29–32] and associated higher hospitalization rates and costs[33,34]. The MAH/branded firm entered in 2005 into a price-volume agreement with the NHIF-OEP, which was renewed in2008. The controlled nature of growth is obvious from theexamination of total milligrams dispensed (Fig. 7) and publicexpenditure including 12% rebate (Table 5). Because the NHIF-OEPdid not resort to international reference pricing once the brandeddrug had entered the market, the price per milligram remainedunchanged until the 12% rebate was mandated. In spite of the

controlled nature of growth, the long-acting atypical antipsy-chotic achieved in Hungary diffusion that was globally surpassedonly by Spain; in April 2011, its market share in the totalantipsychotic market (N05A) was 23.7% (vs. 26.7% in Spain,17.1% in France, 13.4% in Italy, 10.9% in Germany, and 6.8% inthe United Kingdom) [35].

Co-payment surged in 2007, and the absolute amount hasbeen with an average monthly dose of 75 mg around 525 HUF (1.7EUR), which is likely not excessive.

Oncology drugs (ATC5 groups L01XX32 and L01XE01)Oncology drugs are good proxies for examining dynamic effi-ciency; here, we will consider bortezomib (ATC5 group L01XX32).As with the long-acting atypical antipsychotic, its MAHs/brandedfirm entered into price-volume agreements with the NHIF-OEP in2005 (Figure 8). Bortezomib’s consumption per capita was in 2010in Hungary on par with that in Italy and the most developed CEEjurisdiction Slovenia, and although lower than in Spain, France,and Germany, it was markedly higher than in the United King-dom [36].

Since 2007, co-payment amounted to 300 HUF per prescribed3.5-mg bortezomib vial (Table 6). As with the long-acting atypicalantipsychotic, the absence of international reference pricingimplies that public expenditure per milligram changed only in2007 with the introduction of the 12% rebate.

Discussion

To our knowledge, this is the first study to systematicallyevaluate the impact of the PEA in the period between 2007 and

Fig. 7 – Total milligrams of long-acting atypicalantipsychotics dispensed in the Hungarian prescriptiondrug market between 2005 and 2010 [2].

Table 5 – Public expenditure along with out-of-pocket co-payment for the long-acting atypical antipsychoticdispensed in the Hungarian prescription drug market between 2005 and 2010 [2].

2005 2006 2007 2008 2009 2010

Public expenditure 1,093 3,168 3,633 4,067 4,546 4,950Social welfare “co-pay” 0 0 1 3 4 5Co-payment 0 0 24 26 27 29Total market 1,093 3,168 3,658 4,096 4,578 4,984Public expenditure minus 12% rebate 3,197 3,579 4,001 4,356

Public expenditure per milligram (HUF), rebate included 1,133 1,128 984 983 983 983

Co-payment per milligram (HUF) 0.3 0.0 7.2 7.1 6.7 6.4


Fig. 8 – Number of units of bortezomib dispensed in theHungarian prescription drug market between 2004 and 2010[2].


2010 and the results of our study could have potential implica-tions for prescription drug policy in Hungary and other jurisdic-tions within CEE and EU-15 regions.

There is no point denying that the NHIF-OEP managed tocontrol the prescription drug costs by implementing the PEA inJanuary 2007. The PEA caused a substantial decrease in prices ofoff-patent drugs and imposed a 12% rebate on pharmaceuticalfirms. While the share of out-of-pocket co-payments markedlyincreased and the reimbursement was lowered, the concurrentprice decreases often meant that the co-payment per milligramof a given dispensed drug was actually lower than that before theact, thereby benefiting the patient. We also have no reason tobelieve that the cost-containment policies were excluding inno-vation: examples of long-acting atypical antipsychotic and borte-zomib suggest that the Hungarian prescription market under thePEA has enabled diffusion of new technologies, on per-capitabasis comparable to those of G-5 countries.

The NHIF-OEP set up before the PEA an excellent informationtechnology system for tracking the number of prescription drugsdispensed and associated public expenditure and co-payments;while most of the jurisdictions in the EU have in place electronicprescribing systems, the analysis provided to the public by theNHIF-OEP is exemplary in the CEE region and maybe even wider.

With such accolades, it is almost impossible to resist thetemptation in adopting the NHIF-OEP “model” all over Europe.The PEA, however, has some serious shortcomings, which arepromoting inefficiencies. An obvious inefficiency is that the PEAis still financing excessive rents for off-patent drugs, whichremain a global challenge for payers [22]. The Hungarian off-patent market reinforces in particular the general notion thatregulating generic prices may lead to price convergence andinefficiencies [7,37] and that generic competition is not suffi-ciently fierce in a regulated environment.

It may well be that the NHIF-OEP also needed to make indesign of its policies “compromises” with the local pharmaceut-ical industry (e.g., Gedeon-Richter), which are significant localemployers and contributors to local taxes. As an example, NHIF-OEP’s continued efforts of international-nonproprietary-nameprescribing were blocked by the dominant generic playerGedeon-Richter.

A 12% rebate (currently 20% rebate and for drugs older than 5years 30% rebate) imposed on the pharmaceutical industrydeserves a comment. This was indeed a bold move that couldlead to a serious unintended consequence of branded firmsleaving the Hungarian market en masse as their profits andparticularly profit versus net-trade-sales ratios plummeted waybelow the expectations of their shareholders. The fact thatbranded firms stayed behind indirectly speaks for the relativelylow production costs of prescription drugs, which is also thereason why protection of intellectual property rights is needed

[38,39]. In this way, the NHIF-OEP not only collected financialrebates but also optimized social welfare via the optimization ofpayments for drugs. Overall, NHIF-OEP’s 12% or even 20% rebatecould be interpreted as a hostile act against mostly branded andforeign generic firms (as local generic firms are exempt from itbecause of their local R&D investment), which indeed led todownsizing of branded firms’ sales forces and reduction in theirpromotional capabilities. The advantage of the 12% rebate is in itssimplicity of implementation and in avoiding legal hassle whenex-factory prices potentially decrease below the COGS—particu-larly of some older products with usually low expenditure butthat are critical for health care (e.g., fentanyl ampoules). Anotherargument might be that the 12% rebate is just a starting tool thatcould be always accompanied by a reduction in ex-factory priceseither via international reference pricing or the so-called risk-sharing mimicking the UK bortezomib scheme [40,41].

The PEA was launched into the milieu of co-payments whilemany jurisdictions in the CEE region still largely exempt patientsfrom co-payments. When introducing co-payments de novo, theadverse effect on adherence could be severe. Also, co-paymentsin Hungary are collected diligently, while it may well be thatadministration costs of applying co-payments may exceed insome instance the amount of co-payment. Further studies wouldbe needed to address this point.

Our study has some obvious limitations. We have examineda limited number of products and one might well find theproduct or a group of products that would deviate from ourconclusions. In fact, we are planning to extend our analysis tothe wider range of ATC groups, including antihypertensives andproton pump inhibitors and a wider number of innovative drugs.We also plan to examine the impact of the reforms that wereintroduced after July 1, 2011, and which we have referred to inthe text.

One may argue that deriving the marginal cost of productionfrom the assumption of 10% COGS is invalid and may at bestapply to a limited number of drugs. A good counterargument is areal-life case from Romania in which in March 2005 the firstgeneric after the patent expiry entered the antipsychotic marketwith 12% to 15% (depending on strength) of the price of the off-patent originator [42]. Other generics quickly followed by decreas-ing the price even below the 10% mark of the original monopolyprice. Another more recent example also from the Romanianpharmaceutical market is an entry of a generic version of anoncological drug at 7% of the originator’s price [43].

Population health in Hungary is among the worst in Europe,and it may well be that the PEA might adversely contribute to themid-term and long-term health prognosis. Our analysis cannotassess this impact. In the same way, we cannot address whetherlower expenditure on drugs increased the use of other health careservices (e.g., on the secondary or tertiary levels) via the so-called

Table 6 – Public expenditure along with out-of-pocket copayment for bortezomib dispensed in the Hungarianprescription drug market between 2005 and 2010 [2].

2005 2006 2007 2008 2009 2010

Public expenditure 366 1,551 1,729 1,934 2,303 2,562Social welfare “co-pay” 0 0 0.0 0 0 0Co-payment 0 0 1.6 2 2 2Total market 366 1,551 1,731 1,936 2,305 2,565Public expenditure minus 12% tax 1,521 1,702 2,027 2,255

Public expenditure per milligram (HUF), rebate included 88,227 87,792 77,186 77,182 77,182 77,016

Co-payment per milligram (HUF) 20.0 0.0 79.4 84.0 84.8 85.1



squeezed balloon effect because we have not considered expen-ditures outside of the prescription drug market.

Finally, which are the tools of the NHIF-OEP that could beshared with other jurisdictions in the CEE region and beyond?The overall philosophy of focusing on volume rather than price isa good guiding beacon and price-volume agreements appear towork and also provide enough room for innovation to strive. Inother particulars, the tools to be shared without modifications areless obvious due to shortcomings/inefficiencies discussed above.Making data on prescription drug expenditures and associatedco-payments publicly available is an item that should be defi-nitely followed by the surrounding jurisdictions.

Source of financial support: These findings are the result ofwork supported by the author. The views expressed in this articleare those of the author, and no official endorsement by theUniversity of Ljubljana is intended or should be inferred.

R E F E R E N C E S

[1] Pharma Marketletter. Pharma-Economic Act now passed by theHungarian Parliament. December 4, 2006. Available from: http://www.highbeam.com/doc/1G1-155673153.html. [Accessed May 31, 2013].

[2] National Health Insurance Fund – Orszagos Egeszebiztositasi Pentzar.2011. Available from: http://www.oep.hu/portal/page?_pageid=35,21341107&_dad=portal&_schema=PORTAL and http://www.oep.hu/portal/page?_pageid=35,21341860&_dad=portal&_schema=PORTAL(in Hungarian). [Accessed May 31, 2013].

[3] Harris CM, Scrivener G. Fundholders’ prescribing costs: the first fiveyears. BMJ 1996;313:1531–4.

[4] Gosden T, Torgerson DJ. The effect of fundholding on precribing andreferral costs: a review of the evidence. Health Policy 1997;40:103–14.

[5] Walley T, Mossialos E. Financial incentives and prescribing. In:Mossialos E, Mrazek M, Walley T,eds., Regulating Pharmaceuticals inEurope: Striving for Efficiency, Equity, and Quality. Maidenhead: OpenUniversity Press, 2004.

[6] Busse R, Howerth C. Cost containment in Germany: twenty yearsexperience. In: Mossialos E, Le Grand J,eds., Health Care and CostContainment in the European Union. Aldershot: Ashgate, 1999.

[7] Danzon PM, Liu H. Reference Pricing and Physician Drug Budgets: TheGerman Experience in Controlling Pharmaceutical Expenditures.Working paper. Philadelphia, PA: Wharton School, University ofPennsylvania, 1996.

[8] Mrazek M, Mossialos E. Regulating pharmaceutical prices in theEuropean Union. In: Mossialos E, Mrazek M, Walley T,eds., RegulatingPharmaceuticals in Europe: Striving for Efficiency, Equity, and Quality.Maidenhead: Open University Press, 2004.

[9] Pharma Marketletter. Hungary’s new drug act “needs revision” (PharmaEconomic Act). February 26, 2007. Available from: http://www.highbeam.com/doc/1G1-160184516.html. [Accessed May 31, 2013].

[10] Santerre RE, Neun PE. Health Economics: Theories, Insights, and IndustryStudies(5th ed.). Mason, OH: South-Western College Learning, 2010.

[11] Hren R. Overview of economic principles of competition in prescriptiondrug markets. Farmakoekonomika a liekova politika 2011;7:3–11.

[12] Gaal P. Health Care in Transition – Hungary. Copenhagen: EuropeanObservatory on Health Care Systems, 2004.

[13] McGuire A, Raikou M, Kanavos P. Pricing pharmaceuticals: value-basedpricing in what sense. Eurohealth 2008;14:3–6.

[14] Mrazek M, de Jonchere K, Petrova G, Mossialos E. The pharmaceuticalsector and regulation in the countries of Central and Eastern Europe.In: Mossialos E, Mrazek M, Walley T,eds., Regulating Pharmaceuticalsin Europe: Striving for Efficiency, Equity, and Quality. Maidenhead:Open University Press, 2004.

[15] IHS Global Insight. IHS Global Insight Report: Hungary (Healthcare andPharma). Budapest, Hungary: IHS Global Insight, Inc., February 23, 2010.

[16] OECD health data. 2006. Available from: http://www.oecd-ilibrary.org/social-issues-migration-health/data/oecd-health-statistics_health-data-en.[Accessed May 31, 2013].

[17] OECD health data. 2005. Available from: http://www.adbi.org/3rdpartycdrom/2005/06/13/1342.public.health.data/. [AccessedMay 31, 2013].

[18] United Nations, Department of Economic and Social Affairs, PopulationDivision. United Nations Population Division 2005-2010 data. New York,NY: United Nations, 2013.

[19] IHS Global Insight. Hungarian pharma law change will affect industryadversely: Egis warns of R&D freeze. 2011. Available from: http://www.ihs.com/products/global-insight/industry-economic-report.aspx?ID=1065930004. [Accessed May 31, 2013].

[20] National Pharmaceutical Institute – Orszagos Gyogyszerszeti Intezet.2011. Available from: http://www.ogyi.hu/nyitooldal/ (in Hungarian).[Accessed May 31, 2013].

[21] Business Monitor International. Hungary Pharmaceuticals & HealthcareReport Q1 2009. London: Business Monitor International, 2009.

[22] Kanavos P. Generic policies: rhetoric vs. reality. Euro Observer2008;10:1–6.

[23] OECD. OECD Policy Roundtables: Generic Pharmaceuticals. OECD, 2010.Available from: http://www.oecd.org/competition/abuse/46138891.pdf[Accessed May 31, 2013].

[24] Grabowski HG, Vernon JM. Brand loyalty, entry, and price competitionin pharmaceuticals after the 1984 Drug Act. J Law Econ 1992;35:331–50.

[25] Frank R, Salkever D. Generic entry and the pricing of pharmaceuticals. JEcon Manag Strat 1997;6:75–90.

[26] Vandoros S. Generic policies and the ‘generics paradox’. Euro Observer2008;10:7–8.

[27] Folland S, Goodman AC, Stano M. The Economics of Health and HealthCare(6th ed.). London: Pearson, 2010.

[28] Frank RH. Microeconomics and Behavior(8th ed.). New York: McGraw-Hill, 2010.

[29] Wahlbeck K, Tuunainen A, Ahokas A, Leucht S. Dropout rates inrandomised antipsychotic drug trials. Psychopharmacology (Berl)2001;155:230–3.

[30] Valenstein M, Ganoczy D, McCarthy JF, et al. Antipsychotic adherenceover time among patients receiving treatment for schizophrenia: aretrospective review. J Clin Psychiatry 2006;67:1542–50.

[31] Lieberman JA, Stroup TS, McEvoy JP, et al. Effectiveness ofantipsychotic drugs in patients with chronic schizophrenia. N Engl JMed 2005;353:1209–23.

[32] Keith SJ, Kane JM. Partial compliance and patient consequences inschizophrenia: our patients can do better. J Clin Psychiatry2003;64:1308–15.

[33] Gilmer TP, Dolder CR, Lacro JP. Adherence to treatment withantipsychotic medication and health care costs among Medicaidbeneficiaries with schizophrenia. Am J Psychiatry 2004;161:692–9.

[34] Kozma C, Grogg A. Medication compliance and hospitalization inschizophrenia. Paper presented at Sixth Annual Meeting of the College ofPsychiatric and Neurologic Pharmacists, May 1–4, 2003, Charleston, SC.

[35] IMS. Long-acting injectable risperidone data in European Union.Danbury, CT: IMS Health, 2011.

[36] IMS. Bortezomib data in European Union. Danbury, CT: IMS Health,2010.

[37] Danzon PM, Ketcham JD. Reference pricing of pharmaceuticalsfor Medicare: evidence from Germany, the Netherlands andNew Zealand. In: Cutler DM, Garber AM (eds.), Frontiers in Health PolicyResearch Vol. 7. National Bureau of Economic Research andMIT Press, 2004.

[38] Beier FK. The significance of the patent system for the technical,economic and social progress. Intern Rev Indus Prop Copyright1980;11:563-71.

[39] Landes WM, Posner RA. The Economic Structure of IntellectualProperty Law. London: Harvard University Press, 2003.

[40] Towse. Value based pricing, research and development, and patientaccess schemes: will the United Kingdom get it right or wrong? Br J ClinPharmacol 2010;70:360–6.

[41] Jack A. NHS drug adviser questions drug policy. Financial Times,August 24, 2010.

[42] Canamed. Romanian national price registry of medicines. 2005.Available from: http://www.ms-preturi.ro/msf-dgf26/download.htm,historical data available from the Romanian Ministry of Health uponrequest. [Accessed May 31, 2013].

[43] Canamed. Romanian national price registry of medicines. 2011.Available from: http://www.ms-preturi.ro/msf-dgf26/download.htm,historical data available from the Romanian Ministry of Health uponrequest. [Accessed May 31, 2013].



http://www.highbeam.com/doc/1G1-155673153.html



http://www.oep.hu/portal/page?_pageid=35,2134&!QJ;1107&_dad=portal&_schema=PORTAL




http://www.oep.hu/portal/page?_pageid=35,21341860&_dad=portal&_schema=PORTAL









































http://www.oecd-ilibrary.org/social-issues-migration-health/data/oecd-health-statistics_health-data-en

http://www.oecd-ilibrary.org/social-issues-migration-health/data/oecd-health-statistics_health-data-en

http://www.adbi.org/3rdpartycdrom/2005/06/13/1342.public.health.data/

http://www.adbi.org/3rdpartycdrom/2005/06/13/1342.public.health.data/






http://www.ihs.com/products/global-insight/industry-economic-report.aspx?ID=1065930004




http://www.ogyi.hu/nyitooldal/







http://www.oecd.org/competition/abuse/46138891.pdf

















































http://www.ms-preturi.ro/msf-dgf26/download.htm


http://www.ms-preturi.ro/msf-dgf26/download.htm



Performance Assessment of Ga District Mutual Health Insurance Scheme,Greater Accra Region, GhanaEric Nsiah-Boateng, MPH1,*, Moses Aikins, PhD2

1Claims Department, National Health Insurance Scheme, Ashiedu Keteke Submetro Office, Accra, Ghana; 2Health Policy, Planning and Management Department,School of Public Health, University of Ghana, Accra, Ghana

A B S T R A C T

Objective: This study assessed performance of the Ga District MutualHealth Insurance Scheme over the period 2007-2009. Methods: Thedesk review method was used to collect secondary data on member-ship coverage, revenue, expenditure, and claims settlement patternsof the scheme. A household survey was also conducted in the MadinaTownship by using a self-administered semi-structured questionnaireto determine community coverage of the scheme. Results: The studyshowed membership coverage of 21.8% and community coverage of22.2%. The main reasons why respondents had not registered with thescheme are that contributions are high and it does not offer theservices needed. Financially, the scheme depended largely on sub-sidies and reinsurance from the National Health Insurance Authorityfor 89.8% of its revenue. Approximately 92% of the total revenue wasspent on medical claims, and 99% of provider claims were settledbeyond the stipulated 4-week period. Conclusions: There is anincreasing trend in medical claims expenditure and lengthy delay in

claims settlements, with most of them being paid beyond themandatory 4-week period. Introduction of cost-containment meas-ures including co-payment and capitation payment mechanismwould be necessary to reduce the escalating cost of medical claims.Adherence to the 4-week stipulated period for payment of medicalclaims would be important to ensure that health care providers arefinancially resourced to deliver continuous health services to insuredmembers. Furthermore, resourcing the scheme would be useful forspeedy vetting of claims and also, community education on theNational Health Insurance Scheme to improve membership coverageand revenue from the informal sector.

Keywords: claims settlements, Ghana, membership coverage, NationalHealth Insurance Scheme..


Introduction

Many low- and middle-income countries are challenged withhow to finance their health care systems to achieve universalcoverage of health services. In 2005, the member states of theWorld Health Organization adopted a resolution encouragingcountries to develop health financing systems aimed at providinguniversal coverage [1]. This was defined as securing access for allto appropriate promotive, preventive, curative, and rehabilitativeservices at an affordable cost.

In the 1990s, a number of mutual health organizations wereestablished in Ghana, with funding and technical support fromexternal partners. Most of these mutual health organizations,however, primarily focused on providing financial protectionagainst the potentially catastrophic costs of a limited range ofinpatient services for the disadvantaged people in society [2]. TheNational Health Insurance Scheme (NHIS) was introduced in 2004to build on these organizations and provide comprehensivehealth services to all citizens in Ghana [3].

The National Health Insurance Act, Act 650, was passed intolaw in Ghana in 2003 through the Legislative Instrument (LI 1809),

though implementation in terms of access to benefits began inNovember 2005 [3–5]. Its policy objective is that “within the nextfive years, every resident of Ghana shall belong to a healthinsurance scheme that adequately covers him or her againstthe need to pay out-of-pocket at point of service use in order toobtain access to a defined package of acceptable quality healthservices” [6]. The NHIS was designed as a mandatory healthinsurance system, with risk pooling across district schemes,funded from members’ contributions and a levy on the value-added tax charged on selected goods and services [3–6].

As a key social sector initiative to support the Ghana PovertyReduction Strategy II policy objective of ensuring sustainablefinancial arrangements that protect the poor, the NHIS’s perform-ance and long-term sustainability are significant. The perform-ance assessment of the scheme is also key to Ghana’s attainmentof the Millennium Development Goals 1, 3, 4, and 5.

Since its full implementation in 2005, the NHIS has beenfacing major structural and administrative challenges, includingsignificant delays in issuing membership cards; lack of a uni-form contribution system across schemes, which has implica-tions for portability and equity within the national scheme; and






* Address correspondence to: Eric Nsiah-Boateng, MPH, Claims Department, National Health Insurance Scheme, Ashiedu KetekeSubmetro Office, Derby Avenue, Accra, Ghana.


dx.doi.org/10.1016/j.vhri.2013.06.005

dx.doi.org/10.1016/j.vhri.2013.06.005

dx.doi.org/10.1016/j.vhri.2013.06.005


dx.doi.org/10.1016/j.vhri.2013.06.005

considerable delays in provider claims reimbursement [3,7,8]. Anindependent health sector review report shows that at the end of2008, the health facilities had outstanding claims worth GH¢49million [7]. A number of initiatives and activities by researchersand development partners aimed at tackling performance chal-lenges facing the District Mutual Health Insurance Schemes(DMHISs) have taken place. Most of these research and initiatives,however, focused on inventories of DMHISs and access, utiliza-tion, and quality of care and were largely uncoordinated. More-over, key findings and associated recommendations were leftunimplemented [9]. Therefore, little is known about performanceelements of membership coverage, revenue mobilization, expen-diture, and medical claims settlements of the DMHISs. The studyaimed at filling this gap by providing performance assessment ofGa DMHIS.

According to the 2010 NHIS Annual Report, there are 145DMHISs nationwide, with 10 in the Greater Accra region [10]. TheGa DMHIS is the biggest DMHIS in the Greater Accra region interms of catchment area. The Ga district has a large number ofsuburban and rural communities, making it suitable for thisstudy. Madina is the largest cosmopolitan settlement in thedistrict, which also made it appropriate for the household survey.This article reports performance assessment of the scheme forthe 2007-2009 period and recommendations for improving itsoperations.

Conceptual Framework

The conceptual framework for the study was adopted from theWorld Health Organization proposed framework for health sys-tems performance assessment and modified to reflect the Inter-national Labour Organization’s core performance indicators forassessing social health insurance schemes [11,12]. According tothe framework, high membership coverage, high revenue base,low expenditure, and prompt settlement of provider claimsenhance performance ratios such as coverage rate, renewal rate,expense ratio, and claims ratio, which, in turn, results in highperformance and an improved health status of the target pop-ulation (Fig. 1).

Methods

Study Area

The study was conducted at Ga DMHIS and Madina Township, allin the Ga district of the Greater Accra region. The Ga DMHIS has astaff strength of nine and 74 contracted health care providers.The Ga district lies in the northern part of the Greater Accraregion and is bounded in the north by Akuapim South district, inthe east by Tema Municipal, and in the south by Accra Metro-polis. It has three subdistricts, namely, Ga South, Ga East, and GaWest, with 594 communities comprising mixed settlements:urban, periurban, and rural areas. The district has an estimatedpopulation of 891,609 and a growth rate of 4.4%. There are 58health facilities in the district comprising public, private, andChristian Health Association of Ghana facilities. The main eco-nomic activities in the district are public service, trading, farming,and craftsmanship.

Study Design

The study was a cross-sectional survey of households and aretrospective analysis of membership, revenue, and expenditurerecords of the Ga DMHIS for the period 2007-2009. The studypopulation consisted of membership data of Ga DMHIS andselected heads of surveyed households in the Madina Township.

Data Collection Method

Desk reviewDocuments on membership, operational reports, audited reports,financial statements, and claims payment books of the schemewere reviewed. The registration files were reviewed in terms ofthe number of people registered, number of membership cardsissued, and number of renewals for each year under review. Theaudited accounts for 2007-2008 and unaudited accounts for 2009were examined for total contribution collected, subsidies receivedfrom the National Health Insurance Authority (NHIA), donorsupport, and other internally generated funds. Information onadministrative expenditure and medical bills was also collected.The claims submission registers and claims payments for 2009were reviewed to determine the number of days between sub-mission of claims and reimbursement.

Face-to-face interviewA community household survey was conducted in the MadinaTownship to determine the community coverage rate. A multi-stage sampling method was used to select the study subjects. Inall, 376 household heads were sampled on the basis of anestimated prevalence rate of 43% membership coverage, a con-fidence level of 95%, and 5% margin of error. The questionnairecovered background characteristics and membership in the NHIS.The household membership section looked at knowledge on theNHIS, membership status, and reasons for not enrolling into thescheme.

Data Analysis

The performance indicators for the study were analyzed asfollows.

Coverage rateMembership files for the period 2007-2009 were reviewed todetermine the total number of valid card-bearing members ineach year. The coverage rate was determined by dividing the totalnumber of valid card-bearing members in each year by theestimated district population in the same year.

Community coverage rateThis was estimated by dividing the total number of participantswith valid membership cards as of March 2010 by the totalnumber of participants interviewed.

Annual revenueThis was estimated by adding the total amount of moneycollected from contributors, received from the NHIA and otherdonor agencies, and investment returns in each year of theperiod under review (2007–2009).

Annual expenditureThis was estimated by adding administrative expenses andmedical claims expenses in each year of the period 2007-2009.The administrative expenses comprised salaries, membershipcards’ processing cost, and other operational costs, while themedical claims expenses constituted payment of outpatient,inpatient, specialized medical services and essential medicinesin the NHIS minimum benefit package.

Expense ratioThis was estimated by dividing the total administrative expensesincurred in each year by the total amount of contributionscollected in the same year for the period 2007-2009.


Claims ratioThis was estimated by dividing the total medical claims expensesincurred in each year by the total amount of contributionscollected in the same year for the period 2007-2009.

Combined ratioThis was calculated by adding the expense ratio and theclaims ratio.

Promptness of claims settlementsAll provider claims submitted within the October-December 2009period were reviewed to determine the number of days it took tosettle them. The settlements days for each provider claim weregrouped according to a defined schedule of 0 to 28 days, and morethan 28 days. Based on the stipulated period of 4 weeks for claims

settlements, all claims paid within 0 to 28 days were consideredas prompt payment and those paid beyond 28 days as delayedpayment.

Ethical considerationEthical clearance for the study was obtained from Ghana HealthService Ethical Review Committee on Research Involving HumanSubjects. Approval was sought from the Ga DMHIS to use its data,and permission was sought from the Municipal Health Services toconduct the household survey in the Madina Township.

Participants were informed of the objective of the study andthat they were free to participate and to leave at any point.A total of 365 household heads out of the 376 sampled voluntarilyprovided signed informed consent and were interviewed.

High membership coverage

-Registered members with valid ID Cards

High revenue

-Premium

-NHIA subsidy

-Other income

Low expenditure

-Administrative

-Medical bills (claims)

Prompt provider claims settlement

High renewal rate and growth rate

Low expense ratio and claims ratio

Improved health status of the population

High performance

Fig. 1 – Conceptual framework for assessment of Ga DMHIS. DMHIS, District Mutual Health Insurance Scheme; NHIA, NationalHealth Insurance Authority.

Table 1 – Membership by category (2007–2009).

Year Districtpopulation

Formalsector

Informalsector(18–69 y)

SSNITpensioners

Aged(≥70 y)

Youngerthan18 y

Indigent Pregnantwomen

Totalcoverage(%)

2007 749,022 4,957 19,035 237 1,990 25,598 831 NA 7.02008 782,715 9,794 42,318 1,171 4,027 47,240 2,920 11,997 15.32009 817,924 15,943 55,561 3,484 7,909 63,512 2,920 28,594 21.8

NA, not applicable (the free maternal care policy was introduced in 2008); SSNIT, Social Security and National Insurance Trust.


Limitations of the studyFirst, there were no separate data for active and nonactivemembers between the 2007 and 2008 period, making it difficultto determine the true membership coverage of card-bearingmembers in that period. Second, the financial audit report for2009 was not available; hence, revenue and expenditure datawere retrieved from account ledger books and financial state-ments, which might not give the true picture of financial flowsin 2009.

Results

Coverage Rate

In general, the membership coverage of the scheme increasedfrom 7% in 2007 to 21.8% in 2009 (Table 1). The number ofregistered members, however, dropped from 8.3 percentagepoints in the first year (2007–2008) to 6.5 in the second year(2008–2009). In all, the informal sector constituted one-third ofthe registered members; there were no registrations of pregnantwomen in 2007 and indigent in 2009.

Community coverage rateOut of the total number of 365 household heads who participatedin the survey, 81 (22.2%) were registered members with valid IDcards. The main reasons why some of the respondents had notregistered with the scheme were “contribution is expensive” and“does not offer services needed” (Table 2).

Revenue and expenditureIn all, the expenditure of the scheme for the period under studyexceeded the revenue. This occurred in 2007 and 2008 but wasreversed in 2009 by a significant increase in NHIA support (Fig. 2).In 2009, the scheme generated 60.9% of the total revenue andexpended 41.8%. The total revenue generated and the expendi-ture incurred for the period under study were GH¢12.8 millionand GH¢13.7 million, respectively.

Annual revenueWhile proportions of NHIA support in the form of subsidy for theexempt group and reinsurance for claims payment increasedsignificantly from 82.9% in 2008 to 93.9% in 2009, contributionscollected and other income generated showed a decreasing trend(Fig. 3).

Annual expenditureThe main expenditure areas of the scheme were administrativeand medical claims. The total expenditure for the period underreview doubled from GH¢2.2 million in 2007 to GH¢5.8 million in2008, and dropped slightly in 2009 (i.e., GH¢5.7 million) (Table 3).The proportion of administrative expenses showed a downwardtrend from 18.2% in 2007 to 3.6% in 2009, while that of medicalclaims expenses went up from 81.8% in 2007 to 96.4% in 2009.

Claims settlementsAbout 99% of 38,737 medical claims reviewed for the periodOctober-December 2009 were paid beyond the stipulated periodof 28 days for claims settlements.

Expense, claims, and combined ratiosThe expense ratio showed a downward trend from 2.7 in 2007 to0.7 in 2009, while the claims ratio increased from 12.2 in 2007 to18.1 in 2009 (Table 4). The combined ratio, which is the sum of theexpense ratio and the claims ratio, increased from 15.0 in 2007 to18.8 in 2009.

Discussion

The membership of the scheme is categorized into formal sectorworkers, informal sector workers, pensioners, the aged (70 yearsand older), children younger than 18 years, indigent, and pregnantwomen. The membership coverage increased over the periodunder study, driven mainly by the exempt groups: childrenyounger than 18 years, the aged, indigent, and pregnant women;only one-third of the informal sector group who pays contribu-tions is registered. The increasing trend in annual membership

Table 2 – Reasons for not registering with theGa DMHIS.

Reason (n ¼ 284) %

Contribution is expensive 35.3Not sick now 10.5Hospital is too far 1.0Treat elsewhere 9.5Does not offer services needed 33.2Belong to other district scheme (NHIS) 10.5Total 100

DMHIS, District Mutual Health Insurance Scheme; NHIS, NationalHealth Insurance Scheme.

12.2

26.9

60.9

16.3

42 41.8

0

10

20

30

40

50

60

70

2007 2008 2009

Perc

enta

ge (

%)

Years

Revenue (GHc Expenditure (GHc

Fig. 2 – Revenue and expenditure status (2007–2009).


coverage is good for membership growth and development of thescheme as shown in previous studies [11,13,14]. It indicates thatregistered members have accepted the NHIS program and areready to pool their resources to seek a measure of protection fromthe risks that they face [13]. The small proportion of registeredinformal group who pays contributions, however, could affect therevenue base and presents long-term sustainability problems [3].In spite of the increasing trend in annual membership coverage,the number of registered members declined from 8.3 percentagepoints in the first year (2007–2008) to 6.5 in the second year (2008–2009), a trend found in other studies [3,8,15]. The cause of thisdecline in membership registration could be attributed to reasonsfound in the household survey.

The household survey shows a membership coverage of 22%,which gives credence to that obtained from the desk review atthe scheme office. Most of the respondents considered theannual contribution of the scheme, which ranges from GH¢10.00 to GH¢24.00, as expensive, and hence they did not registerfor the scheme. Others believe that the scheme does not offer theservices that they need or are simply not sick; hence, they do notsee the need to register. These reasons show that respondents donot understand the importance of the NHIS; therefore, continu-ous community education would be useful to sensitize people onthe role that health insurance plays in stabilizing their situation.

In relation to revenue, the main sources are support from theNHIA in the form of subsidy and reinsurance, contributions fromthe informal sector, and internally generated funds such asmembership cards’ processing fees and interest on investments.The total revenue of the scheme increased significantly over thestudy period, which is good for growth and long-term sustain-ability as evidenced in a study by Guy [4]. In spite of the annualincrease in revenue, NHIA support was the predominant contri-buting factor, which constitutes about 90% of total revenue of the

scheme. In absolute terms, the contributions collected and otherincome generated increased in the first year (2007–2008) anddeclined in the second year (2008–2009), which might be due to asimilar trend observed in the membership registration, particu-larly informal sector registration. The decline in contributionsfrom the informal sector, if it continues in subsequent years,would reduce the revenue base and cause further days in thepayment of provider claims. In the long term, it could threatenthe financial viability of the scheme if support from the NHIAgoes down. These trends in revenue were also found in the studyby Witter and Garshong [3].

The main expenditure areas of the scheme are administrativeexpenses and medical claims expenses. The annual expenditurewent up in the first year of the study period and dropped slightlyin the second year. The total expenditure, however, exceededrevenue, with medical claims expenses as the major contributingfactor. This trend in the scheme’s expenditure was also seen inthe estimation of the expense and medical claims ratios. Thecombined ratio increased consistently over the study period, withmedical claims ratio as the driving force. The increasing trend inmedical expenses could pose a serious financial challenge to thescheme. In the long term, it may result in diminished socialprotection and value to the insured members. The change inclaims payment mechanism from fee-for-service to DiagnosticRelated Groupings and an upward review of the medicine pricelist in 2008 might partly account for the high medical claimsexpenditure observed in 2008 [3]. According to the 2010 inde-pendent health sector review, claims for medicines totaled 60% ofall claims in 2009 [7]. These factors contributed to a growth indistressed schemes and failure to pay outstanding providerclaims in 2008 [3]. As emphasized earlier, it would be importantfor the scheme to attract more members and to retain them overlong periods during which they consume no or few services [16].The increasing rate of medical bills expenditure as explained bythe management is due to 1) comprehensive benefit package withno co-payment mechanism to control cost; 2) ineffective gate-keeper system, which contributes to increase in health careutilization and cost at secondary and tertiary facilities; and 3)client and health care provider abuse.

The claims settlements pattern of the scheme shows lengthydelays in the payment of provider claims. According to the NHISLegislative Instrument (LI 1809), the stipulated period for vettingand payment of claims is 4 weeks after receipt of the claims from ahealth care provider [5]. About 99% of sampled claims submittedwithin October-December 2009, however, were settled beyond themandatory period. Because the main source of income for health

84.9 82.9 93.9

0%10%20%30%40%50%60%70%80%90%

100%

2007 2008 2009

Per

cent

age

Year

Contributions collected NHIA support Other income

Fig. 3 – Revenue distribution of the scheme (2007–2009). NHIA, National Health Insurance Authority.

Table 3 – Expenditure status of the scheme (2007–2009).

Year Administrativeexpense (%)

Medicalclaimsexpense (%)

Totalexpenditure(GH¢)

2007 18.2 81.8 2,232,163.342008 8.9 91.1 5,757,550.212009 3.6 96.4 5,732,104.33


care providers is the internally generated fund that comes fromclaims reimbursed by the NHIA, this situation would affect theirrevenue base and may result in poor delivery of health service toinsured members [7]. Delays in claims settlements remains aproblem in most of the DMHISs, and this has raised acuteproblems of bankruptcy of certain providers and lack of trust bysuppliers [7]. It has been found that paying claims promptly is animportant aspect of service and good value [17]. It is also anessential element of the financial incentives of health careproviders and as a result, a key factor affecting provider behavior[18]. Early settlement of provider claims will enable health careproviders to render continuous service to insured members. Moreoften, health care providers require funds immediately after theprovision of service; hence, significant delays in claims settle-ments may force them to take other measures that defeat thepurpose of the scheme [17]. For instance, they may indulge inunhealthy practices including extortion of money from insuredmembers and refusing health care services to insured members.These provider-abuse practices may affect the renewal ratebecause dissatisfied insured members see no value in the schemeand as such are less likely to renew their memberships [13].

Since its introduction in 2004, the NHIS has partly accountedfor reduction in availability of Government of Ghana resources foroperations and public health activities of health care providersbecause service claims reimbursed by the NHIS cover some of thefacility-based operational costs [7]. The scheme has also been acontributory factor to delays in the release of funds from thecentral government, which affects the implementation of plannedactivities of health care providers. According to the Ministry ofHealth, by December 2009, the ministry received only 20% ofexpected transfers from the National Health Insurance Fund,mainly because of delays in inflows in the National HealthInsurance Fund [7]. As a result, health care providers largelydepend on claims reimbursement funds to cope with the signifi-cant increase in health cost and workload. Further delays in claimspayment by the DMHISs will therefore cause cash problems forhealth care providers. The factors contributing to delays in claimssettlements as explained by the management of the scheme weredelays in the transfer of subsidies and reinsurance from the NHIA,large volumes of medical claims, inadequate number of personnel,and ineffective claims processing software. The review of themedical claims showed that the scheme receives an averagevolume of 30,000 claims per month, which makes it practicallychallenging to vet and settle payments within the mandatory 28days. These factors were also found to be contributing to delaysassociated with provider claims settlements in 2009 [7].

Conclusions

There are increasing trends in membership coverage andrevenue that are largely driven by the exempt groups andsubsidies from the NHIA, respectively. The medical claimsexpenditure is increasing with significant delays in settlement.Introduction of cost-containment measures including co-payments and capitation payment mechanism would be nec-essary to reduce the escalating cost of medical claims.

Adherence to the stipulated 4-week period for claims paymentwould also be important to ensure that health care providersare financially resourced to deliver continuous health servicesto insured members. Furthermore, the scheme should beadequately resourced to ensure speedy vetting of medicalclaims and also to facilitate community education in thedistrict to improve membership coverage and revenue fromthe informal sector.

Acknowledgments

We acknowledge contributions of the management and staff ofGa DMHIS, field workers, and field supervisors. We thank the GaEast District Director of Health Services for permitting us toconduct the household survey in the Madina Township and thestaff of Madina Polyclinic for their support in the recruitment andtraining of field workers.

Source of financial support: This study was self-financed bythe authors. The views expressed in this article are those of theauthors.

R E F E R E N C E S

[1] World Health Organization. Sustainable Health Financing, UniversalCoverage and Social Health Insurance [A58/33]. Geneva, Switzerland:World Health Organization, 2005.

[2] Atim C, Grey S, Apoya P, et al. A Survey of Health Financing Schemes inGhana. Bethesda, MD: Partners for Health Reformplus Project, AbtAssociates, Inc., 2001.

[3] Witter S, Garshong B. Something old or something new? Social healthinsurance in Ghana (Article). BMC Int Health Hum Rights 2009;9:20.

[4] Republic of Ghana. National Health Insurance Scheme Act, Act 650.Accra, Ghana: Ghana Publishing Corporation, 2003.

[5] Republic of Ghana. National Health Insurance Scheme, LegislativeInstrument, LI 1809. Accra, Ghana: Ghana Publishing Corporation, 2004.

[6] Ministry of Health. National Health Insurance Policy Framework forGhana, Revised Version. Accra, Ghana: Ministry of Health, 2004.

[7] Ministry of Health. Independent Review: Health Sector Programme ofWork 2009, Ghana. Accra, Ghana: Ministry of Health, April 2010:8–19.

[8] Ministry of Health. Independent Review: Health Sector Programme ofWork 2007, Ghana. Report. Accra, Ghana: Ministry of Health, April2008:54–8.

[9] National Health Insurance Authority (NHIA) and Development Partners(DP). Ghana National Insurance Scheme: joint development mission.Aide Memoire 2007. Available at: http://www.snvworld.org/sites/www.snvworld.org/files/publications/5.ghana_national_health_insurance_scheme_joint_dp_mission_aide_memoire.pdf. [Accessed August 30,2013].

[10] National Health Insurance Scheme. Annual Report 2010. Accra, Ghana:National Health Insurance Authority, 2011.

[11] Wealth Health Organization. World Health Report 2000. HealthSystems: Improving Performance. Health Financing Policy Issue Paper.Geneva, Switzerland: Wealth Health Organization, 2000.

[12] International Labour Organization. Performance Indicators of StatutorySocial Insurance Schemes: Global Extension of Social Security. WorkingPaper 2009.

[13] Wipf J, Garand D. Performance Indicators for Microinsurance: AHandbook for Microinsurance Practitioners. Luxembourg: ADA asbl,2008.

[14] Guy C. Social health insurance in developing countries: a continuingchallenge. Int Soc Secur Rev 2002;55:2.

[15] Yevutsey SK, Akins M. Financial viability of District Mutual HealthInsurance Schemes of Lawra and Sissala East Districts, Upper WestRegion, Ghana. Ghana Med J 2010;44(4):130–7.

[16] Dror D. Health insurance for the poor through community schemes—isit viable? Defining an agenda for poverty reduction. Proc First AsiaPacific Forum Poverty 2007;2:193–208.

[17] CGAP Working Group on Microinsurance. Performance Indicators forMicroinsurance Practitioners. Luxembourg: ADA asbl, 2006; Workshopreport. October 16–17, 2006.

[18] Kwon S. Thirty years of national health insurance in South Korea:lessons for achieving universal health care coverage. Health PolicyPlanning 2009;24:63–71.

Table 4 – Expense, claims, and combined ratios(2007–2009).

Year Expenseratio

Claimsratio

Combinedratio

2007 2.7 12.2 14.92008 1.4 14.0 15.42009 0.7 18.1 18.8























http://www.snvworld.org/sites/www.snvworld.org/files/publications/5.ghana_national_health_insurance_scheme_joint_dp_mission_aide_memoire.pdf






























The Process of Privatization of Health Care Provision in PolandKrzysztof Kaczmarek, PhD1, Hannah Flynn, MPH2, Edyta Letka-Paralusz, MA3, Krzysztof Krajewski-Siuda, MD, PhD4,5,Christian A. Gericke, MD, PhD6,7,8,�

1Department of Health Policy, Medical University of Silesia, Katowice, Poland; 2PenCLAHRC, National Institute for Health Research, Plymouth University Schoolsof Medicine and Dentistry, Plymouth, UK; 3Department of Public Health, Medical University of Silesia, Katowice, Poland; 4Sobieski Institute, Warsaw, Poland;5University of Information Technology and Management, Rzeszow, Poland; 6The Wesley Research Institute, Brisbane, Australia; 7University of Queensland Schoolof Population Health, Brisbane, Australia; 8Queensland University of Technology School of Public Health, Brisbane, Australia

A B S T R A C T

Objectives: In January 1999, a new institutional structure for Poland'shealth care system was laid out, instigated by the dramatic change inboth the political and economic system. Following the dissolution ofstate socialism, private financing of health care services was encour-aged to fill an important role in meeting rising consumer demand andto encourage a more efficient use of resources through competitionand private initiative. However, from the outset of the intendedtransformations, systemic limitations to the privatization processhindered progression, resulting in varying rates of privatizationamongst the distinct health care sectors. The aim of this paper is todescribe the privatization process and to analyze its pace and differ-ences in strategic approach in all major health care sectors. Methods:Policy analysis of legislation, government directives, and publishednational and international scientific literature on Polish healthreforms between 1999 and 2012 was conducted. Results: The analysisdemonstrates a clear disparity in privatization rates in differentsectors. The pharmaceutical industry is fully privatized in 2012, and

the ambulatory and dental sectors both systematically increased theirprivate market shares to around 70% of all services provided. However,despite a steady increase in the number of private hospitals in Polandsince 1999, their overall role in the health care system is comparativelylimited. Conclusions: Unclear legal regulations have resulted in a grayarea between public and private health care, where informal paymentsimpede the intended function of the system. If left unchanged, officialhealth care in Poland is likely to become an increasingly residualservice for the worst-off population segments that are unable to affordthe legal private sector or the informal payments which guarantee ahigher quality service in the public sector.

Keywords: health care provision, health care reform, health policy,Poland, privatization.


Introduction

On January 1, 1999, a new institutional structure for Poland’shealth care system was founded, instigated by a dramatic changein both the political system and the economic system [1]. In theyears preceding such change, a state-funded and centralizedhealth care system had operated where the public sector haddominated in terms of both funding and service provision. Thecollapse of state socialism in 1989 because of increased opposi-tion and a failing economy, however, had severe consequenceson the state’s ability to provide health care coverage [2]. Thisresulted in a growing imbalance between the needs expressed bythe population and the system’s ability to meet them, exacer-bated by the ever-increasing cost of health care service provision.In an attempt to address this, Poland transformed the health caresystem and encouraged competition and private initiative [3,4].From the outset of the intended transformations, however,systemic limitations to the privatization process have hinderedprogression. This has resulted in varying rates of privatizationamong the distinct health care sectors and an ambiguousrelationship between public and private health care provision.

Initial Drivers for Health Care Reform

During state socialism, Poland, like many other Soviet blocnations, adopted the Semashko model for health care [5]. State-funded through taxation and heavily centralized, this particularsystem was designed with the intention of guaranteeing egali-tarian health care coverage for the entire population. After thedissolution of the Soviet Union, however, Poland along withmany other Central European countries suffered severe economicdifficulties that significantly affected health care provision [6].Because of cuts in government expenditure and a shortage ofproviders, public health care facilities became overcrowded andhad long waiting lists, scarce medical supplies, and out-of-datetechnologies [7]. Receptive to this, Poland began to allow limitedprivate providers to manage demand for public health services[8]. The principal idea envisaged was to establish a new set ofinstitutions and market-type mechanisms that would ensure amore efficient use of productive assets by creating strongerincentives arising from ownership, thereby increasing productiv-ity and efficiency [3,9,10]. This signaled an initial step towardprivatization, defined as follows: “the transfer of ownership and






� Address correspondence to: Christian Gericke, The Wesley Research Institute, PO Box 499, Toowong QLD 4066, Australia.


dx.doi.org/10.1016/j.vhri.2013.06.001

dx.doi.org/10.1016/j.vhri.2013.06.001

dx.doi.org/10.1016/j.vhri.2013.06.001


dx.doi.org/10.1016/j.vhri.2013.06.001

control of government or state assets, firms and operations toprivate investors” [11].

After this, the public sector gradually began to devolve furtheruntil budgeting of health care services was replaced with aninsurance-based system of financing. Undoubtedly, this radicallychanged the population’s right to health services, as access wasinstantly linked to registration with a mandatory health insur-ance and payment of contributions [5]. As insurance funds wereinitially regional and given autonomy, conditions were set forprivate sector service provision, which then intensified as officialout-of-pocket payments for health services were started [12].Alongside hospitals, clinics, and health centers, foundations orvoluntary associations were established, which accepted pay-ments for performing better quality or difficult-to-access services.This divided health care provision into both public and private,with a gray sphere of informal payments emerging between thetwo [2] that continued a long-standing history of informal pay-ments in the socialist health system.

In the 2007 Stefan Batory Foundation’s Corruption Barometer,78 (9%) of 870 respondents declared that they had made informalpayments in the last year, 52% of which were for informalpayments in health care [13].

In the larger Social Diagnosis panel of 3000 Polish households,1.8% of households declared informal payments in 2007, 1.3% in2009, and 1.7% in 2011 [14]. In 2011, the average informal paymentfor health services was estimated at 1244 Polish Zloty (300 euros)per year and household. Furthermore, 18.1% of householdsdeclared that they refrained from purchasing necessary medi-cines, 17.3% could not afford dental treatment, and 13.9% couldnot afford medical treatment [14].

Legal Basis for Privatization

Between 1989 and 2001, approximately 20 new laws relating tohealth care provision were adopted in Poland, which facilitatedthe development of the private sector. In particular, the law ofJuly 13, 1990, which related to the privatization of state enter-prises [15], and after its abolition the law of August 30, 1996,which related to the commercialization and privatization of stateenterprises [16], were exceptionally influential in instigating theprivatization process. Although these acts did not directly refer tohealth care services, they drew a general framework for theprocess of privatization in Poland after the fall of communism.

The most important and far-reaching legislative acts to affecthealth care were those that shaped the contracting environment.The Health Care Organisation Act passed in 1991 introducedcontracting in place of administrative relationships, allowingprivate surgeries and organizations to sign contracts for theprovision of services to people entitled to care financed frompublic resources [17]. In doing so, categories of entities authorizedto provide health services (including those that are established bynonpublic entities or individuals) were defined, as well as thetechnical requirements that such entities must fulfill.

This was followed by perhaps the most influential act—TheGeneral Health Insurance Act 1999, which introduced a socialhealth insurance system in Poland of 16 regional sickness fundsand 1 sickness fund for employees of military services [18]. Thiscaused a vast increase in the number of private organizationsholding public contracts because the regional sickness fundswere allowed to contract services with private health careinstitutions as long as they met the required conditions andoffered cheaper service costs [8]. This was the first time privateproviders were able to act within the public system of financinghealth services.

In addition to these, a package of laws regulating the com-petences of local self-government units have since been passed,which have gradually transferred the ownership duties of health

care facilities from the central administration units to the localself-governments, enabling them at the same time (under someconditions) to transform those facilities into private entities.These laws include

● the law of March 8, 1990, on local self-government [19];● the law of November 24, 1995, on a change in the range of

responsibilities of some cities on the municipal areas of publicservices [20];

● the law of June 5, 1998, on regional self-government [21]; and● the law of June 5, 1998, on district self-government [22].

None of these legal acts, however, has directly and system-atically regulated the issue of privatization of health care facili-ties. This has resulted in a process that is complicated, legallyunclear, and vulnerable to abuses, particularly in the case ofhospitals that are the most controversial in terms of theirprivatization. During the last decade, successive governmentshave tried on three occasions to establish such a law but none ofthese efforts has been successful, each time being blocked duringthe legislative process, or even earlier, at the stage of preparation.In its first attempt, the Ministry of Health tried to implementobligatory transformation of all health care organizations intocommercial law companies, entitled “Law on Commercialisationand Privatisation of Independent Public Health Care Facilities(2001).” Nevertheless, because of unfavorable political conditions(forthcoming elections, a breakdown of the governing coalition,and a strong political disintegration), the project was withdrawnand replaced with a less radical approach.

Progress of Privatization in Poland

An analysis of the scale of privatization in the Polish health caresystem shows significant disparity between the different healthcare sectors. Changes in the pharmaceutical sector and inambulatory, dental, and hospital care differ in terms of pace,strategic approach, and public resistance. To understandthese fundamental differences, each sector will be discussedseparately.

Pharmaceutical sector.The commercialization of health services began with the priva-tization of the pharmaceutical industry. This was based on theFreedom of Economic Activity Act (1988), which came intofruition at the very beginning of the postcommunist transforma-tion period [23]. Around the same time, the number of privatepharmacies accounted for approximately 43.9% of the totalnumber. Following a program implemented in 1994 devoted tothe privatization of pharmacies, however, almost all pharma-ceutical outlets belonging to the Treasury have subsequentlybeen privatized [24]. This dynamic transformation in pharmacyownership between 1990 and 2006 is illustrated in Fig. 1.

Since the introduction of co-payments for dental care,patients have started to purchase services offered by privatepractices and clinics more willingly, even when required to coverthe total cost of the treatment. In doing so, they are able toreceive a faster and perceived better quality treatment. Becauseof this high acceptability, the private dental sector developedquickly in the early 1990s. After the Law on Social HealthInsurance came into force in 1999, private dental practicesstarted to offer treatment contracted within the Social HealthInsurance system. As a result, the number of facilities offeringservices that are available only for out-of-pocket payments hasstarted to decrease gradually since 1999 [25]. Currently, morethan 80% of the active dentists work in the private sector andapproximately 85% of the services are provided by nonpublicproviders [26].


Ambulatory care.Until 1990, hospitals were responsible for the delivery of special-ist outpatient services as well as for laboratory and imagingdiagnostics. The separation of hospitals and ambulatory care wasthe first step toward the transformation of service provision. Thissubsequently offered a great opportunity to invest in the ambu-latory care infrastructure when previously inpatient services thatare in general more expensive than outpatient services hadreceived financial priority.

In the mid-1990s, the process of privatization was directedtoward outpatient care. In the first phase of the process, nonpublicfacilities were established mainly by individuals (doctors, nurses,and other medical practitioners) and companies, who becameresponsible and accountable for providing health care services [27].Later, the local self-government units joined the process bytransforming the facilities they owned when the package of lawsreforming the system of public administration came into force.These were, however, based on legal regulations that enabled onlylarge cities to do so in 1995 and all other local government unitsfrom 1999 onward. This coincided, however, with the introductionof the Social Health Insurance Law in 1999, which enabled non-public health care providers to enter into contracts with publicinsurers [28]. This allowed self-government units to privatize theirambulatory care facilities in two distinct ways:

1. The first involves holistic transformation of the public entitythrough its liquidation. In this case, the duty to provideservices is transferred to a private entity together with own-ership of the technical infrastructure. Formally, the unit thatopts for such a procedure must adopt a resolution to liquidatethe facility as well as to define the procedure of its transfer tothe nonpublic entity [29].

2. The second method involves separation of the service provi-sion from the structures of the facility and a transfer of theprovision function to a private entity. Contrary to the firstmethod, in this case the self-government unit does not liqui-date the facility but adopts a resolution on its restructuration.Formally, the infrastructure ownership remains public [28].

In view of the above-mentioned legal changes, ambulatory careprivatization has been accelerated since the mid-1990s. This is

illustrated in Fig. 2, which highlights the increase in the numberof private ambulatory care providers since 1990.

As stated previously, the past two decades have seen animmense increase in the number of private ambulatory carefacilities in Poland. In 1990, 4.5% of the ambulatory carefacilities were privately operated; by 2008 this had increasedto 77.8% [26]. The increase in privately owned facilities wasparticularly fast during the period between 1999 and 2002before stabilizing to an approximate increase of 3% each year.During this time, public facilities decreased by approximately47.5% [26].

Currently, the private sector is dominating outpatient care,which is illustrated by the number of services provided. In 2008, atotal of 290,553,000 services were provided; of these, 202,785,000were provided by private facilities, meaning that services pro-vided by public facilities account for only 30% of all servicesprovided [30].

Hospital care.The postwar history of private hospitals in Poland is relativelylinear. Those that resumed activity after 1945 were closed a fewyears later because the government continued to transformboth the economic and the health system into the Semashkomodel. The networks of private providers, however, started toreestablish themselves after the fall of communism in the1990s. The first private hospitals were established between1993 and 1994 and began their operations as single departmentsestablished mostly alongside outpatient health centers, andonly later developed further more advanced services. Theprocess of the development of the private sector in inpatientcare is much slower than in the other sectors previouslydescribed. The main reason behind this largely relates to theextent of the higher financial risk and investment required, andthe widespread existence of political beliefs opposing hospitalprivatization.

Perhaps most significantly, there is often a public and politicalunwillingness toward hospital privatization that is of particularconcern for postcommunist countries. During the communistperiod, health care was generally considered to be a “social service”that should not be determined by economic measures of efficiency

Fig. 1 – Dynamics changes in pharmacy ownership between 1990 and 2006.


[31]. Consequently, strong public opposition remains one of the keybarriers to the privatization of health care, driven by beliefs that

● hospital services are of a specific nature and they should notbe subject to a profit motive;

● private hospitals are acting mainly for the profit of theirowners, which is in conflict with the mission of health carefacilities;

● privatization of hospitals equals charging patients for theservices; and

● nonpublic hospitals are unwilling to provide highly complexmedical procedures because of their unprofitability, whichmay cause a limitation of access to such services.

Furthermore, the lack of a comprehensive legal basis forprivatization constitutes an obstacle for the development of thissector. Despite these limitations, the process is still progressing,largely through the local self-governments, for which the trans-formation of hospital facilities they own is becoming increasinglypopular [32].

Since 1999, self-government units have transformed 130hospital units into commercial companies; this includes 77hospitals and 53 single hospital departments [33]. In many cases,the situation is becoming paradoxical because self-governmentunits still remain formal owners of the privatized hospitals,

which makes them so-called nonpublic self-government-ownedhospitals. The difference between such an organization and atraditional one is that these hospitals are acting as companiesunder commercial law and not as the “independent public healthcare facilities” as it was before they were restructured. Table 1summarizes the dynamics of the development of the nonpublichospital sector in Poland over the past two decades.

Of the 120 nonpublic hospitals currently operating in Poland,77 are those that have been restructured by local self-governmentunits. This constitutes 64.2% of the number of private hospitalsand 10.4% of the total number of private hospitals in Poland [26].In recent years, the general share of private hospitals equates toapproximately 20% of the total, and when taking the number ofbeds into consideration, it is less than 6%. In 2007, nonpublichospitals signed contracts with the social health insurance fundsfor an amount of more than 600 million Polish Zloty (145 millioneuros), which equates to approximately 3.2% of the total publicresources spent on hospital care [26].

Cost-Effectiveness of Privatization

Because of issues regarding the availability of data, a comparisonof the cost-effectiveness of public and private health care poses asignificant problem. Private entities generally do not releaserelevant information, and available sources are generally limited

Table 1 – Structure of ownership of hospital care facilities in Poland (1999–2007).

1999 2000 2001 2002 2003 2004 2005 2006 2007

No. of privatehospitals

21 30 45 61 72 147 170 153 170

No. of beds 446 1,574 2,476 4,221 5,171 7,649 8,215 9,318 10,204Total hospitals 715 716 736 739 732 790 781 742 748Total beds 198,688 190,952 188,234 188,038 186,043 183,280 179,493 176,673 175,023% of private

hospitals2.9 4.2 6.1 8.3 9.1 18.6 21.8 20.6 22.7

% of privatebeds

0.2 0.8 1.3 2.2 2.8 4.2 4.6 5.3 5.8

Fig. 2 – Structure of the ownership of ambulatory care facilities in Poland (1990–2008).


to the presentation of a few selected individuals. Existing data,however, suggest an average growth of within 20% to 30% inprivate sector revenues annually. The overall size of the sectorwas estimated at about 1915 million Polish Zloty (462 millioneuros). This information confirms an analysis by the NationalAssociation of Private Employer Health Care (2009), who con-ducted a study involving nine private companies providingmedical services. The analysis indicated that in 2007 and 2008,the annual growth of the sector revenue amounted to 31.1% and33.1%, respectively, and the value of income in 2008 was esti-mated to be 930.4 million Polish Zloty (225 million euros). Onaverage, survey participants within the health profession indi-cated that their income in 2006 increased by approximately 32.9%in 2007 and 35.3% in 2008. In comparison with the public sector,employers’ contributions to medical care for their employeeswere the dominant source of revenue for private providers. Theseaccounted for nearly 44% of the revenue [34].

Quality of Care—Private versus Public

The private and public sectors in Poland also differ in terms of theperceived quality of care reflected in patients’ opinions. Severalopinion polls dedicated to this issue share the same conclusion—in general, private services are considered to be of a higherquality, with the most significant differences stated to be thephysician’s manner toward patients. In a survey conducted bythe marketing research and opinion poll company Partner inBusiness Strategies in 2007 [35], respondents compared publicand private services with respect to atmosphere, doctor’s com-mitment, and respect to patient privacy. For these three criteria,private providers were considered significantly better than thepublic sector. With respect to atmosphere, 54% of the respond-ents declared that private providers are better, with only 10%declaring the opposite. Results for doctors’ commitment andrespect for patient’s privacy were also in favor of the privatesector (51% and 47% for private providers and 9% and 8% forpublic providers, respectively). The general quality of serviceswas also recognized as being better in the private sector, and ofall respondents, 45% declared the private sector as being superior,with only 10% having the opposite opinion. One of the few areasin which the study showed no significant advantage to theprivate sector was with respect to the experience of the doctors.Although 19% of the respondents thought that more experiencedand better qualified doctors worked in private health care, 17% ofthe respondents considered the quality of doctors better in thepublic sector. It is worth noting at this point that almost two-thirds of the respondents had no opinion on this matter, whichmay suggest that differences in knowledge and experience aredifficult to judge for patients.

Conclusions

A free health care market does not exist in Poland, nor is itplanned to be one. Since the dissolution of state socialism,private financing of health care services has increased substan-tially and has filled an important role in meeting the increasingconsumer demand in some areas of health care and, to someextent, encouraging a more efficient use of scarce resourcesthrough competition and private initiative for public healthservices. This step toward privatization is mostly evident withinthe pharmaceutical industry, which is now fully privatized. Inaddition, the ambulatory sector has systematically increased itsprivate market share to approximately 70% of all services pro-vided and is continually moving toward full privatization. Despitethe regular increase in the number of private hospitals inPoland, however, their role in the health care system remains

comparatively limited. This is largely due to complicated andunclear legal regulations, resulting in a gray area between publicand private health care where informal payments impede theintended function of the system. If this is left unchanged, theofficial public health care sector in Poland is likely to become anincreasingly residual service for the worst-off patients, who areunable to afford the legal private sector or the informal paymentsthat guarantee a higher quality service in the public sector.

Of prime importance, however, as with many postcommunistcountries, one of the key barriers to health care privatization isthe traditional belief that health care should be a “public service,”not determined by economic measures of efficiency. With theheritage of the communist period hard to overcome, strongpublic opposition will remain one of the key barriers to theprivatization of health care.

Source of financial support: This study was supported by theNational Institute for Health Research Collaboration for AppliedHealth Research and Care for the South West Peninsula (to H.F.).

R E F E R E N C E S

[1] McMenamin I, Timonen V. Poland’s health reform: politics, marketsand informal payments. J Soc Policy 2002;31:103–18.

[2] Mishtal J. Neoliberal reforms and privatisation of reproductive healthservices in post-socialist Poland. Reprod Health Matters 2010;18:56–66.

[3] Ostrowski T, Wdowiak L. Prywatyzacja opieki zdrowotnej w Polsce[Characteristics of the privatization of health care in the province].Zdrowie Publiczne 2002;112:86–92.

[4] Jaworska-Łuczak B. Prywatyzacja w ochronie zdrowia w Polsce-analizastanu obecnego i prognoza na przyszłosc [The privatization of healthcare in Poland—analysis of the current state and outlook for thefuture]. Zdrowie Publiczne 2002;112:93–103.

[5] Kuszewski K, Gericke C. Health Systems in Transition: Poland (Vol 7).Copenhagen, Denmark: World Health Organization Regional Office forEurope on behalf of the European Observatory on Health Systems andPolicies, 2005:1–106.

[6] Maarse H. The privatization of health care in Europe: an eight-countryanalysis. J Health Polit Policy Law 2006;31:981–1014.

[7] McLaughlin D, Smith D. Doctors go west in Polish brain drain. TheObserver May 15, 2005.

[8] Brandt T, Schulten T. Liberalisation, Privatisation and Regulation in thePolish Healthcare Sector/Hospitals: Privatisation of Public Services andthe Impact on Quality, Employment and Productivity (PIQUE).Düsseldorf, Germany: Wirtschafts- und SozialwissenschaftlichesInstitut, 2006.

[9] Kruk W, Wdowiak L. Prywatyzacja w opiece zdrowotnej [Analysis of thecurrent state and outlook for the future]. Medycyna Ogólna2006;12:171–85.

[10] Tymowska K. Prywatyzacja opieki zdrowotnej w teorii i praktyce[Prywatyzacja opieki zdrowotnej w teorii i praktyce]. Prawo Medycyna2000;5:138–46.

[11] Organisation for Economic and Co-operation Development (OECD).Glossary of Industrial Organisation Economics and Competition Law. Paris,France: Organisation for Economic and Co-operation Development, 1993.

[12] Chawla M, Berman P, Kawiorska D. Financing health services inPoland: new evidence on private expenditures. Health Econ 1998;7:337–46.

[13] Kopiak A. Barometr korupcji 2007. Raport z badań [2007 CorruptionBarometer. Research Report]. Warsaw: Stefan Batory Foundation, 2007.Available from: http://www.batory.org.pl/doc/barometr-korupcji-2007.pdf [Accessed May 8, 2013].

[14] Białowolski P, Czapiński J, Grabowska I, et al. Warunki życiagospodarstw domowych. Diagnoza Społeczna 2011. Warunki i JakoścŻycia Polaków [Social diagnosis 2011. Objective and subjective qualityof life in Poland]. Contemp Econ 2011;5:45–159.

[15] Polish Parliament. Ustawa z dnia 13 lipca 1990 r. o prywatyzacjiprzedsiębiorstw państwowych. Dz. U. 51, item 298 [The Act of 13 July1990 on the privatization of state enterprises. No. 51, item 298]. SEJM ofthe Republic of Poland, 1990.

[16] Polish Parliament. Ustawa z dnia 30 sierpnia 1996 r. o komercjalizacji iprywatyzacji przedsiębiorstw państwowych. Dz.U. 118, item 561 [TheAct of 30 August 1996 on commercialisation and privatization. No. 118,item 561]. SEJM of the Republic of Poland, 1996.

[17] Polish Parliament. Ustawa z dnia 30 sierpnia 1991 r. o zakładach opiekizdrowotnej. Dz. U. 91, item 408 [The Act of 30 August 1991 on healthcare. No. 91, item 408]. SEJM of the Republic of Poland, 1991.







































http://www.batory.org.pl/doc/barometr-korupcji-2007.pdf

http://www.batory.org.pl/doc/barometr-korupcji-2007.pdf
















[18] Watson P. Fighting for life: health care and democracy in capitalist. CritSoc Policy 2011;31:53–76.

[19] Polish Parliament. Ustawa z dnia 8 marca 1990 r. o samorządzieterytorialnym. Dz. U. 16, item 95 [The Act of 8 March 1990 on thelocal government. No. 16, item 95]. SEJM of the Republic of Poland,1990.

[20] Polish Parliament. Ustawa z dnia 24 listopada 1995 o zmianie zakresudziałania niektórych miast oraz o miejskich strefach usługpublicznych. Dz. U. 141, item 692 [The Act of 24 November 1995 tochange the scope of some of the cities and urban areas of publicservices. No. 141, item 692]. SEJM of the Republic of Poland, 1995.

[21] Polish Parliament. Ustawa z dnia 5 czerwca 1998 r. o samorządziewojewództwa. Dz. U. 91, item 576 [The Act of 5 June 1998 on theprovincial self-government. No. 91, item 576]. SEJM of the Republic ofPoland, 1998.

[22] Polish Parliament. Ustawa z dnia 5 czerwca 1998 r. o samorządziepowiatowym. Dz. U. 91, item 578 [The Act of 5 June 1998 on the countygovernment. No. 91, item 578]. SEJM of the Republic of Poland, 1998.

[23] Polish Parliament. Ustawa z dnia 23 grudnia 1988 o działalnościgospodarczej. Dz. U. 41, item 324 [The Act of 23 December 1988 oneconomic activity. No. 41, item 324]. SEJM of the Republic of Poland, 1988.

[24] Trzecia W, ed. Zespół do przygotowania raportu: Finansowanieochrony zdrowia w Polsce—Zielona Księga [Financing Health Care inPoland—Green Paper]. 2005.

[25] Izabella-Michalewicz M. Sektor prywatny w systemie ochrony zdrowiaw Polsce [Private Sector in the Health Care System in Poland]. Warsaw,Poland: Wydział Analiz Ekonomicznych i Społecznych, KancelariiSejmu, 2004.

[26] Central Statistical Office. Mały rocznik statystyczny [Polish StatisticalYearbook]. Warsaw, Poland: Central Statistical Office, 2009.

[27] Kachniarz M. Komercjalizacja samodzielnego publicznego zakładuopieki zdrowotnej. Kluczowe warunki osiągnięcia sukcesu[Commercialization of Independent Public Health Care. Key Conditionsfor Success]. (1st ed.). Krakow, Poland: ABC - Wolters Kluwer Polska,2008.

[28] Król ZJ. Prywatyzacja ambulatoryjnej opieki zdrowotnej [Privatizationof ambulatory health care]. Zdrowie Zarządzanie II 2000:29–32.

[29] Kowalska K. Managed care and a process of integration in health caresector: a case study from Poland. Health Policy 2007;84:308–20.

[30] Golinowska S, Kozierkiewicz A. Quality in and Equality of Access toHealthcare Services: country report for Poland. Brussels, Belgium:European Commission, 2008.

[31] Profaska J. Prywatyzacja samodzielnego zakładu opieki zdrowotnejjako decydujący element przekształceń własnościowych w ochroniezdrowia [Privatization of health care themselves as a decisive elementof privatization in health care]. Przewodnik Menedżera Zdrowia2001;1:15–9.

[32] Wojciechowski A. Prywatyzacja szpitali. [The privatisation ofhospitals]. Przegląd Urologiczny 2001;7:83–94.

[33] Ministry of Health. Informacja o przekształceniach własnościowych wsektorze ochrony zdrowia, przeprowadzonych decyzją jednosteksamorządu terytorialnego w latach 1999–2009 (I półrocze). [Informationabout Ownership Transformations in the Health Sector, Carried Out theDecision of Local Government Units in the Years 1999–2009]. Warsaw,Poland: Ministerstwo Zdrowia RP, 2009.

[34] Mrozowicki A. Poland: Industrial Relations in the Health Care Sector.Wroclaw, Poland: Institute of Public Affairs, University of Wroclaw,2011.

[35] Partner in Business Strategies. Służba zdrowia — publiczna a prywatna[Healthcare—Public and Private]. Warsaw, Poland: Partner in BusinessStrategies, 2007.

































































Transforming Public Servants’ Health Care Organization in Greecethrough the Implementation of an Electronic Referral ProjectKyriakos Souliotis, PhD1,�, Vasiliki Mantzana, PhD2, Manto Papageorgiou, MSc3

1Faculty of Social Sciences, University of Peloponnese, Corinth, Greece; 2Department of Digital Systems, University of Piraeus, Piraeus, Greece; 3National School ofPublic Health, Athens, Greece

A B S T R A C T

Objective: The Greek Public Servants’ Health Care Organization aim-ing to organize, monitor, and enhance the health care servicesprovided to 1,500,000 public servants decided to respond to thenational alert of the economic crisis through the reduction of costscaused by diagnostic tests (€300,000,000 claims for 2008), to improveworking conditions of contracted physicians and laboratories, and toenhance services provided to insured members. In September 2010,the Greek Public Servants’ Health Care Organization initiated a pilotproject that electronically records the prescription process of thediagnostic tests, which is Web-based, is open source, and wasprovided for free to the contracted physicians and diagnostic centers.Methods: In this article, we present some interesting findings result-ing from the implementation of the pilot electronic referral project byexamining a 9-month period. Results: Fifty-eight percent of thephysicians had the necessary equipment for the operation of thesystem, more than 3600 physicians used it, 17,495 public servants had

been served through the system, and 178,456 paraclinical examina-tions had been prescribed with a cost of €1,394,980. In addition, theanalysis revealed that the implementation of an electronic referralsystem could provide significant benefits, such as a faster referralprocess, valid and coherent information, minimization of the risk ofmisinterpreting the electronic referral due to illegibility of handwrit-ing, and improvement in quality of services. Conclusions: The Greekelectronic referral system was one of the first attempts towardcreating the basis of a society of transparency and cost control. Thelessons learnt from this article should not be ignored in the process ofredesigning and improving the electronic referral system for Greece.Keywords: E-health, electronic referral project, Greece, insurancefunds, physician.


Introduction

Health care systems consist of large and complex processes thataffect and get affected by multiple interacting actors, such asphysicians, nurses, patients, citizens, medical suppliers, andhealth insurance providers, with different backgrounds, knowl-edge, organizational beliefs, interests, and culture. Health caresystems all over the world, however, face challenges, such as1) increasing demand for health and social services due to anaging population; 2) higher expectations by citizens who requestprovision of high-quality health care services and reduction ofinequalities; 3) management of large amounts of health data thatshould be available on request safely at any place and should beprocessed for administrative purposes; 4) need to provide thebest possible health care services in conditions of budgetaryconstraints; and 5) reduction of medical errors that lead to lossof life or causing irreparable injury.

It has been reported that the implementation and use ofe-health systems, such as e-prescribing, electronic referralsystems, and electronic health care records, could support gov-ernments and organizations to confront the aforementionedchallenges. Although hundreds of billions of dollars and

countless number of hours have been invested in the develop-ment of electronic health, more needs to be done.

In Greece, the adoption of e-health systems is happeningat an exceptionally slow pace. According to records of theHealth Ministry, the actions included in the second CommunityStructural Fund relative to the sector of information technologywere of small scale and limited to the planning level. For thisreason, in the third Community Structural Fund, the main focuswas on the integration of information systems in health careorganizations. Handwritten prescriptions and referral letters hadbeen the main method for physicians to communicate decisionson diagnostic tests and for diagnostic centers to conduct the testswhile being a source of information for the payer at thesame time.

In an effort to put an end to directed prescribing, to elevatethe medical expense in great heights, and to crack down thephenomenon of overprescribing, Greek social insurance fundsdecided to implement and adopt an electronic referral system.Currently, the electronic referral project in Greece is at thetransitional stage between paper and Web. The transition fromthe traditional process to the new electronic era offers uniqueopportunities and challenges.






�Address correspondence to: Kyriakos Souliotis, Faculty of Social Sciences, University of Peloponnese, Damaskinou & Kolokotroni Street,Corinth 20100, Greece.


dx.doi.org/10.1016/j.vhri.2013.06.003

dx.doi.org/10.1016/j.vhri.2013.06.003

dx.doi.org/10.1016/j.vhri.2013.06.003


dx.doi.org/10.1016/j.vhri.2013.06.003

The introduction of electronic referral systems has beenproved to have a positive effect on prescribing and dispensingprocesses, in the sense that an electronic referral system canimprove safety, quality, efficiency, and cost-effectiveness [1].

The purpose of this article was to present some interestingfindings and outcomes that the pilot electronic referral projectinduced to Greek Public Servants’ Health Care Organization(GPSHCO) by examining a 9-month period (September 1, 2010,to June 1, 2011). In doing so, the authors present a review of theliterature on e-health systems as well as the methodology usedto conduct this research. Afterward, they analyze GPSHCO’s effortto enhance health care services provided to 1.5 million insuredpublic servants and reduce costs that are provoked by diagnostictests prescribed to GPSHCO members (€300,000,000 claims for2008) by launching the pilot electronic referral project on May 1,2010. Moreover, they present the outcomes that this projectinduced to the organization and draw conclusions.

Literature Review

Information systems (ISs) play an increasingly crucial role in thehealth care sector by providing an infrastructure to integratepeople, processes, and technologies [2,3]. Health care informationsystems (HISs) are defined as “computerized systems designed tofacilitate the management and operation of all technical (bio-medical) and administrative data for the entire health caresystem, for a number of its functional units, for a single healthcare institution, or even for an institutional department or unit”[4]. The use of computers in the health care sector can be tracedto the 1960s. The first attempts to adopt HISs were made duringthe early 1970s [5]. In the past, however, HIS initiatives werelimited to the automation of business processes related toadministration and health care tools as well as techniquesrelated to various medical procedures such as diagnostic, ther-apeutic, and surgical procedures. During the 1980s, innovativepatterns in database designs and applications related to HISs ledto developments in planning and administration of the healthcare data. In parallel, HISs also introduced low-cost financialsystems for hospitals fewer than 200 beds in size [6]. It should benoted that the early computerized systems were limited to bighospitals and government projects (military).

As the information technology industry flourished, the HIStechnology was populated with various network applications.The net period with the Internet, intranet, and extranet affectedthe communication of data in hospitals, especially in the 1990s.In the middle of the same decade, the interface engine emergedas a product to support the integration of applications, as best-of-breed applications became harder to manage [7].

E-health systems such as electronic prescribing [8], electronicreferral systems [9], personal health records [10,11], asynchro-nous health care communication systems [12], and picturearchiving communication systems [8] have been applied inhealth care to improve the capabilities of physicians and clinicalstaff and provided increased services to patients, caregivers, andcitizens in general. E-health is an emerging field in the inter-section of medical informatics, public health, and business,referring to health services and information delivered orenhanced through the Internet and related technologies.

World Health Organization defined e-health as being “thecost-effective and secure use of information and communicationtechnologies in support of health and health-related fields,including health-care services, health surveillance, health liter-ature, and health education, knowledge and research.” Health-care Information and Management Systems Society’s E-HealthSIG defined e-health as “the application of Internet and otherrelated technologies in the health care industry to improve the

access, efficiency, effectiveness, and quality of clinical andbusiness processes utilized by health care organizations, practi-tioners, patients, and consumers to improve the health status ofpatients” [13]. In a broader sense, the term characterizes not onlytechnical development but also a way of thinking, an attitude,and a commitment for networked, global thinking to improvehealth care locally, regionally, and worldwide by using informa-tion and communication technology [14].

In a study conducted by Delloite/Ipsos in 2011 for the Euro-pean Commission regarding e-health systems in Europe, it wasstressed that integrated systems for electronic referral are amongthe e-health systems that need the greater attention. In the samestudy, it was reported that an integrated system able to send orreceive electronic referral letters is currently available in 34% ofthe 900 acute hospitals in the 30 countries surveyed [15].

E-referral projects among various countries differ a lot. It isnot likely that there is one solution that proves to be beneficial forall countries because the solution depends on legislation, organ-ization of the health care system, and cultural differences.Electronic referral systems seem to have a large potential foreconomic savings as a whole, but it takes a longer time thanexpected to realize this potential.

Research Methodology

The authors have developed an empirical research methodologyto study electronic referral projects. This methodology is basedon three development stages: 1) research design development,2) case study data collection, and 3) case study data analysis. Theresearch design adopted was the first independent part of theempirical research methodology. The starting point was to reviewthe literature to develop an understanding of the research areaunder investigation. From the literature review, several researchissues were highlighted for a more focused study. For the purposeof this article, a single case study strategy was used to exploreand understand the topic under research. A qualitative casestudy strategy can offer a “holistic” view of the processesinvolved as well as a realization of the topic under research[16]. To collect data, we used various techniques, such as ques-tionnaires, documentation, and observation.

Case Study

In 1983, the Greek National Health System (ESY) was establishedunder the Law 1397/1983 [17]. The Ministry of Health and Welfare

Fig. 1 – Health care provision infrastructure; data from theWorld Health Organization [18].


is responsible for the health care provision as well as for thenational health policy and strategy development. The ESY con-sists of the following three subsystems (Fig. 1), which operatealmost independently [19].

The National Health System

The ESY comprises public hospitals, health centers, and theNational Centre of Emergency Care. It provides inpatient careand emergency prehospital care on a universal basis. It aims toprovide free and comprehensive health care coverage. The ESYconstitutes 123 general and specialized public hospitals (36,621beds) and 9 psychiatric clinics (3,500 beds). Moreover, 32 of the123 hospitals provide tertiary and highly specialized care [17].Health care services are also provided by public hospitals, whichinclude 13 military hospitals financed by the Ministry of Defense,5 hospitals of the Social Security Institution, and 2 universityhospitals operating under the authority of the National andKapodistrian University of Athens. Emergency prehospital careis provided by the National Centre of Emergency Care, which is aNational Health System agency. Health centers also provideemergency services, short-term hospitalization, and follow-upof recovering patients, dental treatment, family planning serv-ices, vaccination services, and health education. In addition,health care services are provided extensively by private healthcare organizations (26%). Nowadays, 234 private hospitalsand clinics operate in Greece, with a total capacity of 15,397 beds[17]. It appears that the ESY is a mixed system of public-privatefunding and provision of health care services.

Social Insurance Funds

Health care services in Greece are offered through hospitals,private clinics, and physicians as long as one has been registeredwith a social insurance fund. Every individual (employer,employee, pensioner, dependent family member, and child)living in Greece has a unique personal Social Security Number,named AMKA, to benefit from social security and to obtain ahealth booklet. There are several social security institutions inGreece, such as the GPSHCO and the Social Security Institution.Registration with such institutions requires part-time or full-timeemployment. Social security benefits are available only tothose who are registered and contributing to the social securitysystem. Social security contributions are paid by both theemployer and the employee or by the individual, in case one isself-employed.

Private Health Sector

The private health sector has numerous diagnostic centers,private clinics, laboratories, and so forth.

This case study was conducted at the GPSHCO whose aim is toorganize, monitor, control, and enhance the health care servicesprovided to public servants in terms of quality and funding.The structure of the GPSHCO includes the Central Office and57 regional health care services for the insured public servants(YPAD). The GPSHCO has approximately 1,500,000 insuredmembers (public servants and family members). It offers healthcare services to public servants. More precisely, the GPSHCO hascontracted partners (10,000 pharmacists, 13,000 physicians, 3,000diagnostic centers and private clinics, and 700 physiotherapists)that provide health care services to the insured members.

The remuneration system for social security funds is based onGreek law. The GPSHCO receives revenues from the contributionsof the insurers that are paid on a monthly basis. In particular, theGPSHCO receives 2.55% of the monthly wage of each insuredmember paid by the employee plus an amount equal to 5.1% ofthe monthly wage of each insured member paid by the employer,that is, the state. These contributions are collected via theMinistry of Finance that subsidizes the GPSHCO to pay itsobligations. The aforementioned contracted partners are paidvia the regional offices of the GPSHCO.

The GPSHCO has a fully integrated IS, which was imple-mented in early 2003 and consists of the following subsystems:

� protocol and price management of hospital stays of 57 YPAD;� protocol and economic management of GPSHCO’s Central

Office;� insured members’ registry for the Central Office and YPAD;� warehouse for the Central Office and YPAD;� management of insured members’ hospitalization in private

clinics;� payroll of GPSHCO’s employees;� Management information system—Business intelligence

based on budget figures for the Central Office and YPAD;� Safety management system (create users’ accounts, assign

roles, etc.); and� Disaster recovery system that covers all GPSHCO’s systems.

Although the GPSHCO has a quite advanced IS, it is noteffective enough in identifying abuse of the system by theinsured members. In particular, the insured member, presentingwith his or her health booklet, can visit a physician contractedwith the GPSHCO. The latter can refer the insured member fordiagnostic tests using the health booklet, and the insuredmember can then visit a physician contracted with the GPSHCO

Have joinedthe pilotproject66.6%

Will join in theimmediate

future16.9%

Refused tojoin

15.4%

Other reasonsfor not joining

1.1%

Fig. 3 – Acceptance of the electronic referral system fordiagnostic tests.

Fig. 2 – Digital behavior of the medical community.


or diagnostic center to execute the referral and conduct the tests.The physician or the diagnostic center, at the end of each month,submits claims to the GPSHCO. The GPSHCO registers to the IS

the claims made by each physician or the diagnostic center on aninvoice basis without registering any information regarding thediagnosis or the type and cost of tests ordered. As a result, it is

Fig. 4 – Number of physicians who used the system from October 1, 2010, to July 7, 2011.

Fig. 5 – Number of electronic referrals issued from October 1, 2010, to July 7, 2011.

Fig. 6 – Cost of electronic referrals (€) from October 1, 2010, to July 7, 2011.


impossible for GPSHCO’s management to monitor and control theincreasingly rising health care expenditure. More specifically, in2003, the cost of diagnostic tests and visits was €155,494,142.8whereas in 2006, it was €266,391,245.4 and in 2009, €297,051,270.7.At this point, it should be mentioned that during this period, thenumber of insured members and their health status have notchanged.

In an effort to 1) control and reduce expenses that are provokedby diagnostic tests and visits to physicians by GPSHCO’s insuredmembers, 2) improve the working conditions of contracted physi-cians and laboratories, and 3) enhance the services provided toinsured public servants of the organization, GPSHCO’s CommitteeBoard initiated a pilot project on September 1, 2010, that electroni-cally records the prescription process of the diagnostic tests. Theplatform is a Web-based open source and was provided for free tothe contracted physicians and diagnostic centers, which are itsmain users. The contracted physicians use the platform to refer thepatient (insured member) for diagnostic tests by using a coded listof diagnostic tests and diagnoses according to International Classi-fication of Diseases, Tenth Revision, and diagnostic centers execute thediagnostic tests and charge the GPSHCO.

The project has been assisted by a call center that promotesthe implementation of the system, communicates and supportsphysicians, and provides information as well as technical support

for the whole project. Furthermore, a Web site providing infor-mation on the application and answering frequently askedquestions posed by the stakeholders concerning the manage-ment of electronic referrals has been developed.

Discussion

The electronic referral project started on a voluntary pilot basisin September 2010, and after 9 months (June 2011) its usebecame mandatory for all physicians contracted with theGPSHCO. During the first month of the pilot project implemen-tation, a survey was conducted to study the digital behavior ofthe medical community. According to the results of this survey,44.5% of the physicians mentioned that they have all thenecessary equipment for the operation of the system, 22.5%stated that they have part of the necessary equipment (usuallyjust a PC and a printer), and 29.1% said that they do not haveany of the necessary equipment (Fig. 2). Moreover, it should benoted that 66.6% of them accepted to use the electronic referralsystem (more than 3600 physicians) while 16.9% of the con-tracted physicians said they would start using it in the imme-diate future (Fig. 3).

In the following paragraphs, we present data regarding theuse of the system on a monthly basis for the pilot period understudy (September 2010 to June 2011).

People react differently toward new ideas, practices, or objectsas a result of variations in innovativeness. The diffusion ofinnovation theory operationalizes innovative individuals as earlyadopters of innovations in contrast with majority and lateadopters [20]. The electronic referral project started on a volun-tary pilot basis in September 2010, and after 9 months (June 2011)its use became mandatory for all physicians contracted with theGPSHCO. At the beginning of the project, the number of physi-cians was relatively low, which started increasing in the middleof the pilot project and decreased in the last 3 months of theproject (Fig. 4). When the GPSHCO’s Committee Board realizedthis, and as during the pilot phase the system was continuouslyimproving, they made the use of the system mandatory for all thecontracted physicians.

Although the number of contracted physicians decreased inthe last 3 months of the pilot project, the number of insuredmembers asking for an electronic referral increased constantlybecause the system provided significant advantages to them.

Fig. 7 – Mean cost of electronic referrals (€) per patient from October 1, 2010, to July 7, 2011.

Table 1 – Comparing the cost of diagnostic testsbetween 2009 and 2010.

Month Cost of diagnostic tests per year (€)

2009 2010January 8,984,396.21 7,294,334.95February 7,319,735.93 5,846,991.50March 8,195,129.24 6,421,289.59April 8,079,929.95 6,495,760.75May 8,234,257.64 6,189,177.16June 7,510,277.69 5,586,522.53July 8,824,800.53 5,754,905.08August 9,048,589.53 4,230,721.10September 5,850,758.07 4,577,497.58October 8,047,418.45 4,882,519.44November 7,562,862.22 4,169,340.00December 8,083,606.29 4,167,922.29


More specifically, the advantages of the implementation of theelectronic referral system included faster referral process, validand complete (coherent) information, and minimization of therisk of misinterpreting the electronic referral due to illegibility ofhandwriting. Moreover, such an implementation can increase thequality of services provided to insured members. It should benoted that in the 9-month pilot period, 88,154 insured publicservants had been served through the system and 41,200 elec-tronic referrals had been issued (Fig. 5).

Moreover, 178,456 diagnostic tests had been prescribed, with atotal cost of €1,979,538.59 (mean cost of €48.05) (Figs. 6 and 7). Themean cost of the electronic referrals per patient became approx-imately €47 compared with the mean cost on January 2010, whenthe project was initiated, which was approximately €58 (Table 1).

Moreover, it should be mentioned that the cost for diagnostictests decreased by approximately 32% in 2010 (January to Decem-ber) compared with 2009 (January to December), not only because ofthe use of the electronic system but also because of administrativedecisions that the committee decided to implement (Table 1).

The vast majority of physicians using the system seem to havenormal prescribing behavior. Approximately 60% of them induceless demand for diagnostic tests than the average, 16% of thephysicians induce greater demand than the average because oftheir medical (Table 2) specialization, and 40% of the physicianscause demand below the average, regardless of their specialty.

Conclusions

The Greek electronic referral system was one of the first attemptstoward creating the basis of a society of transparency and costcontrol. Similar to other projects, the launch of the electronicreferral system aims at supporting Greece in tackling the eco-nomic crisis, and thus it is important to note that the memo-randum of economic and financial policies published mentionedthat “[the] Government ensures that the e-prescribing system fordiagnostic tests currently piloted by GPSHCO is extended to allsocial security funds” [21].

Moreover, the empirical evidence indicates the lack of techno-logical education among physicians and their denial to invest innew technologies. Inevitably, physicians have to deal with manychanges in their work environment as the progress of medicalscience and technology is rapid. It is imperative for them toimprove their skills and requalify themselves at all times. Inaddition, the analysis revealed that the implementation of anelectronic referral system could provide significant benefits, suchas faster referral process, valid and complete (coherent) informa-tion, and minimization of the risk of misinterpreting the electronicreferral due to illegibility of handwriting, and can also increase thequality of services provided to insured members. Moreover, suchan implementation can decrease the mean cost of the electronicreferrals per patient, which became approximately €47 comparedwith the mean cost on January 2010 when the project was initiated,which was approximately €58. In addition, it can decrease the cost

for diagnostic tests. In this case study, the cost has decreased byapproximately 32% in 2010 (January to December) compared withthat in 2009 (January to December), not only because of the use ofthe electronic system but also because of administrative decisionsthat the committee decided to implement.

The lessons learned from the international experience shouldnot be ignored in the process of redesigning and improving theelectronic referral system for Greece. Electronic referral in Greeceis a reality against all predictions. Electronic referral can andshould be the beginning for the widespread of e-health, giventhat the electronic referral administration affects public healthand economy. No claim for generalization is made, and thelessons learned are a result of the description provided and donot seek to be prescriptive. These lessons might be helpful tothose willing to implement electronic referral projects as well asto researchers and practitioners.


R E F E R E N C E S

[1] Franklin BD, O’Grady K, Donyai P, et al. The impact of a closed-loopelectronic prescribing and administration system on prescribing errors,administration errors and staff time: a before-and-after study. Qual SafHealth Care 2007;16:279–84.

[2] Wanless D, Charlesworth A, Walker I, et al. In: H. Treasury, ed., SecuringOur Future Health: Taking a Long-Term View (Vol. 2004). London: HMTreasury, 2002.

[3] Ragupathi W. Healthcare Information Systems. Commun ACM1997;28:81–2.

[4] Rodrigues RJ, Gattini G, Aalmeida G. Setting up Healthcare ServicesInformation Systems: A Guide for Requirement Analysis, ApplicationSpecification, and Procurement. Washington, DC: Essential Drugs andTechnology Program, Division of Health Systems and ServicesDevelopment, PAHO/WHO, 1999.

[5] Hodge JG, Gostin LO, Jacobson PD. Legal issues concerning electronichealth information: privacy, quality, and liability. JAMA1999;282:1466–71.

[6] Straggers N, Thompson CB, Snyder-Halpern R. History and trends inclinical information systems in the United States. J Nurs Sch2001;33:75–81.

[7] Beaver K. Healthcare Information Systems (2nd ed.). Boca Raton, FL.Auerbach Publications, 2002.

[8] Menachemi N, Burke DE, Ayers D. Factors affecting the adoption oftelemedicine—a multiple adopter perspective. J Med Syst2004;28:617–32.

[9] Nicholson C, Jackson CL, Wright B, et al. Online referral and OPDbooking from the GP desktop. Aust Health Rev 2006;30:397–404.

[10] Lafky DB, Tulu B, Horan TA. A user-driven approach to personal healthrecords. Commun Assoc Inf Syst 2006;17:1028–9.

[11] Lafky DB, Horan TA. Prospective personal health record use amongdifferent user groups: results of a multi-wave study. Paper presented atthe Proceedings of the 41st Hawaii International Conference on SystemSciences, Waikoloa, Big Island, HI, 2008.

[12] Wilson EV. Asynchronous health care communication. Commun ACM2003;46:79–84.

[13] Griskewicz M. HIMSS SIG develops proposed e-health definition(Vol. 13). Chicago, IL: Healthcare Information and ManagementSystems Society, 2002: 12.

Table 2 – Summarizing data regarding electronic referral system from January 2011 to June 2011.

Month No. of physiciansusing the system

No. of insuredmembers visitingphysician

Visits No. of electronicreferrals

Cost of electronicreferrals (€)

Mean cost ofelectronicreferrals (€)

January 2011 1,701 54,571 55,249 13,996 807,272.49 57.68February 2011 2,859 94,088 94,787 19,806 1,072,562.93 54.15March 2011 3,559 130,945 136,720 43,205 2,039,216.66 47.20April 2011 2,880 109,972 122,047 37,216 1,755,829.23 47.18May 2011 2,482 95,296 131,647 37,804 1,875,444.88 49.61June 2011 2,204 88,154 129,354 41,200 1,979,538.59 48.05








































[14] Eysenbach G. What is e-health? J Med Internet Res 2001;33:E20.[15] Deloitte/Ipsos. eHealth Benchmarking (Phase III): Final Report. Brussels:

Deloitte/Ipsos, 2011.[16] Zmud RW, Olson MH, Hauser R. Field experiment in MIS research.

Boston, MA: Harvard Business School Research Colloquium, 1989:97–111.

[17] Hellenic Association of Pharmaceutical Companies. ThePharmaceutical Market in Greece: Facts and Figures. Athens, Greece:Hellenic Association of Pharmaceutical Companies, 2003.

[18] World Health Organization. WHO Highlights on Health in Greece.Copenhagen, Denmark: World Health Organization, 2004.

[19] Koutsouris D, Aggelidis P, Berler A, Tagaris A. Integration of HealthcareInformation Systems. Athens, Greece: ebusiness Forum, Z3 ResearchTeam, 2005.

[20] Rogers EM. Diffusion of Innovations (4th ed.). New York: The Free Press,1995.

[21] Hellenic Ministry of Finance. Memorandum of understandingbetween the European Commision acting on behalf of the Euro-area member states, and the Hellenic Republic, 2010, p. 43.Available from: http://ec.europa.eu/economy_finance/eu_borrower/mou/2012-03-01-greece-mou_en.pdf. [AccessedNovember 22, 2010].





















http://ec.europa.eu/economy_finance/eu_borrower/mou/2012-03-01-greece-mou_en.pdf

http://ec.europa.eu/economy_finance/eu_borrower/mou/2012-03-01-greece-mou_en.pdf



CONCEPTUAL PAPER

Recommendations for Reporting Pharmacoeconomic Evaluations in EgyptGihan H. Elsisi, MSc1,�, Zoltán Kaló, MSc, MD, PhD2, Randa Eldessouki, MSc, MD3,4, Mahmoud D. Elmahdawy, PharmD5,6,Amr Saad, MSc, PhD7, Samah Ragab, MPA8, Amr M. Elshalakani, MD, MBA9, Sherif Abaza, MBA10

1Pharmacoeconomic Unit, Central Administration for Pharmaceutical Affairs, Cairo, Egypt; 2Health Economics Research Centre, Eötvös Loránd University,Budapest, Hungary; 3Scientific and Health Policy Initiatives, International Society for Pharmacoeconomics and Outcomes Research, NJ, USA; 4Faculty of Medicine,Fayoum University, Fayoum, Egypt; 5Hospital Pharmacy Administration, Central Administration for Pharmaceutical Affairs, Cairo, Egypt; 6Misr InternationalUniversity, School of Pharmacy, Cairo, Egypt; 7Pharmacovigilance Center, Central Administration for Pharmaceutical Affairs, Cairo, Egypt; 8Head Technical Office,Central Administration for Pharmaceutical Affairs, Cairo, Egypt; 9Health Economics Unit, Ministry of Health, Cairo, Egypt; 10Market Access, Roche, Cairo, Egypt

A B S T R A C T

Objective: Introduction of economic evaluations for pharmaceuticalsor other health technologies can help the optimization of outcomesfrom resource allocations. This article aims to provide recommenda-tions for researchers in presenting pharmacoeconomic evaluations inEgypt with special focus on pricing and/or reimbursement applica-tions of pharmaceuticals. Methods: The Minister of Health approvedthe initiative of establishing a focus group of decision makers thatincluded academic and industry experts with experience in healtheconomics, pharmacovigilance, and clinical pharmacy. The focusgroup has reviewed 17 economic evaluation guidelines available onthe Web site of the International Society for Pharmacoeconomics andOutcomes Research for reporting health economic evaluations. Todevelop core assumptions before preparing a draft report, focus groupmeetings were held on a regular basis starting June 2012. Therecommendations were developed by using the Quasi-Delphi method,taking into account current practices and capacities for conductingpharmacoeconomic evaluations in Egypt. Conclusions: Worldwide,health care decision makers are challenged to set priorities in anenvironment in which the demand for health care services outweighsthe allocated resources. Effective pharmaceutical pricing and

reimbursement systems, based on health technology assessment(HTA) that encompasses economic evaluations, are essential to anefficient sustainable health care system. The Egyptian Ministry ofHealth and Population was encouraged to establish a pharmacoeco-nomic unit, as an initial step, for the support of pricing and reim-bursement decisions. We anticipate that standardization of reportingwould lead to a progressive improvement in the quality of submis-sions over time and provide the Egyptian health care system withhealth economic evidence often unavailable in the past. Therefore,recommendations for pharmacoeconomic evaluations provide anessential tool for the support of a transparent and uniform processin the evaluation of the clinical benefit and costs of drugs that do notrely on the use of low acquisition cost as the primary basis forselection. These recommendations will help inform health caredecisions in improving health care systems and achieving betterhealth for the Egyptian population.Keywords: economic evaluation, Egypt, recommendations, reporting.


Introduction

Egypt’s general budget devotes limited amounts to the healthsector. In the period 2008 to 2009, Egypt spent LE 61.4 billion(Egyptian pounds) on health, which represents 5.9% of thecountry’s gross domestic product. Out of the total health careexpenditure, pharmaceutical expenses constitute a large portion,34%. [1]. In addition, over the past 16 years, the share of out-of-pocket spending in total health spending has increased dramat-ically from 51% to 72% [2]. These numbers suggest the increasingneed for optimizing the limited resources available. With thegrowing public demand for improving health care services andreducing the out-of-pocket expenses, economic evaluations of

pharmaceuticals and health technologies are critical for efficientallocation of the limited resources.

To better allocate resources and with the growing awarenessof the importance of health technology assessment (HTA), theMinistry of Health and Population (MOHP) established a pharma-coeconomic unit to support and inform pricing and reimburse-ment decisions [3]. No economic evaluations guidelines orstandards, however, have been set up yet.

This article provides recommendations based on reviewingother countries’ national guidelines for economic evaluation aswell as experts’ opinions. Other factors influencing the feasibilityof conducting such studies in Egypt, including the complexity ofthe health sector, the availability of data on health care outcomes





E-mail: [email protected]; [email protected].

� Address correspondence to: Gihan H. Elsisi, Central Administration for Pharmaceutical Affairs, 21 Abd Elaziz Alsoud Street, PO 11451,Elmanial, Cairo, Egypt.


dx.doi.org/10.1016/j.vhri.2013.06.014

dx.doi.org/10.1016/j.vhri.2013.06.014

dx.doi.org/10.1016/j.vhri.2013.06.014



dx.doi.org/10.1016/j.vhri.2013.06.014

and the costs data, and current capacities for conducting phar-macoeconomic evaluations, were put into consideration in devel-oping these recommendations.

Objective

This article aims to provide recommendations for researchers topresent pharmacoeconomic evaluations in Egypt with specialfocus on pricing and/or reimbursement applications of pharma-ceuticals. Policymakers are encouraged to consider these recom-mendations in developing the national guidelines for theeconomic evaluation of pharmaceuticals.

Methods

As a self-initiated activity by government personnel, with theapproval of the Ministry of Health at the time, a focus group wasformed. The aim of the focus group was to develop a set ofrecommendations and standards for economic evaluation stud-ies used in applying for reimbursement and coverage to 1)promote the concept of combining efficacy, safety, effectiveness,and economic evaluation in the decision-making process; 2)provide instructions for drug manufacturers: how to supplyinformation directly to health care decision makers to supportthe use of their products; and 3) emphasize that simple assess-ment of acquisition cost is not a sufficient approach for thecontrol of overall health care expenditures.

To develop the recommendations, two steps were under-taken. The first step was to review the available nationaleconomic guidelines. It included a review of 17 recently publishednational economic evaluation guidelines for conducting andreporting of economic evaluations (Table 1) that included anEnglish version available on the Web site of the InternationalSociety for Pharmacoeconomics and Outcomes Research [4].

The second step was to solicit inputs and feedback from keyleaders and stakeholders through focus groups. For a compre-hensive representation of key stakeholders in health care, focusgroups included decision makers experienced in health econom-ics, pharmacovigilance, and clinical pharmacy, health providersas well as researchers and experts selected from both industryand academia, as shown in Table 2.

A consensus approach developed by using the Quasi-Delphimethod consisted of an iterative series of meetings and inter-rogations. Anonymous responses were synthesized into a seriesof statements. Then, the synthesized statements were submittedto the focus group members for comment until convergence orstasis of opinion was identified in the third round.

Starting June 2012, focus group meetings were held on aregular basis to develop core assumptions before preparing adraft report. The discussions were recorded in written minutes.The recommendations were developed by consensus approach,taking into account current practices and capacities for conduct-ing pharmacoeconomic evaluations in Egypt.

Developing Recommendations for ReportingPharmacoeconomic Evaluations

Disease and Product Background

Economic evaluations should provide information about the epi-demiology of the disease and treatment pathways according tomost recent treatment guidelines. Data on the product shouldinclude pharmacological class, proposed dosing regimen, route ofadministration, and results of clinical studies performed to date [5].

Study Design

The study question should address the needs of the decisionmakers by clearly establishing the context of the study. It shouldprovide details of the study perspective, the proposed productand its comparator(s), the target population, and the effect onspecific subgroups where appropriate. Secondary questions thatrelate to the primary study question should be clearly stated [6].

Perspective should be relevant to the research question andadapted to benefits gained by the health care system. Theperspective adopted should maximize the health gain for thepopulation while representing the most efficient use of the finiteresources available to the Ministry of Health [7]. It should includedirect medical costs as well as additional costs, savings, or otherbenefits when data are available.

The proposed product should be used primarily in theapproved indications with detailed information about its

Table 1 – Focus group members’ information.

Member of FocusGroup

Degree Title Organization GovernmentEmployee

Gihan H. Elsisi MSc Head of PharmacoeconomicUnit

Central Administration forPharmaceutical Affairs, Cairo, Egypt

Yes

Randa Eldessouki MSc, MD Director, Scientific and HealthPolicy Initiatives/Lecturer

International Society forPharmacoeconomics and OutcomesResearch, NJ, USA/Faculty ofMedicine, Fayoum University, Egypt

No

Mahmoud D. Elmahdawy PharmD Manager of Hospital PharmacyAdministration/Part TimeLecturer of clinical pharmacy

Central Administration forPharmaceutical Affairs/MisrInternational University, Cairo,Egypt

Yes

Amr Saad MSc, PhD Head of PharmacovigilanceCenter


Yes

Samah Ragab MPA Director of the TechnicalSupport Office


Yes

Amr M. Elshalakani MBBch,MSc, MBA

Head of Health Economics Unit Ministry Of Health, Cairo, Egypt Yes

Sherif Abaza MBA Market Access & GovernmentalAffairs Manager

Hoffmann-La Roche Ltd. Cairo, Egypt No


technical characteristics (to differentiate it from its comparators),regulatory status, and the specific application.

The selection of the comparator has to be justified. Compara-tors should be policy relevant; therefore, widely used and reim-bursed health care technology for a given patient group andindication is the preferred option. If no such technologies arereimbursed in the tender list at the time the assessment isconducted, the investigated product can be compared with themost frequently used technologies to treat the same patientgroups. If a new product is used as first-line, second-line, orthird-line therapy, it should be compared with first-, second-, orthird-line therapies, respectively.

The targeted population should include both those who areinsured by the Egyptian health system and those who are unin-sured. Parameters to define the population include baseline dem-ographic characteristics, disease characteristics, treatment setting,the context of past treatment, and any confounders adjusted [5].

Specific subgroups should be identified for those for whomclinical effectiveness and cost-effectiveness may be expected todiffer from those of the overall population. Stratified analysis usedto quantify the differences in cost-effectiveness that may exist indifferent subgroups is recommended because it may contributeimportant information to the final advice. The evidence support-ing the clinical plausibility of the subgroup effect should be fullydocumented, including details of statistical analysis [8].

Appropriate Pharmacoeconomic Method

The choice of method of analysis depends on the researchquestion and must be justified. If the compared health technol-ogies result in equal health gain, cost minimization analysis isthe preferred analytical approach.

If at least one of the compared health technologies is betterthan the other, and the clinical benefit can be aggregated andinterpreted as naturalistic clinical outcomes, cost-effectivenessanalysis (CEA) is the preferred method. CEA, where an inter-mediate marker is chosen, must have a validated, well-established link with an important hard end point (e.g., patientsurvival, heart attack, and bone fracture) [9]. Because the meas-ure of primary clinical outcome may differ in different therapeu-tic areas, CEA cannot be used to compare or rank the cost-effectiveness of a broad set of products.

If the quality of life of patients is an important clinicaloutcome in the treatment course of patients, cost-utility analysisis the preferred analytical approach. In cost-utility analysis, thehealth gain is expressed in a combined single measure of life-years and health-related quality of life (HRQOL), for example, inquality-adjusted life-years (QALYs) [10]. Ignoring quality-of-lifedifferences among products would provide less than completedata to decision makers to address the health care dilemma ofwhere to allocate resources [11]. Adherence to the reference case

Table 2 – The national health economic guidelines reviewed by the focus group members.

Title of the document Source, Country PublishedYear

Guidelines for the Economic Evaluation of Health Technologies: Canada Canadian Agency for Drugs and Technologies inHealth (CADTH), Canada

2006

Guidelines of Methodological Standards for PharmacoeconomicEvaluations in Taiwan

Taiwan Society for Pharmacoeconomics andOutcomes Research, Taiwan

2006

Prescription for Pharmacoeconomic Analysis: Methods for Cost-utilityAnalysis

The Pharmaceutical Management Agency(PHARMAC), New Zealand

2007

Guidance to Manufacturers for Completion of New Product AssessmentForm

Scottish Medicines Consortium, Scotland 2007

Guide to the Methods of Technology Appraisal National Institute for Health and ClinicalExcellence (NICE), England and Wales

2008

Guidelines for Preparing Submissions to the Pharmaceutical BenefitsAdvisory Committee

Pharmaceutical Benefits Advisory Committee(PBAC), Australia

2008

Guidelines for Pharmacoeconomic Evaluations in Belgium Belgian Health Care Knowledge Center (KCE),Belgium

2008

Health Technology Assessment Guideline Journal of the Medical Association of Thailand,Thailand

2008

The Academy of Managed Care Pharmacy Format for FormularySubmissions

Academy Of Managed Care Pharmacy, UnitedStates

2009

General Methods for the Assessment of the Relation of Benefits to Costs German national institute for quality andefficiency in health care (IQWiG), Germany

2009

Guidelines for conducting Health Technology Assessment Poland Agency for Health TechnologyAssessment, Poland

2009

Decree of the Ministry of Social Affairs and Health on applications andprice notifications made to the Pharmaceuticals Pricing Board–Appendix: Guidelines for Preparing a Health Economic Evaluation

Ministry of Social Affairs and Health, Finland 2009

Procedure for Clinical and Economic Evaluation of Drug Lists That AreSubmitted for Reimbursement Coverage from Public Health CareBudget.

ISPOR Russia HTA Regional chapter, RussianState Medical University, Russian Federation

2010

The Guidelines for Pharmacoeconomic Evaluations of Medicines andScheduled Substances

National Department of Health, South Africa 2010

Guidelines for the Submission of a Request to Include a PharmaceuticalProduct in the National List of Health Services

Pharmaceutical Administration, Israel 2010

Guidelines for the Economic Evaluation of Health Technologies inIreland

Health Information and Quality Authority, TheNational Centre for Pharmacoeconomics,Ireland

2010

Guidelines on How to conduct Pharmacoeconomic Analyses Norwegian Medicines Agency, Norway 2012


approach for estimating QALYs for inclusion in economic evalu-ations would facilitate comparability [12].

Time Horizon

In choosing the time horizon, it should be ensured that the chosenoutcome and the resource consumption of the treatment alterna-tives are observable in this period to reflect the course of the diseaseand the effects of the interventions. The same time horizon shouldbe applied to both costs and outcomes [9]. A decision to use a shortertime frame should be justified. When extrapolating data beyond theduration of the study, assumptions regarding future treatmenteffects and disease progression should be clearly outlined. Censoringmight be used to account for the incomplete information [13].

Choice of Outcome Measure

The choice of outcome parameters depends both on the indica-tion and on the research question. Primary outcome measures arethe first choice whenever possible. When an intermediate endpoint is used, it must have a high degree of predictability of thefinal end point.

HRQOL is an appropriate outcome indicator for the evaluation ofhealth status. HRQOL can be measured by using generic question-naires, disease-specific questionnaires, or preference-based meas-ures. If HRQOL is to be included in the study design, this variablemust be measured by validated instruments. The direct use of theEuroQol five-dimensional questionnaire, six-dimensional healthstate short form (derived from short-form 36 health survey), orsimilar generic measures is recommended, because they are easy touse and interpret and are based on preferences of the generalpublic. If the use of disease-specific HRQOL instruments increasesthe sensitivity of measurement, mapping of disease-specific HRQOLresults with the EuroQol five-dimensional questionnaire or similargeneric measures can be useful to translate the findings into QALYs.

Information on the changes in the health state should bereported directly by the patient or the caregiver. A valuation ofthese changes in the health state should then be reported for thegeneral population. The outcome parameter chosen must besensitive, valid, and consistent [14].

Synthesis of Clinical and Economic Evidence

Evidence synthesis has to be based on objective, systematic, andreproducible search criteria. Estimation of health gain must bebased on scientific literature review and/or results of primarydata collection, and the best available evidence should beconsidered. Meta-analysis based on large randomized controlledtrials is the highest hierarchy of evidence with the heterogeneityof data accounted for. If compared drug therapies differ inadherence or persistence of patients, then these factors shouldbe incorporated in calculating the relative effectiveness. In caseof orphan drugs where randomized controlled clinical studieshave not been conducted, the results of uncontrolled clinicalstudies can be accepted, including studies with small samplesize. All product safety data need to be included whether fromclinical studies or from national and foreign pharmacovigilancecenters and patient registries with attention given to those thatdiffer substantively among the products being compared [15].Economic evidence should be synthesized from systematicreview of the local data sources and the best available evidence.

Costs Determination

Resource use data should be obtained mainly from primary datacollection (e.g. health care providers or non-interventional stud-ies) from Egypt; if not available, secondary data sources such aslocal administration, accounting data, or patient chart review data

can be used. Official sources of unit cost data for products(e.g., tender lists) are preferable. In the absence of a publishedtender list price, the price submitted by a manufacturer for aproduct may be used. The quality, validity, relevance, and gen-eralizability of local data should be clearly described. Bothestimated consumption of resources and their unit prices mustreflect real-world settings in Egypt because relative and absoluteprice levels differ among countries [16].

Resource use and costs should be identified, measured in theirnatural units and values [17]. The primary perspective for thesestudies is the overall health care services. Therefore, the resour-ces that should be considered are direct medical costs, whichinclude drugs, medical devices, medical services including pro-cedures, laboratory, or diagnostic tests, hospital services andemergency department visits, and primary care visits. Otherdirect nonmedical and indirect costs paid by patients, includinglost productivity costs, might be included only in the sensitivityanalysis. If indirect costs are included in the analysis, the ration-ality of the costs and how they are estimated should beexplained. Current and future costs arising as a consequence ofa product, and occurring during the specified time frame of thestudy, should also be included. Mean values should be used.Different costs or costs of the same resources that are used indifferent quantities should be included in the analysis [18].

Out of the two general approaches to determine costs, micro-costing and macro-costing, macrocosting is preferred [19]. Thesource of cost data must be reported in detail. Data should be themost recently available, with the cost year specified. Retrospectiveinput costs should be inflated to the most recent calendar year byusing the Consumer Price Index for health [20]. The drug cost usedshould reflect the formulation and pack size that gives the lowestcost. For drugs available in the outpatient pharmacies, the fullpublic price should be used for calculating costs. For hospitalproducts, the wholesale price should be used for cost calculations.Future costs should be calculated at constant current costs; there-fore, results are not subject to uncertainty in future inflation rates.

Modeling

Economic modeling based on prospectively collected data is thepreferred method by decision makers in an increasing number ofcountries to aggregate the expected costs and health effects forall options relating to appropriate population and subpopula-tions, based on the full range of existing evidence [21]. The majoraim of applying modeling techniques is to aggregate short- andlong-term outcomes in the most appropriate time horizon.

The results of economic modeling studies presented should takeinto account the following requirements: 1) the model should bedescribed in detail and should correspond to real practice of patientmanagement; 2) the model should be as simple as possible, andeasily understood; and 3) to facilitate assessment of the outputs of amodel, full documentation of the structure, data elements, andvalidation of the model should be addressed in a clear manner,with justification provided for the options chosen and presentedthrough diagrams (e.g., decision trees and Markov models) [22].

In addition, the model should be adapted to exclude clinicalevents not expected to differ among the comparator products[20]. For state transition models, such as Markov models, thecycle length should be sufficiently short to ensure that multiplechanges in disease, treatment decisions, or costs do not occurwithin a single cycle. Heterogeneity in the population should beaccounted by disaggregating the population into clinically plau-sible subgroups that require different structural assumptions.The internal validity of the model should be tested before usingto ensure that the model is robust. The external validity shouldbe tested by comparison of the results with those generated byother models and explaining differences if they exist.


Discounting

Discounting should be made according to the time horizon. Anycosts or outcomes occurring beyond 1 year should be discountedby using standard methods [19]. For comparability of resultsacross evaluations, it is important that a common discount rateis used. Because constant prices and outcomes are used in theeconomic evaluation, there is no need to take into accountinflation in the discount rate. A real discount rate of 3.5% peryear should be used for both costs and health gains. The discountrate should be varied from 2% to 6% in the sensitivity analysis.

Uncertainty

Data for a health economic analysis are derived from varioussources, and this may be incomplete and affected by uncertain-ties. In a sensitivity analysis, critical component(s) in the calcu-lation should be varied through a relevant range or from theworst case to the best case, and the results recalculated [13].

Probabilistic sensitivity analysis (PSA) is an appropriatemethod for exploring uncertainty around the true mean valuesof cost and efficacy inputs in decision-analytic modeling. In PSA,however, probability distributions are applied by using specifiedplausible ranges for the key parameters rather than the use ofvaried point estimates for each parameter. Its results are difficultto interpret for decision makers, while the stochastic approach,such as deterministic sensitivity analysis (DSA), examines howparameter variables (included as point estimates) affect themodel output [23]. We propose, given the difficulty in interpretingthe PSA, that DSA should be required, while PSA remainsoptional.

To avoid potential bias and uncertainty that arise from themodeling process, assumptions about the model structure shouldbe clearly stated and justified and their impact on cost-effectiveness explored though a series of plausible scenarioanalyses so that whether the study results will be changed canbe observed. All choices and the ranges of the parameters, andthe method used in sensitivity analysis, should be clearlyexplained.

Present Study Results

Total costs and health outcomes must be reported separately,and the aggregated result be explained. All parameters used inthe estimation of clinical effectiveness and cost-effectivenessshould be itemized in a tabular form with data sources trans-parently. Negative results should be reported. Incremental cost-effectiveness ratio has to be calculated, unless one of thecompared health technologies dominates the other one. Inaddition, the potential impact of the introduction of the newtreatment on the society needs to be assessed [24].

Where more than two products are being compared, theresults should be presented in the order of increasing costs andthe incremental cost-effectiveness ratio calculated by comparingeach product with the one above it, excluding those products thatare dominated. Equity issues, affordability, and resource con-straints should be considered in judging the cost-effectiveness ofa product for reimbursement [20].

Tornado diagrams are useful tools to display DSA. If PSAs areperformed, the probability that the intervention is cost-effectiveat a range of threshold values should be reported and the datashould be displayed graphically to facilitate the uncertaintyinterpretation [9].

Equity and Generalizability Issues

To meet the needs of the decision makers, an attempt should bemade to include equity considerations in the study report. The

equity assumption of the basic case in economic evaluationsmeans that all patients should have a fair participation oppor-tunity and obtain the expected treatment outcomes.

To determine equity in economic evaluation, we propose thatall lives, life-years, or QALYs should be valued equally, regardlessof the age, gender, or socioeconomic status of individuals in thepopulation [12]. The equity assumption should be included inevery model and analytical method of economic evaluations andmust be clearly stated.

Analysts must consider two specific areas of concern regard-ing the generalizability of clinical and economic data in theassessment of technologies. The first area of concern is theextent to which the clinical efficacy data are representative ofthe likely effectiveness and similarly the extent to which eco-nomic data are representative of the costs and resource utiliza-tion [8]. The second area of concern is the generalizability of theeconomic and clinical data across different patient ages andgenders as well as regional differences in health care practicewithin Egypt. These areas of concern should be identified anddiscussed, and the likely effect on the results and conclusions ofthe report should be highlighted [25].

Discussion

There is an increasing need for justification of resource alloca-tions and policy decisions, especially with the scarcity of publicresources. Fig. 1 shows that among all middle-income countriesin the region, Egypt invests a smaller proportion of its grossdomestic product on health care [1]. Investments in economicevaluation studies and development of pharmacoecomic guide-lines and expertise will help in allocating these limited resourcesin the most efficient way to improve health care services.

The current reimbursement decisions in Egypt are based onthe lowest price after clinical review and approval of efficacy andsafety of the medication by the Procurement Technical Commit-tee. The Procurement Technical Committee reviews all MOHPhospitals and primary care units’ needs of medications andapplications submitted by drug manufacturers. It then decideswhether this medication is to be listed or not, according topharmacokinetics, pharmacodynamics, safety, and efficacy.Then, the applications go to the Committee for Financial Offersat the MOHP to review the financial issues and decide which drugmanufacturer or wholesaler, the one that presents the lowestprice for each active ingredient (medication), is to get reimburse-ment. Drug manufacturers or wholesalers who submitted iden-tical price levels for the same active ingredient are givenreimbursement by an equal process. Fig. 2 presents the decisionmakers and influencers in the Egyptian pricing and reimburse-ment decision-making processes [26].

There is a growing need to incorporate high-quality economicevaluation studies into the reimbursement decision-makingprocess to adequately evaluate clinical and economic benefitsof medications in addition to the assessment of their acquisitioncosts. These evaluations will improve decision making withprioritizing our resources, which results in reducing our hugeexpenditure on pharmaceuticals and save these resources to beallocated to other cost-effective health technologies. In Egypt,there is no limited budget that should be allocated for drugcoverage only but is allocated to the whole health sector.

The submission of an economic evaluation is currently rec-ommended in Egypt. And an economic evaluation guideline tostandardize the process and provide a transparent and uniformapproach was approved by the Ministry of Health. There is a bigchance that these recommendations will be implemented inEgypt. Policymakers are encouraged to consider these


recommendations in developing the national guidelines for theeconomic evaluation of pharmaceuticals.

The Canadian guidelines reported that by providing standardsfor conducting and reporting of economic evaluations, the cur-rent limitations of evaluations can be addressed and lead tobetter study [9]. It is important to note that the standardization ofreporting and other policies in the United States shared in thebulk of the estimated $2 trillion savings [27]. We anticipate thatthe standardization of reporting would lead to a progressiveimprovement in the quality of submissions over time and providethe Egyptian health care system with data often unavailable inthe past.

In developing those recommendations, we chose to build onthe learning experience from other countries and modify andadapt the knowledge acquired to fit the Egyptian setting. In doingso, duplication of efforts and use of resources much neededelsewhere are avoided. As a rule, certain elements of HTA reportsare transferable, but adjustment to local data is absolutelynecessary [16]. Copying recommendations based on internationalHTA without local adjustment may do more harm than good.Putting this into consideration, our recommendations were tail-ored to the current settings and environment in Egypt while usingthe current guidelines as an initial benchmark. Our starting pointis built on many years of experiences and expertise worldwide.

0.00%

1.00%

2.00%

3.00%

4.00%

5.00%

6.00%

7.00%

8.00%

9.00%

10.00%

Overall Health Care Spending as Percentage of Gross Domes�c Product

Fig. 1 – Egypt in comparison with other middle-income countries in overall health care spending as percentage of grossdomestic product. Data from USAID [1].

Fig. 2 – The Egyptian pricing and reimbursement decision-making processes [26]. Bold boxes, decision-making bodies; boxes,decision influencer bodies; bold arrows, required step in the decision-making process; arrows, may or may not affectdecision.


Table 3 – Key elements of the recommendations for reporting pharmacoeconomic evaluations in Egypt.

Key elements The Egyptian recommendations Differences/similarities across

the nationalguidelinesreviewed

Rationale for inclusion in the Egyptiansetting

Perspective It should be relevant to the researchquestion and adapted to benefits gainedby the health care system.

Common It is a common agreed-upon element thatcaptures all the benefits when data areavailable representing the most efficientuse of the finite resources.

Indication It should be used in the approvedindications.

Common According to the Egyptian Ministry of Healthregulations, the use of the product inunapproved indications is forbidden.

Choice ofcomparator

Comparators should be policy relevant. Thewidely used and reimbursed health caretechnology for a given patient group is thepreferred option.

Different Because of policy-related problems such asdrug supply shortage, we have to use theavailable technologies.

Targetpopulation

Both those who are insured and uninsuredby the Egyptian health care system.

Different Because of the existing widespreaduninsured population that is covered byother forms of health coverage, there is aneed to assess the effectiveness amongdifferent categories of access to healthcare.

Subgroupanalysis

Only for those for whom clinicaleffectiveness and cost-effectiveness maybe expected to differ from that of theoverall population.

Common When a distinct group differs from theoverall population, a subgroup analysis isessential to reflect the actual clinicalbenefit and provide an accurate estimateof the cost-effectiveness of the therapyacross all population groups to betterinform decision on reimbursement.

Preferredanalyticaltechnique

Any of CMA, CEA, and CUA considered. Different It depends on the research question. Whenthe clinical benefit is interpreted asnaturalistic clinical outcomes, CEA is thepreferred method while CUA is the secondoption because the concept of usingQALYs might not be well understood bythe majority of decision makers.

Time horizon It should be ensured that the chosenoutcome and the resource consumption ofthe treatment alternatives are observablein this period.

Common To accurately reflect the course of thedisease and the total effects of theinterventions.

Choice ofoutcomemeasure

Primary outcome measures are the firstchoice. CEA, where the intermediatemarker is chosen, must have a validated,well-established link with an importanthard end point. In CUA, outcomes aremeasured in QALYs gained.

Common It depends both on the indication and on theresearch question.

Preferredmethod toderive utility

The direct use of the EQ-5D questionnaire,SF-6D, or similar generic measures isrecommended.

Common They are easy to use and interpret and arerelevant to the Egyptian publiceducational level and preferences. After aperiod of time allowing knowledgebuilding and according to the learningcurve, more sophisticated instrumentsmight be considered.

Synthesis ofclinical andeconomicevidence

Evidence synthesis has to be based onobjective, systematic, and reproduciblesearch criteria. The results of meta-analysis are preferable with theheterogeneity of data accounted for.Economic evidence should be synthesizedfrom systematic review of the local datasources and the best available evidence.

Common Meta-analysis based on large randomizedcontrolled trials is the highest hierarchy ofclinical evidence and is recommended forclinical benefit evidence. However, econo-mic benefit evidence is nontransferableamong the countries and should beobtained from local data sources and thebest available evidence.


Key elements common across all national guidelinesreviewed were included in the recommendations. Other keyelements differed between the various guidelines such as thechoice of comparator, preferred analytical technique, targetpopulation, costs to be included, and uncertainty. Through aconsensus approach between all focus group membersfor these elements, we recommended the best fit to Egyptiansettings that are applicable to the current Egyptian environ-ment. A summary of the key elements of the recommendationsand the rationale for their inclusion within the Egy-ptian setting are presented in Table 3 highlighting the elementsthat were common across all guidelines and the ones thatvaried.

HTA implementation in Egypt, however, is significantly chal-lenged by the diversity and heterogeneity of the health caresystem, limited tradition for national treatment guidelines, andlimited availability of epidemiological, health outcomes, and costdata. Data for economic evaluations are low quality, region andprovider specific, unavailable in electronic records, and, in mostcases, not updated. So, the common opinion is “HTA cannot beimplemented in Egypt.” In fact, there are no perfect data forhealth care research; we have to assess realistically how wrongthey have to be not to be considered useful. Having some data isbetter than having no data at all; conducting pharmacoeconomicevaluations and outcomes research in Egypt would also greatlyimprove the quality of current data.

Table 3 – continued

Key elements The Egyptian recommendations Differences/similarities across

the nationalguidelinesreviewed

Rationale for inclusion in the Egyptiansetting

Costs to beincluded

Direct medical costs as well as additionalcosts, savings, or other benefits when dataare available.

Different In most cases, data and information onindirect costs are lacking in Egypt;therefore, direct costs estimation isrecommended.

Sources ofcosts

Primary data collection; if unavailable,secondary data sources can be used suchas local administration, accounting data,and patient chart review. Official sourcesof unit cost data for products (e.g., tenderlists) are preferable.

Common Both estimated consumption of resourcesand their unit prices must reflect real-world settings in Egypt because relativeand absolute price levels differ amongcountries.

Modeling Modeling options include decision trees andMarkov models. The model should bedescribed in detail and should correspondto real practice of patient management.

Common These models are easy to use, interpret, andaggregate the expected costs and short-and long-term outcomes captured in themost appropriate time horizon relating tothe Egyptian population and subpopula-tions.

Discountingcosts andoutcomes

A discount rate of 3.5% per year should beused for costs and outcomes.

Common Because constant prices and outcomes areused in the economic evaluation, there isno need to take into account inflation inthe discount rate.

Uncertainty Critical component(s) in the calculationshould be varied through a relevant rangeor from the worst case to the best case.DSA should be required, while PSAremains optional.

Different With the current level of knowledge, resultsof PSA are difficult to interpret bypersonnel reviewing the studies forcoverage decisions.

Equity issues All lives, life-years, or QALYs should bevalued equally, regardless of the age,gender, or socioeconomic status ofindividuals in the population.

Common All patients should have a fair participationopportunity and obtain the expectedtreatment outcomes.

Generalizability The generalizability and the extent to whichthe clinical efficacy data and theeconomic data are representative shouldbe identified and discussed.

Common Because of the presence of wide regionaldifferences in health care practice amongurban and rural areas, generalizability ofthe studies should be discussed in detail.

Presentingresults

Total costs and health outcomes must bereported separately, and the aggregatedresult be explained. ICER has to becalculated. The probability that theintervention is cost-effective at a range ofthreshold values should be reported anddisplayed graphically.

Common Detailed information should be provided tofacilitate the interpretation of results,thus allowing for a more transparent anduniform process for the final coveragedecision.

CEA, cost-effectiveness analysis; CMA, cost minimization analysis; CUA, cost-utility analysis; DSA, deterministic sensitivity analysis; EQ-5D,EuroQol five-dimensional; SF-6D, six-dimensional health state short form (derived from short-form 36 health survey); ICER, incremental cost-effectiveness ratio; PSA, probabilistic sensitivity analysis; QALYs, quality-adjusted life-years.


Therefore, it will be important to evaluate the effect of theimplementation of these recommendations on reporting infuture economic evaluations in a manner similar to ConsolidatedHealth Economic Evaluation Reporting Standards (CHEERS):ISPOR Task force report [28]. As methods for the conduct ofeconomic evaluation continue to evolve, it will also be importantto revisit our recommendations. These recommendations werepresented to the Assistant Minister of Health, and initial stepsrequired to start building the capacity of the pharmacoeconomicunit are underway. A young generation of government personnelis enthusiastic to enter this field; recent graduates from the firsthealth economic diploma program in the Middle East are keen tofacilitate the implementation of HTA in Egypt.

Conclusions

Worldwide, health care decision makers are challenged to setpriorities in an environment in which the demand for health careservices outweighs the allocated resources [29]. Effective pharma-ceutical pricing and reimbursement systems, based on HTA thatencompasses economic evaluations, are essential to an efficientsustainable health care system [30]. The MOHP was encouraged toestablish a pharmacoeconomic unit, as an initial step, for thesupport of pricing and reimbursement decisions. We anticipatethat the standardization of reporting would lead to a progressiveimprovement in the quality of submissions over time and providethe Egyptian health care system with health economic evidenceoften unavailable in the past. Therefore, recommendations forpharmacoeconomic evaluations provide an essential tool for thesupport of a transparent and uniform process in the evaluation ofthe clinical benefit and costs of drugs that do not rely on the use oflow acquisition cost as the primary basis for selection. Eventually,these recommendations will help inform the health care decisionsin improving health care systems and achieving better health forthe Egyptian population.


R E F E R E N C E S

[1] USAID. Summary key findings national health accounts 2008/2009. 2009.Available from: http://egypt.usaid.gov/en/procurement/Documents/keyfindings_nha2008_09EN.pdf. [Accessed May 31, 2012].

[2] Egypt’s Central Agency for Public Mobilization and Statistics incollaboration with Egypt’s Ministry of Health and Population.Household Health Care Utilization and Expenditure Survey 2010. Cairo,Egypt: Ministry of Health and Population, 2010.

[3] International Society for Pharmacoeconomics and Outcomes Research.Pros and cons of pricing and reimbursement systems in Egypt.Available from: http://www.ispor.org/meetings/WashingtonDC0512/releasedpresentations/Updated-Final-Egypt-ISPOR-presentation_06032012.pdf. [Accessed August 15, 2013].

[4] International Society for Pharmacoeconomics and Outcomes Research.Pharmacoeconomic guidelines around the world. Available from:http://www.ispor.org/PEguidelines/index.asp. [Accessed August 15,2013].

[5] Spooner JJ, Gandhi PK, Connelly SB. AMCP Format dossier requests:manufacturer response and formulary implications for one large healthplan. J Manag Care Pharm 2007;13:37–43.

[6] International Society for Pharmacoeconomics and Outcomes Research.Pharmacoeconomic guidelines around the world, Austria. Availablefrom: http://www.ispor.org/PEguidelines/source/Guidelines_Austria.pdf. [Accessed August 15, 2013].

[7] Drummond MF, Schwartz JS, Jonsson B, et al. Key principles for theimproved conduct of health technology assessments for resource

allocation decisions. Int J Technol Assess Health Care 2008;24:244–58.

[8] Fry RN, Avey SG, Sullivan SD. The academy of managed care pharmacyformat for formulary submissions: an evolving standard—a foundationfor managed care pharmacy task force report. Value Health2003;6:505–21.

[9] International Society for Pharmacoeconomics and Outcomes Research.Pharmacoeconomic guidelines around the world, Canada. Availablefrom: http://www.ispor.org/PEguidelines/source/HTAGuidelinesfortheEconomicEvaluationofHealthTechnologies-Canada.pdf.[Accessed August 15, 2013].

[10] Drummond M. Introduction to pharmacoeconomics. Eur J Hosp PharmPract 2008;14:17–9.

[11] Nord E, Daniels N, Kamlet M. QALYs: some challenges. Value Health2008;12:S10–5.

[12] Drummond M, Brixner D, Gold M, et al. Toward a consensus on theQALY. Value Health 2009;12(Suppl.):S31–5.

[13] Berger ML, Bingefors K, Hedblom EC, et al. Health Care Cost, Quality,and Outcomes: ISPOR Book of Terms. (1st ed.). Lawrenceville,New Jersey: ISPOR, 2003.

[14] Elliot R, Payne K. Essentials of Economic Evaluation in Healthcare. (1sted.). London: Pharmaceutical Press, 2005.

[15] International Society for Pharmacoeconomics and Outcomes Research.Pharmacoeconomic guidelines around the world, Russia. Availablefrom: http://www.ispor.org/PEguidelines/source/Russia_PE_Recommendations_english_fnal_13_03.pdf. [Accessed August 15, 2013].

[16] Drummond M, Barbieri M, Cook J, et al. Transferability of economicevaluations across jurisdictions: ISPOR Good Research Practices TaskForce Report. Value Health 2009;12:409–18.

[17] Hay JW, Smeeding J, Carroll NV, et al. Good research practices formeasuring drug costs in cost effectiveness analyses: issues andrecommendations: the ISPOR Drug Cost Task Force Report—part I.Value Health 2009;13:3–7.

[18] International Society for Pharmacoeconomics and Outcomes Research.Pharmacoeconomic guidelines around the world, Taiwan. Availablefrom: http://www.ispor.org/PEguidelines/source/2006_PEG_EN_2009.pdf. [Accessed August 15, 2013].

[19] Rascati KL. Essentials of Pharmacoeconomics. Philadelphia: LippincottWilliams and Wilkins, 2009.

[20] International Society for Pharmacoeconomics and Outcomes Research.Pharmacoeconomic guidelines around the world, Ireland. Availablefrom: http://www.ispor.org/PEguidelines/source/Ireland_Economic_Guidelines_2010.pdf. [Accessed August 15, 2013].

[21] Bodrogi J, Kaló Z. Principles of pharmacoeconomics and their impact onstrategic imperatives of pharmaceutical research and development. Br JPharmacol 2010;159:1367–73.

[22] Weinstein MC, Brien BO, Hornberger J, et al. Principles of good practicefor decision analytic modeling in health-care evaluation: report of theISPOR Task Force on Good Research Practices—modeling studies. ValueHealth 2003;6:9–17.

[23] Gold MR, Siegel JE, Russell LB, et al. Cost-effectiveness in Health andMedicine. New York: Oxford University Press, 1996.

[24] Ngorsuraches S. Defining types of economic evaluation. J Med AssocThailand 2008;91(Suppl.):S21.

[25] Drummond M, Brown R, Fendrick AM, et al. Use of pharmacoeconomicsinformation: report of the ISPOR Task Force on use ofpharmacoeconomic/health economic information in health-caredecision making. Value Health 2003;6:407–16.

[26] International Society for Pharmacoeconomics and Outcomes Research.ISPOR global healthcare systems road map for pharmaceuticals inEgypt. Available from: http://www.ispor.org/HTARoadMaps/EgyptPH.asp. [Accessed August 15, 2013].

[27] Commonwealth Fund. Confronting costs: stabilizing U.S. healthspending while moving toward a high performance health care system.Available from: http://www.commonwealthfund.org/Publications/Fund-Reports/2013/Jan/Confronting-Costs.aspx?page=all. [Accessed August 15, 2013].

[28] Husereau D, Drummond M, Petrou S, et al. Consolidated HealthEconomic Evaluation Reporting Standards (CHEERS)—explanation andelaboration: a report of the ISPOR Health Economic EvaluationPublication Guidelines Good Reporting Practices Task Force. ValueHealth 2013;16:23–50.

[29] Gibson JL, Martin DK, Singer PA. Priority setting for new technologies inmedicine: a transdisciplinary study. BMC Health Serv Res 2002;2:article14.

[30] Eldessouki R, Smith MD. Health care system information sharing: astep toward better health globally. Value Health Regional 2012;1:118–29.



http://egypt.usaid.gov/en/procurement/Documents/keyfindings_nha2008_09EN.pdf

http://egypt.usaid.gov/en/procurement/Documents/keyfindings_nha2008_09EN.pdf







http://www.ispor.org/meetings/WashingtonDC0512/releasedpresentations/Updated-Final-Egypt-ISPOR-presentation_06032012.pdf





http://www.ispor.org/PEguidelines/index.asp






http://www.ispor.org/PEguidelines/source/Guidelines_Austria.pdf

http://www.ispor.org/PEguidelines/source/Guidelines_Austria.pdf











http://www.ispor.org/PEguidelines/source/HTAGuidelinesfortheEconomicEvaluationofHealthTechnologies-Canada.pdf

http://www.ispor.org/PEguidelines/source/HTAGuidelinesfortheEconomicEvaluationofHealthTechnologies-Canada.pdf














http://www.ispor.org/PEguidelines/source/Russia_PE_Recommendations_english_fnal_13_03.pdf

http://www.ispor.org/PEguidelines/source/Russia_PE_Recommendations_english_fnal_13_03.pdf










http://www.ispor.org/PEguidelines/source/2006_PEG_EN_2009.pdf

http://www.ispor.org/PEguidelines/source/2006_PEG_EN_2009.pdf





http://www.ispor.org/PEguidelines/source/Ireland_Economic_Guidelines_2010.pdf

http://www.ispor.org/PEguidelines/source/Ireland_Economic_Guidelines_2010.pdf


















http://www.ispor.org/HTARoadMaps/EgyptPH.asp

http://www.ispor.org/HTARoadMaps/EgyptPH.asp



http://www.commonwealthfund.org/Publications/Fund-Reports/2013/Jan/Confronting-Costs.aspx?page=all
















Note from the Editors

A journal article is sometimes an unfinished piece, which isfinished not after the peer review process [1], but after its originalpublication and the letters sent to the journal by readers and thewider audience. An example of this is the case of the article byAzevedo et al. [2] published in Volume 1, Issue 2, of Value in HealthRegional Issues (VIHRI) (focusing on Latin America). After receivinga Letter to the Editors by Dr. Jose Elias Rizk Aziz regarding thisarticle, the editorial team reviewed the article and sent the letter,as well as other comments, to the authors, who promptlyresponded with a response Letter to the Editors, as well as withan erratum of the original article. We think that these letters andthe new version of the article, published in this VIHRI issue,contribute to the iterative process of improving publicationthrough the interaction of authors, with the readers and editors.

Best regards,

Federico Augustovski, MD, MSc, PhDBon-Ming Yan, PhDDan Greenberg, PhD

Value in Health Regional Issues Co-Editors-in-Chief



R E F E R E N C E S

[1] Smith R. Peer review: a flawed process at the heart of science andjournals. J R Soc Med 2006;99:178–82.

[2] Azevedo VF, Sandorff E, Siemak B, Halbert RJ. Potential regulatory andcommercial environment for biosimilars in Latin America. Value HealthRegional 2012;1:228–34.

V A L U E I N H E A L T H R E G I O N A L I S S U E S 2 ( 2 0 1 3 ) 3 2 8

dx.doi.org/10.1016/j.vhri.2013.05.006

dx.doi.org/10.1016/j.vhri.2013.05.006

dx.doi.org/10.1016/j.vhri.2013.05.006






dx.doi.org/10.1016/j.vhri.2013.05.006



ERRATUM

In the article, “Potential Regulatory and Commercial Environment for Biosimilars in Latin America,” by Valderilio Feijó Azevedo, ErikSandorff, Brian Siemak, Ronald J. Halbert, which appeared in Value in Health Regional Issues, Volume 1, Issue 2 (December 2012), thefollowing items are to be addressed and clarified:

1) On page 231, Table 3, the correct table should be:

2) Also on page 231, the Potential Issues section should read:

Potential issuesAlthough there was no regulatory pathway until recently, biosimilars have been available in Mexico for many years. Because ofthe increasing number of biosimilars coming into the market, the new biosimilar pathway was designed to increase access tobiosimilars while maintaining quality, efficacy, and safety. As the regulatory pathway is still fluid, there are a few potential issuesthat could affect future regulation and subsequent utilization. Thus far only relatively simple biosimilars have been approved foruse in Mexico. As the regulatory pathway will most likely depend on the complexity of the biosimilar, more complex biosimilarsmay be subject to the same clinical trial requirements as the originator biologic, as described in other biosimilar regulatoryguidance [5].

3) On page 232, the Potential Issues section should read:Chile has local production capacity for biologics. As the current regulatory guidelines for biosimilars are in development, thefuture regulatory landscape remains open. Chile’s eventual biosimilar pathway, however, will most likely follow EuropeanMedicines Agency or WHO guidelines. Our findings indicate that there may be standardized guidelines for all biosimilars withspecific requirements depending on the complexity of the molecule. Safety issues could also influence the future regulatoryenvironment. Although potential safety issues might slow down the progress toward developing biosimilar regulatory guidelines,our primary research indicates that the Chilean government seems determined to press forward with a distinct regulatorypathway for biosimilars.

Table 3 – Attributes potentially affecting the development of biosimilar policies in five Latin Americancountries.

Attribute Brazil Argentina Chile Mexico Venezuela

Market size (population) Large(205 million)

Midsized(42 million)

Small(17 million)

Large(115 million)

Midsized(28 million)

Interaction with internationalthought leaders

High High Medium High Medium

Local production capabilities Yes Yes Yes Yes NoSafety issues with biosimilars

in the pastYes No Yes Yes No

Expected differences inregulatoryrequirements based on thecomplexity of the molecule

Yes NA Yes Yes Yes

Regulatory pathway specific forbiosimilars

Yes Yes No Yes No

Year of biosimilar regulatorypathway�

2010 2005/2011 ≥2012 2009 (indevelopment)

None

NA, not available/applicable; WHO, World Health Organization.� WHO guidelines were finalized in 2009.


dx.doi.org/10.1016/j.vhri.2013.05.004

4) On page 233, the second and third paragraph in the Discussion section should read:Of these two pathways, the comparability pathway is fairly conventional in its similarity to the requirements brought forward bythe WHO Similar Biotherapeutic Products guidelines. The individual development pathway, however, includes reducedrequirements, opening the door to lower complexity biosimilar products. Such a pathway deviates from the vision for biosimilarsdefined by the WHO and is reflective of a hypothesis that has emerged from this research that a conceptual bias may exist withinBrazil regarding the appropriate threshold for regulatory requirements for follow-on biologic products.

Safety problems highlight the need for distinct pathways to regulate review, approval, and pharmacovigilance processes forbiosimilars, and argue for greater transparency of government actions to incentivize the domestic production of biologics.Observed safety issues, however, have apparently not had a major impact on governmental actions to increase access tobiosimilars in the region to date.




LETTERS TO THE EDITOR

Response to “Potential Regulatory and Commercial Environment forBiosimilars in Latin America” by Azevedo et al.

I refer to the article titled “Potential regulatory and commercialenvironment for biosimilars in Latin America” written by Aze-vedo et al., published in Value in Health Regional Issues 1 (2012):228–34, which contains incorrect information about biosimilars’regulation and market environment in Mexico.

It must be emphasized that until now there have been noreported cases of anaphylactic shock related to Kikuzubam inMexico or elsewhere and that there are no safety concerns withthe product. The authors based their assertion on a newspaperarticle that consisted of an interview with a competitor’s execu-tive at its facility in Palo Alto, CA [1]. The mentioned rash andanaphylactic shock cases never existed and were never reported.

This is why Probiomed would like to explain to your distin-guished readers that there are no Kikuzubam’s safety dataconcerns published by any scientific journal or official commu-nication from the Mexican health authorities supporting thoseclaims against the product.

In addition, we identified that the Mexican current situationdescribed by Sandorff et al. has several inaccuracies regarding theregulatory framework for biotechnology drugs. The authors omitto mention that Mexican health authorities issued a decreeamending and adding various provisions of the regulation ofhealth products in October 2011 [2].

Although Sandorff et al. mentioned that phase III comparativetrials may not be required for biosimilars and the main factorinfluencing the decision will be the product type, the regulationcited above describes how biotechnology products require phar-macokinetics, pharmacodynamics, safety, and efficacy clinicaltrials. Kikuzubam complied with all these requirements.

The safety and scientific requirements endorsed in the Emer-gency Mexican Official Standard NOM-EM-001-SSA1-2012 forbiotechnology drugs and biopharmaceuticals published in Sep-tember 2012 include Good Manufacturing Practices compliance,technical and scientific safety, efficacy and quality proof, labeling,and requirements for biocomparability studies and active phar-macovigilance [3].

Mexican health authorities’ initiatives confirm that the newbiosimilar registration pathway has been designed to align theMexican regulatory framework with International Conference ofHarmonization requirements, instead of just pretending toincrease access to biosimilars as Sandorff et al. claimed.

Finally, it is a known fact that biotechnology generic drugs (asthey were called before the terms “biosimilars” or “biocompar-ables” were coined) have been in the Mexican health system forover 20 years.

Probiomed has developed such products since 1996; theportfolio includes recombinant proteins, cytokines, hormones,recombinant vaccines, monoclonal antibodies, and fusion pro-teins. In approximately 15 years, Probiomed has supplied morethan 76 million doses of biotechnology products to healthinstitutions, health care professionals, and patients in Mexicoand several countries, without receiving any safety or efficacycomplains.

More than 15 million patients have been exposed to Probio-med’s biotechnology generic drugs as detailed in Table 1.

We appreciate and thank you for giving us the opportunity tomake these clarifications.

Jorge Revilla Beltri, MDMedical Director, Probiomed, Mexico D.F., Mexico



R E F E R E N C E S

[1] Milenio. Medicamento biotecnológico del ISSSTE desata pleito legal.Available from: http://jalisco.milenio.com/cdb/doc/impreso/9043982.[Accessed April 24, 2013].

[2] DECRETO por el que se reforman y adicionan diversas disposiciones delReglamento de Insumos para la Salud. Available from: http://dof.gob.mx/nota_detalle.php?codigo=5214882&fecha=19/10/2011. [AccessedApril 24, 2013].

[3] Norma Oficial Mexicana Emergente NOM-EM-001-SSA1-2012,Medicamentos biotecnológicos y sus biofármacos. Buenas prácticas defabricación. Características técnicas y científicas que deben cumpliréstos para demostrar su seguridad, eficacia y calidad. Etiquetado.Requisitos para realizar los estudios de biocomparabilidad yfarmacovigilancia. Available from: http://dof.gob.mx/nota_detalle.php?codigo=5269530&fecha=20/09/2012. [Accessed April 24, 2013].


dx.doi.org/10.1016/j.vhri.2013.05.007

dx.doi.org/10.1016/j.vhri.2013.05.007

dx.doi.org/10.1016/j.vhri.2013.05.007


http://jalisco.milenio.com/cdb/doc/impreso/9043982


http://dof.gob.mx/nota_detalle.php?codigo=5214882&fecha=19/10/2011











dx.doi.org/10.1016/j.vhri.2013.05.007

Table 1 – Patients treated with biosimilars manufactured by Probiomed.

Probiomed biosimilarname

A P I Original brand/laboratory

Approvalyear

Units distributed per year Total units Treatedpatients†

2010* 2011 2012 2013

Gramal Molgramostim Leucomax/Sch-P 1996 157,633 3,652 227 3,654 165,166 23,595Proquiferón Alfa 2a IFN Roferón/Roche 1996 169,492 – – – 169,492 1,412Urifrón Alfa 2b IFN Intrón-A/Sch-P 1997 977,341 170,035 183,316 49,745 1,380,437 11,504Bioyetin Eritropoyetin

(rHu-EPO)Recormón/Roche 1998 12,764,557 2,198,892 2,317,365 4,887,978 22,168,792 213,161

Probivac pediatric HBv Vaccine Engerix-B/Glaxo 2000 4,532,144 3,875,000 3,927,069 7,500,000 19,834,213 6,611,404Probivac adult HBv Vaccine Engerix-B/Glaxo 2000 17,030,081 4,969,950 1,296,456 2,500,000 25,796,487 8,598,829Filatil Filgrastim Neupogen/Roche 2001 436,274 79,072 84,133 317,639 917,118 131,017Protophin Somatropin Humatrope/Lilly 2001 130,294 47,424 44,016 18,000 239,734 7,882Glinux Insulin Humulin/Lilly 2001 3,709,686 224,231 55,330 23,400 4,012,647 131,923Uribeta Beta 1b IFN 8 MUI Betaferon/Bayer 2002 360,720 106,083 121,742 154,669 743,214 4,764Emaxem Beta 1a IFN 12

MUIRebif/Merck-Se 2004 188,442 13,922 14,544 16,391 233,299 1,496

Jumtab Beta 1a IFN 6 MUI Avonex/Bayer 2006 111,244 20,883 25,852 13,575 171,554 3,299Kikuzubam Rituximab Mabthera/Roche 2010 – 12,685 40,593 63,461 116,739 7,708Infinitam Etanercept Enbrel/W-Pfizer 2012 – – – 248,000 248,000 2,583

Total biosimilarunits

76,166,892

Total patientstreated

15,750,577

Note. For more than 15 years, Probiomed has offered more than 76 million doses of biosimilar products for the treatment of more than 15 million patients in Mexico and in the global market.Probiomed has never received a report of immunogenicity or adverse events on its biosimilars, different from those reported by the original product. All Probiomed products are under strictpharmacovigilance programs.� From year of registration up to 2010.† The number of patients was calculated according to current posological schemes.

VALUE

INH

EALTH

REGIO

NAL

ISSUES

2(2

013)331–332

332



Response to Letter from Dr. Jorge Revilla Beltri dated 22nd March, 2013

The letter from Dr. Revilla Beltri, a Medical Director at Probiomed,raises two points.

First, that we cited a newspaper article interview with asource who appears to have been an industry lawyer. On thisbasis, we withdraw the paragraph in question.

Second, Dr. Revilla Beltri states that there are inaccuracies inthe published article with regard to the Mexican regulatoryframework surrounding biosimilars. We stand by our assertionin the article that a new regulatory pathway had only beenrecently created in Mexico to reflect the increasing number ofbiosimilars coming into that market and at the time the articlewas submitted for publication, that regulatory pathway was stillevolving. In fact, the October 2011 Decreto referenced inDr. Revilla Beltri’s letter specifies that biosimilar regulation in

Mexico should be addressed by a subcommittee on a case-by-case basis but did not specify the regulatory structure thatwould surround such subcommittees. The second regulationquoted by Dr. Revilla Beltri was enacted after our article wasaccepted for publication.

Erik Sandorff, MA, MBAPriceSpective, Blue Bell, PA, USA

Copyright & 2013, International Society for Pharmacoeconomicsand Outcomes Research (ISPOR). Published by Elsevier Inc.


V A L U E I N H E A L T H R E G I O N A L I S S U E S 2 ( 2 0 1 3 ) 3 3 3

dx.doi.org/10.1016/j.vhri.2013.05.005

dx.doi.org/10.1016/j.vhri.2013.05.005

dx.doi.org/10.1016/j.vhri.2013.05.005

dx.doi.org/10.1016/j.vhri.2013.05.005

GUIDE FOR AUTHORSValue in Health Regional Issues is a peer-reviewed publication of the International Society for Pharmacoeconomics and Outcomes Research, focusing on Asia, Latin American, Central & Eastern Europe, Western Asia, and Africa. Its mission is to provide a forum for the advancement and dis-semination of knowledge and research in pharmacoeconomics and the healthrelated outcomes of disease and treatment processes, and there-fore solicits original contributions in health care policty analysis, outcomes research (clinical, economic, and patient-reported) empirical studies, methodological studies, and manuscripts on health care resources in the region. Health care decision-maker commentaries from health outcomes researchers and policy makers are welcome. Commentaries are expected to include discussion on how researchers can better respond to the needs of those making clinical and fi nancial decisions in health care.

Value in Health Regional Issues does not consider papers reporting data series or data sets that do not include appropriate statistical confi -dence intervals and/or other measures of statistical imprecision. Value in Health Regional Issues also does not consider papers reporting mod-eling results that do not include sensitivity analysis of key and infl uential model parameters.

Authors for whom English is a second language that would like to submit a manuscript in English may choose to have their manuscript professionally edited before submisstion or during the review process. Authors wishing to employ a professional English-language editing service (English manuscript submissions only) should make contact and arrange payment with the editing service of their choice. For more details regarding the recommended services, see http://support. elsevier.com/.

Value in Health Regional Issues articles may be published in a language of the region other than English (i.e. Chinese, Korean, Japanese, Thai, Indonesian, etc. for Asia; Spanish, Portuguese, etc. for Latin America; and Russian, Polish, Arabic, etc. for Central & Eastern Europe, Western Asia and African region). All manuscripts accepted for publication in Value in Health Regional Issues must have an abstract and keywords in English as well as the language of the manuscript (if not English). Languages for a specifi c issue are decided by the Value in Health Regional Issues Editorial Board.

The criteria for a manuscript to be considered for Value in Health Regional Issues are as follows:

• First author must reside in the region to which a specifi c regional issue is dedicated; and

• The empirical study manuscript submitted to Value in Health Regional Issues must include subjects from population(s) in the region to which a specifi c regional issue is dedicated.

For further details concerning the types of articles and eligible regions suitable for Value in Health Regional Issues, see: http://www.ispor.org/ publications/VIHRI/VIHRImain.asp.

I. ETHICS IN PUBLISHINGFor information on Ethics in Publishing and Ethical guidelines for journalpublication see http://www.elsevier.com/publishingethics and http:// www.elsevier.com/ethicalguidelines.II. CONFLICT OF INTERESTAll authors must disclose any fi nancial and personal relationships with other people or organisations that could inappropriately infl uence (bias) their work. Examples of potential confl icts of interest include employ-ment, consultancies, stock ownership, honoraria, paid expert testimony, patent applications/registrations, and grants or other funding. See also http://www.elsevier.com/confl ictsofi nterest.

III. SUBMISSION DECLARATIONSubmission of an article implies that the work described has not been published previously (except in the form of an abstract or as part of a published lecture or academic thesis), that it is not under consideration for publication elsewhere, that its publication is approved by all authors and tacitly or explicitly by the responsible authorities where the work was carried out, and that, if accepted, it will not be published elsewhere including electronically in the same form, in English or in any other lan-guage, without the written consent of the copyright-holder.

IV. RETAINED AUTHOR RIGHTSAs an author you (or your employer or institution) retain certain rights; for details you are referred to: http://www.elsevier.com/authorsrights.

V. FUNDING BODY AGREEMENTS AND POLICIESElsevier has established agreements and developed policies to allow authors whose articles appear in journals published by Elsevier, to comply

with potential manuscript archiving requirements as specifi ed as condi-tions of their grant awards. To learn more about existing agreements and policies please visit http://www.elsevier.com/fundingbodies.

VI. MANUSCRIPT SUBMISSION AND SPECIFICATIONSSubmissions received eight months prior to the publication date will be considered by the upcoming issue. Publication timeline for a specifi c regional issue is below:

• Value in Health Regional Issues for Asia is published in May.• Value in Health Regional Issues for Central & Eastern Europe, Western

Asia and Africa is published in August.• Value in Health Regional Issues for Latin America is published in

December.To submit a manuscript to Value in Health Regional Issues, please

go to: http://www.ispor.org/publications/VIHRI/submission_instruction.asp. For assistance, authors may contact the Value in Health Regional Issues editorial offi ce at: [email protected].

If submissions are larger than 500 KB, they should be compressed using PKZIP or WINZIP.

Authors will be required to assign copyright of their papers. Copyright assignment is a condition of publication and papers will not be passed to the publisher for production unless copyright has been assigned. An appropriate copyright assignment form can be found at the following address: http://www.ispor.org/publications/VIHRI/submission_instruc-tion.asp. A faxed copy of this completed and signed form is acceptable; fax to 609-586-4982 or email to: [email protected].

If excerpts from other copyrighted works are included, the author(s) must obtain written permission from the copyright owners and credit the source(s) in the article. Elsevier has preprinted forms for use by authors in these cases: please consult: http://www.elsevier.com/permissions.

Each Submission should contain separate documents as follows:i. COVER LETTER. The cover letter should include: 1) title of the manu-script; 2) name of the document fi le(s) containing the manuscript and the software (and version) used; 3) name and all contact information for the corresponding author and a statement as to whether the data, models, or methodology used in the research are proprietary; 4) names of all sponsors of the research and a statement of all direct or indirect fi nancial relationships the authors have with the sponsors; and 5) if applicable, a statement that the publication of study results was not contingent on the sponsor’s approval or censorship of the manuscript.For a sample cover letter, see:http://www.ispor.org/publications/VIHRI/SampleCoverLetter.doc.ii. TITLE PAGE. The title page should contain the following: 1) title; 2) full names (fi rst and surname) of all authors including academic degrees and affi liation(s); 3) name, mailing and email addresses, telephone and fax numbers of corresponding author (with whom all correspondence will take place unless other arrangements are made); 4) all sources of fi nancial or other support for the manuscript (if no funding was received, this should be noted on the title page); 5) at least four key words for indexing purposes; and 6) a running title of not more than 45 characters including spaces.For a sample title page, see:http://www.ispor.org/publications/VIHRI/SampleTitlePg.doc.iii. MANUSCRIPTS. Manuscripts must be typed in either Microsoft Word (Version 5.0 or later) or WordPerfect (version 5.1 or later). Manu-scripts should be in 8.5x11-inch page format, double-spaced with 1-inch margins on all sides and size 10 font (Arial or Times New Roman fonts are preferred). Minimal formatting should be used, i.e., no justifi ca-tion, italics, bold, indenting, etc. There should be no hard returns at the end of lines. Double-spacing after each element is requested (e.g., headings, titles, paragraphs, legends). The ‘Uniform Requirements for Manuscripts Submitted to Biomedical Journals’ should be consulted for specifi c style issues not addressed here (www.acponline.org, Ann Intern Med 1997;126:36-47). The word count should be less than 4500.

a. ABSTRACT. An abstract of 250 words or less is required, summarizing the work reported in the manuscript. Original research manuscripts should use a structured format for the abstract, i.e., Objectives, Methods, Re-sults, and Conclusions. A non-English research manuscript requires an abstract in English as well as in the language of the manuscript.b. TEXT. The body of the manuscript should be divided into sec-tions that facilitate reading and comprehension of the material. This should normally include sections with the major headings: Introduc-tion, Methods, Results, Conclusions, Acknowledgments (if needed), and References. There should be no footnotes. Figures (inclusive of fi gure legends) and Tables must be submitted each as separate documents.

GUIDE FOR AUTHORS – continuedc. REFERENCES. References should be listed in a separate section and numbered consecutively with Arabic numerals in the order in which they are cited in the text. Citing unpublished or non-peer-reviewed work such as abstracts and presented papers is discouraged. Personal communications may be indicated in the text as long as written ack-nowledgment from the authors of the communications accompanies the manuscript. Reference style should follow that of Index Medicus. Spell out single-word journals and abbreviate all others according to the style of Index Medicus. If there are more than four authors, use only the names of the fi rst three, followed by et al.

The three most common types of references are illustrated below for example.Journal article: Surname and initials of author(s), title of article, name of journal, year, volume number, fi rst and last page.

Arocho R, McMillan CA. Discriminant and criterion evaluation of the U.S.-Spanish version of the SF-36 Health Survey in a Cuban-American population with benign hyperplasia. Med Care 1998;36:766–72.Book: Surname and initials of author(s)/editor(s), title and subtitle, volume, edition (other than fi rst), city, publisher, year.

Johnston J. Econometric Methods (3rd ed.). New York: McGraw-Hill,1984.Chapter in Book: Surname and initials of author(s), title of chapter,author(s)/editor(s) of book, title of book, volume, edition (other thanfi rst), city, publisher, year.

Luce BR, Manning WG, Siegel JE, et al. Estimating costs in cost-effectiveness analysis. In: Gold MR, Siegel JE, Russell LB, et al., eds., Cost-effectiveness in Health and Medicine. New York: Oxford University Press, 1996.

Website: Title of article on web. www.document. Available from: http://www... [Accessed Month day, year].

International Society for Pharmacoeconomics and Outcomes Re-search (ISPOR) Good Outcomes Research Practices Index. Avail-able from: http://www.ispor.org/workpaper/practices_index-.asp. [Acessed January 1, 2011].For a sample manuscript, see:http://www.ispor.org/publications/VIHRI/SampleManuscript.pdf.

iv. TABLES. Tables should be clearly labeled, neatly typed, and easy to understand without reference to the text. Each should be doublespaced and presented on a separate page. Statistical estimates should indicate parameter estimates and, as appropriate, t ratios or standard error, statistical signifi cance, sample size, and other relevant information. All abbreviations must be explained below each table. Each table should be numbered and have a self-explanatory title.For a sample table, see:http://www.ispor.org/publications/VIHRI/SampleTbls.doc.v. FIGURES. Figures should each be submitted as a separate image fi le, not embedded in the manuscript document or in a slide presentation. Cite fi gures consecutively, as they appear in the text, with Arabic num-bers (Figure 1, Figure 2, Figure 3A, etc.). If, together with your accepted article, you submit usable color fi gures then the Journal will ensure, at no additional charge, that these fi gures will appear in color on the Web (e.g., ScienceDirect and other sites) regardless of whether or not these illustrations are reproduced in color in the printed version. There will be a charge for color reproduction in print; you will receive information regard-ing the costs from Elsevier after receipt of your accepted article. Please indicate your preference for color in print or on the Web only. Each fi gure must be assigned a brief title (as few words as possible, and reserving ab-breviations for the legend) as well as a legend. The corresponding legend should be typed double-spaced on a separate page. All symbols, arrows, and abbreviations must be explained in the legend. Please submit fi les with a resolution of at least 300 DPI. Line artwork should contain a resolu-tion of least 1000 DPI. Elsevier recommends submitting fi gures in the following formats: TIFF, JPG, EPS, and PDF. Please be sure to delete any identifying patient information such as name, social security number, etc. Photographs in which a person’s face is recognizable must be accompa-

nied by a letter of release from that person explicitly granting permission for publication in the Journal. For any previously published material, writ-ten permission for both print and electronic reprint rights must be ob-tained from the copyright holder. For further explanation and examples of artwork preparation, see Elsevier’s Author Artwork Instructions at www.elsevier.com/artwork.For a sample fi gure, see:http://www.ispor.org/publications/VIHRI/SampleFigure.doc.vi. SUPPLEMENTARY MATERIAL OR SUPPLEMENTAL DATA. You may submit appendixes that describe either methods or results in more detail if these are needed for clarity of understanding by either peer reviewers or readers. If appropriate, materials suitable for Web publica-tion but not print publication (eg, audio or video fi les, see below) can also be submitted. If you do so, indicate the particular reasons for the appendix and whether you are submitting it for possible Web publication or simply for peer review purposes. Value in Health Regional Issues ac-cepts audio and video fi les as ancillaries to the main article. Audio fi les should be in .mp3 format; the recommended upper limit for the size of a single fi le is 10 Mb. Video fi les should be submitted in .mpg or .mp4 format, the recommended upper limit for the size of a single fi le is 10 Mb. Any alternative format supplied may be subject to conversion (if technically possible) prior to online publication.vii. SURVEY INSTRUMENT. For papers analyzing preferences, Value in Health Regional Issues requires the submission of a copy of the sur-vey instrument used to generate the preference data. This is to help in the review process and the survey instrument need not appear in a fi nal publication. If the authors wish the questionnaire to be published with the paper, it should be submitted as part of the paper. If they do not wish it to be published, it should be submitted as Supporting Infor-mation and then will be sent to the reviewers as a reviewer’s appendix.

VII. DATA, MODELS, AND METHODOLOGYAll authors must agree to make their data available at the Editor’s request for examination and re-analysis by referees or other persons designated by the Editor. All models and methodologies must be pre-sented in suffi cient detail to be fully comprehensible to readers.

VIII. AUTHOR ANONYMITYIt is the policy of Value in Health Regional Issues that peer review of submitted manuscripts is double blinded, i.e., the reviewers do not knowthe names of the authors of manuscripts and the authors do not know the names of the reviewers. Blinded reviews are common practice at many important scientifi c and medical journals.

IX. THE REVIEW PROCESSAll manuscripts deemed appropriate for Value in Health Regional Issues after initial screening will be reviewed by at least two peer reviewers. The objective of the journal is to complete peer review and reach editorial decision within ten to twelve weeks of submission, at which time the corresponding author will receive written notifi cation, including anonymous reviewer commentary.

X. AUTHOR TRACKING SERVICESAuthors may track accepted manuscripts (English only) at:http://www.elsevier.com/trackarticle and set up e-mail alerts to inform them when an article’s status has changed. Contact details for questions arising after acceptance of an manuscript, especially those relating to proofs, will be provided by the publisher. For manuscripts not submitted in English, authors may query the Value in Health Regional Issues Editorial offi ce at: [email protected], [email protected], or [email protected].

XI. PROOFSProofs are to be sent electronically to Authors to review for printer’s errors. Substantive changes or additions to the edited manuscript cannot be allowed at this stage. Corrected proofs should be returned to the publisher within 2 days of receipt.

XII. OFFPRINTSThe corresponding author, at no cost, will be provided with a PDF fi le ofthe article via e-mail. For an extra charge, paper offprints can be orderedvia the offprint order form which is sent once the manuscript is acceptedfor publication. The PDF fi le is a watermarked version of the publishedarticle and includes a cover sheet with the journal cover image and adisclaimer outlining the terms and conditions of use.