on

Ra

oli &

cou

talit

s for

el n

n

ous

dur

e of

ally

a s

this

literature on epidemics control and logistics operations aiming at stimulating further interest in the

area of epidemics control supply chain management.

& 2012 Elsevier B.V. All rights reserved.

1. Introduction

Among disasters, outbreaks of epidemical, inclIV arey devpublicd of timics incause

ndemicedicto

of the population affected (Watson et al., 2007). Finally, deliberate

matevirusemic

nessed rapid urbanization of the worlds population along with a

Contents lists available at SciVerse ScienceDirect

w.e

Int. J. Productio

Int. J. Production Economics 139 (2012) 393410density could potentially lead to a pandemic of unprecedentedpappis@unipi.gr (C.P. Pappis), nraxan@unipi.gr (N.P. Rachaniotis).substantial growth in general population could lead to acceler-ated epidemic outbreaks, rendering socioeconomic systems evenmore vulnerable. A possible outbreak combined with changes indemographic conditions like population distribution, size and

0925-5273/$ - see front matter & 2012 Elsevier B.V. All rights reserved.

http://dx.doi.org/10.1016/j.ijpe.2012.05.023

n Corresponding author. Tel.: 30 2104142150; fax: 30 2104142629.E-mail addresses: dasaklis@unipi.gr (T.K. Dasaklis),prevalent infectious diseases after natural disasters and all of themare closely related to unsanitary health conditions and malnutrition

range and this could be attributed to changes in climate patterns(Gould and Higgs, 2009). Apart from climate change, the wit-int/csr/don/2010_05_28/en/index.html). Epidemic outbreaks arealso very common in the aftermath of natural disasters. Acuterespiratory infections, measles, malaria and diarrhea are the most

yellow fever, dengue and cholera are re-emerging due to clichange among other factors (Shope, 1991). Specic arbodiseases have recently emerged outside their usual end2005). Epidemic outbreaks may occur in the context of naturalcauses such as the recent outbreak of novel inuenza A(H1N1) viruswhere, according to theWorld Health Organization (WHO), from thestart of the pandemic until May 28, 2010, the virus had alreadyspread over 214 countries causing 18,114 deaths (http://www.who.

have arisen and certain drivers like climate change, populationdensity and urbanization could serve as catalysts for the accel-eration of pandemic incidents. Climate change is expected to playa crucial role in the birth and transmission of specic diseases(McMichael, 2003). Many studies suggest that diseases such asdamages of human and material capitdeaths. Polio, smallpox, cholera and Hcontinue to pose a threat for mancountries. Human history is full ofpandemics occurred in a certain periodeath. For example, plague epidemcharacterized by high mortality ratesthe 19181919 Spanish inuenza pato 50 million people worldwide (Bens account for excessiveuding a great number ofamong the diseases thateloping and developedhealth incidents wheree causing suffering andlate Medieval Europe

d many fatalities, whilekilled an estimated 20

w, 1987; Tumpey et al.,

bioterrorist actions and the release of biological warfare agentscould also lead to epidemic outbreaks. According to Henderson(1999) smallpox and anthrax are considered to be among the twomost feared biological agents that could be used in a probablebioterrorist attack as they have the potential to be grown reason-ably easily and in large quantities and are sturdy organisms that areresistant to destruction. The anthrax attacks of 2001 in the UnitedStates demonstrated the threat of a possible bioterrorist action andits severe impacts.

The importance of addressing epidemic outbreaks nowadays iseven greater as the general framework in which they may occurhas dramatically changed during the last years. New challengesEpidemics control and logistics operati

Thomas K. Dasaklis n, Costas P. Pappis, Nikolaos P.

University of Piraeus, Department of Industrial Management and Technology, 80 Kara

a r t i c l e i n f o

Article history:

Received 15 September 2010

Accepted 17 May 2012Available online 28 May 2012

Keywords:

Epidemics control

Emergency supply chains

Epidemics control logistics

a b s t r a c t

Outbreaks of epidemics ac

also a major cause of mor

an epidemic outbreak call

and well-trained personn

available information and

uncontrollable or disastr

emergency supply chain

focuses on dening the rol

epidemic outbreaks, critic

of the selected literature

proposed. In conclusion,

journal homepage: wws: A review

chaniotis

Dimitriou Str., 18534 Piraeus, Greece

nt for a great number of deaths. Communicable or infectious diseases are

y in the aftermath of natural or man-made disasters. Effective control of

a rapid response. Available resources such as essential medical supplies

eed to be deployed rapidly and to be managed in conjunction with

ancial resources in order to contain the epidemic before it reaches

proportions. Therefore, the establishment and management of an

ing the containment effort are of paramount importance. This paper

logistics operations and their management that may assist the control of

reviewing existing literature and pinpointing gaps. Through the analysis

eries of insights are derived and several future research directions are

paper provides both academics and practitioners with an overview of

lsevier.com/locate/ijpe

n Economics

SiddharthaHighlight

SiddharthaHighlight

SiddharthaHighlight

supply chain and enhance its understanding. To look for trends in epidemics control supply chain manage-

ma

levreding

dem

T.K. Dasaklis et al. / Int. J. Production Economics 139 (2012) 393410394proportion where available capacities and resources could bestrained to their limits.

The control of an epidemics outbreak calls for a promptresponse. Certain control protocols should be followed and hugeamounts of supplies together with the necessary humanresources (medical and other personnel) should be available inorder to be utilized during the containment effort. For example, ifa smallpox attack happens, vaccination of the affected populationshould take place within 4 days while in the case of an anthraxoutbreak the distribution of antibiotics should take place within 2days of the event (Lee, 2008). As a consequence, any control of anepidemics outbreak should rely on the establishment of anemergency supply chain as a plethora of logistics issues is raisedaccording to the control strategy adopted and the very nature ofthe agent triggering the outbreak. All the logistics operations suchas transportation of medical supplies and commodities or thedeployment of medical personnel must be managed in conjunc-tion with available information and nancial resources in order tocontain the epidemic before it reaches critical proportions. This isthe reason why leading international health organizations likeWorld Health Organization and the Pan American Health Organi-zation explicitly recognize the importance of logistics operationsto any successful health task undertaken for the control of anoutbreak. A basic component of the World Health OrganizationsEpidemic and Pandemic Alert and Response program addresseslogistic issues in order to provide operational assistance in theongoing management of logistics required for epidemic and pandemic

preparedness and response and for the rapid deployment of medical

and laboratory supplies, transport, communications as well as the

rapid deployment of outbreak response teams (http://www.afro.who.int/en/divisions-a-programmes/ddc/epidemic-a-pandemic-alert-and-response/programme-components/logistics.html). Inter-national aid organizations, like the United States Agency forInternational Development (USAID), also pay attention to logis-tical issues in the case of epidemic outbreaks control. TheUSAID9DELIVER PROJECT, Task Order 2-Supply Chain Manage-ment for Outbreak Response supports USAIDs efforts to mitigateexisting and emerging pandemic threats by procuring, stock-piling, and distributing outbreak response commodities (http://deliver.jsi.com/dhome/topics/health/outbreakresponse).

Although logistics operations are very important for control-ling an epidemics outbreak, the scientic community has yet toproduce a large amount of well-established approaches thatexplicitly incorporate epidemics logistics features. Even further,issues of appropriately managing epidemics logistics operationshave been paid limited attention. So far, epidemics logisticsoperations have been seen through the lens of resource allocationor have been implicitly incorporated as qualitative variables intopreparedness and response plans. For example, a considerablevolume of scientic research has been conducted in the case ofresources allocation for the control of infectious diseases(Brandeau, 2005; Rachaniotis et al., 2012; Zaric and Brandeau,2001,2002) or in the case of developing preparedness andresponse plans for health-care facilities to deal with epidemicoutbreaks (Ammon et al., 2007; Ippolito et al., 2006; Rebmannet al., 2007; Savoia et al., 2009; Webby and Webster, 2003). Whileresource allocation models and contingency plans provide astrong insight towards epidemics control, they often overlooksome critical aspects: allocation of resources cannot be accom-plished unless the availability of these resources is assured at theright time, right place and right quantity. In addition, contingencyplans often overlook critical logistical parameters like patientow logistics and the availability of workforce and, therefore,several decision variables related to possible bottlenecks orworkforce shortages are omitted. Results of many full-scale

exercises have shown that logistics poses a real challenge whenthat will ensure the provision of medical supplies like antiretro-viral drugs, antibiotics, clean water/adequate sanitation andbetter nutrition conditions in order that the multiplication ofthe infectious agent be reduced. Control measures could beadopted with the aim either to prevent the spread after theinitiation of an infectious disease (as pre-event measures) or tocontrol a conrmed outbreak (post-event measures). In the rstcase, a certain level of medical supplies should be kept in order tobe utilized immediately at the initiation of an epidemic. In thesecond case the deployment of all the available resources shouldrapidly take place providing either treatment to those alreadyinfected or prophylaxis to those susceptible to the agent triggeringovethein objectives of any containment effort. Such control measuresand the launching of vaccination or quarantine programsr certain geographic regions. They also call for interventionsmapted at international, national, provincial or even communityel. Reducing the rate by which susceptibles become infected,ucing the mortality rate for those already infected and increas-the immunization capacity of the population comprise theadoEpidemics control and emergency supply chainnagement

The control of infectious diseases may be based on measures2.ment academic research and to propose areas for furtherinvestigation.

The remainder of the paper is structured as follows: in Section 2some key concepts of emergency and humanitarian supply chainmanagement are outlined. Section 3 provides an overview of theresearch methodology utilized for the purposes of the review andthe boundaries of our research. In Section 4 an inventory of thelogistics operations taking place during the control of an epidemicis developed. Section 5 deals with the analysis and classication ofthe selected literature regarding epidemics logistics operations. InSection 6 the main ndings of the review are discussed. Finally, thepaper ends with some concluding remarks and suggestions forfuture research directions.controlling an epidemic outbreak (Aaby et al., 2006). Logisticalbarriers could also be apparent even when establishing a simplequarantine program in terms of transferring supplies acrossquarantine lines, recruiting qualied medical personnel etc.(Barbera et al., 2001).

The aforementioned gap in the literature concerning epidemicslogistics operations and their management has been our main driverfor conducting this review paper. Our scope is to shed light into thegeneral context of emergency supply chain management in the caseof epidemics containment. Thus the objectives of the paper may besummarized as follows:

To dene and inter-correlate the logistics operations takingplace during the containment of an epidemic.

To nd to what extent literature has produced a critical massof scientic work in terms of methodologies applied andresearch techniques utilized regarding the issues of epidemicscontrol logistics.

To determine whether the methodologies applied in businesslogistics problems could be utilized in the context of epidemicscontrol logistics (e.g. Operational Research methods utilizationetc.).

To develop a more robust denition of the epidemics controloutbreak.

SiddharthaHighlight

SiddharthaHighlight

SiddharthaHighlight

SiddharthaHighlight

SiddharthaHighlight

SiddharthaHighlight

SiddharthaHighlight

SiddharthaHighlight

SiddharthaHighlight

SiddharthaHighlight

It is therefore evident that the control of any epidemic out-break should be based on the establishment of an emergencysupply chain as physical movement of large amounts of medicalsupplies takes place. For example, manufacturers should producevaccines, antiretroviral drugs and complementary medical sup-plies. Governments and public health institutions should pur-chase and stockpile well in advance a plethora of such supplies fora possible outbreak. During the containment effort transportationand distribution of these supplies from central warehouses toregional store sites and then to local Points of Dispensing (POD)will have to take place. In the case of vaccines, a cold supply chainmust be established to assure that vaccines are transported,stored and packed in accordance with manufacturers instructions(stable temperature). Affected people will proceed to treatment

the point of origin to the point of consumption for the purpose ofalleviating the suffering of vulnerable people. As an example, oneof the notable aspects of the relief effort following the 2004 Asiantsunami was the public acknowledgment of the role of logistics ineffective relief (Thomas and Kopczak, 2007).

The management of emergency and/or humanitarian supplychains has recently attracted the attention of both practitionersand researchers. Such supply chains have much in common withcommercial supply chains but at the same time they posesignicant challenges as they operate under uncertain, and manytimes, chaotic conditions. Research methodologies widely utilizedin solving business logistics problems could be adopted in thecontext of emergency supply chain operations. Similarities exist-ing between commercial and emergency supply chains offer the

nitarian supply chain operations has been targeted towards

T.K. Dasaklis et al. / Int. J. Production Economics 139 (2012) 393410 395centers where patient ow operations along with dispensingactivities of the medical supplies should be managed appropri-ately. Reverse logistics activities should also take place as dan-gerous wastes must be treated carefully or disposed of in such away that they do not pose a threat for the medical personnel andpeople engaged in the containment effort. In addition, coordina-tion issues across the entire emergency supply chain arise.Manufacturers, governments, primary health care institutes andpossibly military agencies are few among many players thatshould be coordinated during the control effort. Finally, managingthe information regarding the demand for medical supplies aswell as the ow of funds is also critical. In Fig. 1 an End-to-Endapproach of the epidemics control supply chain is depicted. Notethat the reverse ow may refer to only a part of the medicalsupplies.

The outbreak of an epidemic can be related to a high rate ofmortality, thus it may be characterized as a disaster. DisasterOperations in general are a set of activities that are performedbefore, during and after a disaster with the goal of preventing lossof human life, reducing its impact on the economy and returningto a state of normalcy (Altay and Green Iii, 2006). Disastermanagement is an applied science which seeks, by the systematicobservation and analysis of disasters, to improve measures relating to

prevention, mitigation, preparedness, emergency response and recov-

ery (Pettit and Beresford, 2005).Recent studies have demonstrated the contribution of appro-

priate management of disasters to save or offer relief to as manypeople as possible. According to Kovacs and Spens (2007) anemergency or humanitarian supply chain encompasses a range ofactivities, including preparedness, planning, procurement, trans-port, warehousing, tracking and tracing and customs clearance.Humanitarian logistics is an umbrella term for a mixed array ofoperations, from disaster relief to continuous support for devel-oping regions, and could be dened as the process of planning,implementing and controlling the efcient, cost effective ow andstorage of goods and materials as well as related information fromFig. 1. Materials ow of the epidemics contrinventory management, facility location, transportation problemsand performance measurement. For example, Beamon andKotleba (2006) and Clay Whybark (2007) examined the problemof inventory management in case of response to emergencies.Balcik and Beamon (2008) examined the facility location problemof a humanitarian relief chain responding to quick-onset disas-ters. Balcik et al. (2008) studied the last mile distributionproblem of the relief chain, which refers to the distribution ofemergency supplies from local distribution centers to nal ben-eciaries affected by disasters. Issues of performance measure-ment in humanitarian supply chains are examined in Beamon andBalcik (2008) and Van Der Laan et al. (2009). Finally, coordinationand knowledge sharing issues or even the role of all the playerstaking part in an emergency and/or humanitarian supply chainhave also been the subject of research (Balcik et al., 2010; Kovacsand Spens, 2010; Pettit and Beresford, 2005).

3. Research methodology and boundaries

The literature regarding epidemics containment is vast. Threebasic streams of research conducted so far may be identied. Therst stream consists of research addressing pharmaceutical inter-ventions for the control of an epidemic. In this case, public healthtreatment programs (like vaccination campaigns) take placewhere limited resources are utilized. These resources could beopportunity of transferring knowledge from the business sector tohumanitarian organizations (Maon et al., 2009). Even at a long-term level, strategies adopted in commercial supply chains couldbe also adopted in the case of emergency and/or humanitariansupply chains in an effort to match supply with demand(Oloruntoba and Gray, 2006; Taylor and Pettit, 2009). A verycomprehensive description towards the issues of humanitarianlogistics can be found in Van Wassenhove (2006).

Research conducted so far towards emergency and/or huma-ol supply chain (End-to-End approach).

SiddharthaHighlight

SiddharthaHighlight

SiddharthaHighlight

SiddharthaHighlight

SiddharthaHighlight

SiddharthaHighlight

SiddharthaHighlight

SiddharthaHighlight

SiddharthaHighlight

SiddharthaHighlight

SiddharthaHighlight

SiddharthaHighlight

initially identied. Our classication scheme was based on athree-level assessment framework presented in Fig. 2. In the rstlevel of assessment the various logistical features incorporated ineach reference are classied according to the time framework inwhich they take place (pre-event or post-event). In the secondlevel of assessment a context-specic classication is made wherethe logistics features are correlated to the nature of the epidemicsoutbreak. Finally, in the third level of assessment the logisticalfeatures and the methodologies applied (qualitative or quantita-tive) for solving the problem tackled are classied. It is worthmentioning that in the classication scheme the incorporatedlogistical features of each reference have been utilized as themain driver for the classication process. Such an approach notonly provides the means for better synthesizing and analyzing theselected literature but also gives us the opportunity to betterunderstand the interdependencies between logistics operationsand the general control context in which they take place. As aconsequence, the classication scheme highlights several attri-butes of epidemics logistics operations and enables us to develop

2001). In the sequence, an inventory of all the logistics operationstaking place during the various phases of an epidemics contain-

T.K. Dasaklis et al. / Int. J. Production Economics 139 (2012) 393410396either discrete (like vaccines, antibiotics or antiretroviral drugs,etc.) or continuous (like funds). The second stream consists ofresearch approaches addressing non-pharmaceutical interven-tions like the closure of schools, voluntary quarantines over awide area, social distancing and travel limitations. Finally, thethird stream consists of research approaches where the pharma-ceutical and non-pharmaceutical interventions are combined andharmonized control actions are proposed.

From the aforementioned literature our focus has been onpharmaceutical and harmonized approaches where certain logis-tics operations features are incorporated into the decision makingprocess (either quantitative or qualitative). Therefore, this reviewcovers the logistical aspects that support the pre-event as well aspost-event logistics operations in the case of emergencies like anunprecedented pandemic outbreak, a bioterrorist attack, epi-demic outbreaks during mass gatherings or epidemic outbreaksin the aftermath of natural or man-made disasters. As a conse-quence, this review does not take into consideration logisticsaspects arising in the context of seasonal inuenza outbreaks orrelevant diseases outbreaks which may be expected, anticipatedor last for a long period of time (like HIV outbreaks).

It is worth mentioning that this review focuses on the opera-tional issues arising in relation to the control of epidemics. Thuspapers addressing problems such as resource allocation or generalcontingency plans have not been included in the main referencelist. For the interested reader an additional reference list can befound after the main reference list, where several resourceallocation papers as well as contingency plan papers are cited.Furthermore, modeling approaches not calibrated to the epi-demics control context have not been included in the classica-tion framework of this review. A relevant survey regardingmodeling approaches in the generic context of emergency supplychain management can be found in Caunhye et al. (2012). Apartfrom the aforementioned limitations, in this review two exemp-tions have been made regarding the papers of Chick et al. (2008)and Ak et al. (in press). The justication of these exemptions isprovided in the relevant section where both papers are cited.Finally, this review cannot by any means be considered as anexhaustive one.

In order to review the available literature our rst task was todene our sources and to establish the appropriate search criteria.Epidemics control handbooks and relevant information fromseveral databases (governmental and international agencies suchas the World Health Organization, the Pan-American HealthOrganization and the European Centre for Disease Preventionand Control etc.) were studied in order to determine the logisticsoperations taking place during the control of an outbreak. Key-word searches were made through several scientic databasesincluding Springer, Science Direct and Scopus. During the selec-tion process all abstracts or list of references from the paperswere examined in order to locate additional relevant researchworks. A total number of 73 papers were nally selected.

We are aware of at least one review regarding OR/MS researchin disaster operations management (Altay and Green Iii, 2006), inwhich epidemics control issues were excluded and one reviewregarding bioterrorism response logistics (Bravata et al., 2004a).In addition, Brandeau et al. (2009) made an inventory of healthsector disaster response models in which logistical constraintswere incorporated in the case of epidemic outbreaks, whileFerguson et al. (2003) reviewed the usage of available mathema-tical models when planning for smallpox outbreaks. The need fora more integrated incorporation of logistical constraints intoexisting modeling approaches was among the ndings of thereview.

For the classication of the selected literature all the logistical

variables and relevant logistics aspects of each reference werement effort is provided. Generally, these phases could be classi-ed as follows (John Hopkins Bloomberg School of Public Healtha more robust research agenda towards epidemics control supplychain management.

4. Epidemics control and logistics operations

Governmental agencies and health institutions should beprepared in advance for the control of epidemic outbreaks. Thismeans that they should have in place robust contingency plansaddressing issues like the availability of emergency medicalstocks and well-trained personnel, their appropriate deployment,the availability of different types of vehicles for the transportationof essential medical supplies and commodities etc. Generally, itremains very difcult to dene whether the needs for producingand distributing vaccines in the case of e.g. a possible pandemicinuenza outbreak can be met (Fedson, 2003; Webby andWebster, 2003) by existing capacities. Consequently, any attemptto contain an epidemic outbreak demands real-time solutionsthat should ensure the effective management of all the logisticsactivities taking place, since sometimes these activities maybecome a real nightmare if not managed properly (Osterholm,Fig. 2. A framework for epidemics logistics literature classication.

SiddharthaHighlight

SiddharthaHighlight

SiddharthaHighlight

SiddharthaHighlight

SiddharthaHighlight

SiddharthaHighlight

SiddharthaHighlight

SiddharthaHighlight

SiddharthaHighlight

SiddharthaHighlight

SiddharthaHighlight

SiddharthaHighlight

SiddharthaHighlight

4.1

bre

ins

ility and scheduling of personnel etc.)

of

bre200

tem200

tiororet

T.K. Dasaklis et al. / Int. J. Production Economics 139 (2012) 393410 3974.2. Outbreak investigation

Outbreak investigation consists of the detection of any sus-

pecAvailability of funds.

designated areas, availab

tems and their capacavailable resources so as to reduce morbidity and mortality whenan epidemic outbreak occurs. This means that pharmaceuticalsand supplies should remain accessible or kept in large quantities(Richards et al., 1999) in order to assist a prompt response, ifnecessary. Procurement of vaccines and medical supplies andtheir exact storing location play a crucial role for the outcome ofany containment effort. For instance, the Strategic National Stock-pile (SNS) program in the United States is an indicative prepared-ness program with the objective to maintain large quantities ofmedicine and medical supplies and to provide these materials tostates and communities within twelve hours in the event of alarge-scale public health emergency (Esbitt, 2003). In addition, acertain amount of vaccines should be available for the immuniza-tion of control teams and health-care workers. This is of greatimportance as medical personnel will treat the very rst infectedpersons and should be protected against the disease that causesthe outbreak. Among the most important logistics operationstaking place and relevant logistics-oriented decisions to be madeduring the phase of preparedness are the following (John HopkinsBloomberg School of Public Health and the InternationalFederation of Red Cross and Red Crescent Societies, 2008; WorldHealth Organization, 2005, 2008):

Identication of sources for the procurement of medicalsupplies and relevant commodities

Contract management for all the materials procured Inventory management for all the essential medical supplies

(vaccines, antibiotics, antiretroviral drugs) and supplementarymedical commodities (personal protective supplies) kept

Periodical review and updating of medical supplies Facility location and capacity determination for stockpiling

centers Network design for transportation/distribution activities and

selection of appropriate means for transportation/distributionactivities

Selection of appropriate vaccination facilities/health care sys-ity (size, availability of rooms andt pandemics, strengthened outbreak communications, greateright on disease spread etc. (World Health Organization, 2009).Epidemic preparedness aims at maintaining a certain level ofpas

light of new developments regarding increased understanding ofdemthemeasures required for the successful containment of an out-ak. The World Health Organization has published several pan-ic preparedness guidelines since 1999 and it updates them inthe

ns range from community to national level and they include allredplaMany organizations around the world have established prepa-ness plans in the case of epidemic or pandemic outbreaks. Such. PreparednessResponseEvaluation.and the International Federation of Red Cross and Red CrescentSocieties, 2008; World Health Organization, 2005):

Preparedness Outbreak investigationted outbreak and its conrmation through laboratory testing.1990; Jansson et al., 2005). It is worth mentioning that thedevelopment of a surveillance system to detect epidemic outbreaksthat occur during emergency situations (like a humanitarian crisis)may necessitate taking into consideration some context-specicfeatures like the target population, the political context, the poorinfrastructure and, nally, the presence of multiple partners in theeld (Mikanatha et al., 2007). Among the logistics activities thatsupport the detection and conrmation mechanisms of a suspectedoutbreak are (U.S. Agency for International Development, 2009;World Health Organization, 2005):

The provision of all the appropriate materials like report sheetsto hospitals, emergency medical services and local publichealth departments that will be used for the collection ofprimary data regarding initial cases

The training of clinical workers to recognize unexpectedpatterns of the occurrence of specic diseases and to promptlyidentify and report suspected cases using standard denitions

The provision of all the necessary commodities and resourcesto the outbreak response team that will facilitate and ensureits operational deployment

The collection of specimens and their labeling The secure transportation of specimens to the appropriate

laboratory (using cold boxes and coolant blocks) The appropriate storage of specimens in the laboratory (kept

within a specic temperature range) The procurement, handling, storing and distribution of labora-

tory commodities, their classication, their quality assuranceand quality control etc.

It is clear that any successful attempt to contain an epidemicoutbreak is closely related to the services provided by labora-tories. These services rely on a huge number of materials andcommodities that laboratories utilize and they necessitateincreased inventory management capabilities. Additionally, dur-ing epidemic outbreaks laboratories must ensure that they havethe capacity to cope with increased testing demands (Crawfordet al., 2010). A good reference regarding laboratory logistics canoutbeearchers have studied relevant issues arising during the detec-n and conrmation of diseases outbreaks attributed to bioter-ist attacks (Bravata et al., 2004b; Buehler et al., 2003; Loberal., 2002; Pavlin et al., 2003; Platt et al., 2003) or epidemicbreaks related to specic agents (Arita et al., 2004; Dietz et al.,ress have been developed by the scientic community (Dato et al.,4; Krause et al., 2007; Lombardo et al., 2003) and many(WDisease Prevention and Control, June, 2009), epidemic out-aks following natural disasters (World Health Organization,5) or even possible disease outbreaks during mass gatheringsorld Health Organization, 2008). Additionally, surveillance sys-for

tems covering cases like pandemic outbreaks (European Centresysanization, 2005).Leading world health organizations have developed surveillanceOrgease and depends on the infectiousness, other determinantstransmission and local endemicity levels (World HealthdisIn order to detect and conrm a suspected outbreak, surveillancesystems must be put in place in order to provide the decisionmakers of the health agencies in charge with the essentialinformation regarding any unexplained infection increases seenover a period of time through the systematic analysis of datacollected. Surveillance systems provide adequate information thatfacilitates the development of an initial response frameworkwhere the type and magnitude of the containment effort couldbe determined once epidemic thresholds have been reached. Theterm epidemic threshold refers to the level of disease abovewhich an urgent response is required. It is specic for eachfound in U.S. Agency for International Development (2009).

SiddharthaHighlight

SiddharthaHighlight

SiddharthaHighlight

In this section the classication and analysis of the selected

Dis

fra

me

the

see

avama

T.K. Dasaklis et al. / Int. J. Production Economics 139 (2012) 393410398During the phase of response laboratory logistics activitiestake place as it is very important for the parties involved to have aclear understanding of how the epidemic evolves over space andtime (rate of spread among subpopulations). This will allow themto proceed to the necessary adjustments or modications of themeasures initially adopted in order for the new measures to becompliant with the data analysis from laboratories tests (WorldHealth Organization, 2005). Other logistics-oriented activitiesnecessitated during this phase may be the safe disposal of bodyfuels or even the handling of dead bodies, the presence of troopsto keep order etc.

4.4. Evaluation

After the epidemic has been contained, decision makers andpublic health policy makers engaged in the control efforts shouldproceed to the evaluation of all the measures undertaken duringthe previous phases. Generally, the evaluation phase is very usefulas it provides strong insights towards a series of modicationsthat need to be made in order to increase the resilience of thecontrol mechanisms in future epidemic outbreaks. Despite thefact that the evaluation phase entails limited physical movement

ofthe management of patients in triage centers (clinical owlogistics).adjustments to the capacity of health care facilities to hospitalizeinfected peoplethe scheduling of available vehicles to be used for transporta-tion and distribution purposes

number of people that can vaccinate per day)the daily/weekly capacity of available personnel to performmass vaccination campaigns (for example the maximumthe establishment of a cold supply chain for the provision ofessential medical supplies like vaccinesthe procurement of supplies/resources once depletedthe dispensing of medical supplies, supplementary materialsand commodities to the publicthe transportation/distribution of supplies and commoditiesfrom central warehouses to local PODthe periodical review and updating of supplies andcommodities needed4.3. Response

Once leading health agencies have conrmed an epidemicoutbreak, measures and control strategies must be implementedas soon as possible at a regional or national level. Treatmentcenters should be established and available resources such asmedical supplies and personnel should be deployed rapidly inorder to contain the epidemic before it reaches uncontrollableproportions. Vaccination of susceptible groups or isolation andquarantine of those infected are considered standard interven-tions for the containment of an epidemic. All measures takenmust be based on a clear understanding of the agents naturetriggering the outbreak as some diseases necessitate speciccontrol protocols to be followed (World Health Organization,2005). This in turn calls for the availability of additional infra-structure and medical supplies within health care premises suchas isolation rooms with good ventilation systems, respiratoryequipment etc. The logistics operations and relevant decisions tobe made during the phase of response to a conrmed outbreakrefer to (John Hopkins Bloomberg School of Public Health and theInternational Federation of Red Cross and Red Crescent Societies,2008; World Health Organization, 2005, 2008):

the selection of facilities to serve as PODsmedical supplies and complementary commodities, it remainsof 73) and they mostly refer to natural outbreaks (15 out of 22)like inuenza epidemics. Integrated approaches covering bothpre-event and post-event control measures are limited. In thecase of epidemics control in the aftermath of natural disasters atotal of 5 articles were retrieved. Some papers address issues ofepidemics control logistics without explicitly specifying the con-37)tex37 out of 73 research studies deal with the deployment of theilable resources once the outbreak has occurred and thejority of them are closely related to bioterrorism (22 out of. Research efforts towards pre-event strategies follow (22 outrature according to the aforementioned thresholds can ben in Table 1.litedistributing and dispensing of these supplies take place onceoutbreak has occurred. The classication of the selectedtheckpiling of medical supplies for a future outbreak is a proactiveasure and takes place well before the outbreak occurs whileStoherings such as the Olympic Games. In addition, the timeme in which control measures apply may also be different.gatpulations are dislocated or subject to a humanitarian crisis.ease outbreaks may also occur during big events and masspoliterature takes place in accordance with the classication schemepresented in Fig. 2.

5.1. Time frame and type of the outbreak

Epidemic control measures may vary depending on the natureof the outbreak and the time horizon in which they are imple-mented. For example, a country may face a bioterrorist attack or anatural inuenza outbreak. Communicable disease outbreaks mayalso occur in the aftermath of natural disasters where certain sub-important from a logistical point of view. Many useful conclu-sions can be drawn with respect to logistics control operationssuch as (AHRQ, 2004; John Hopkins Bloomberg School of PublicHealth and the International Federation of Red Cross and RedCrescent Societies, 2008):

the identication and assessment of possible bottlenecks ordelays that hindered the deployment of the available medicalsupplies

the evaluation of the timeliness that should have beenrespected during the control of the epidemic

the follow-up and monitoring of patients for antibiotic effec-tiveness or vaccine immunoresponse

the identication of patients requiring dose modication oralternative treatment regimen due to adverse effects

the development of indicators regarding the performance ofthe logistics control operations

the assessment of coordination issues risen among the partiesinvolved

the establishment and operation of rehabilitation proceduresin the case of epidemic outbreaks in the aftermath of naturaldisasters.

All the above should lead to clear conclusions and, therefore,recommendations that will enhance the capabilities of the partiesinvolved and will reduce vulnerabilities of the control mechan-isms. Finally, the dissemination of knowledge and the lessonslearned should take place among all the parties involved, frompublic health policy makers and health agencies to localcommunities.

5. Literature classication and analysist of the outbreak. We refer to this type of literature as

SiddharthaHighlight

SiddharthaHighlight

SiddharthaHighlight

SiddharthaHighlight

SiddharthaHighlight

aks

6; A

al.,

Laur

Dhankhar et al., 201

2010; Hashikura an

Lugner and Postma,

mun

adigu es, 2009; Ekici et al., 2008; (Aaby et al., 2006; Lee et al.,

2009a, 2009b; Li and Jie, 2010;

Liu, 2007; Pietz et al., 2009;

Wang et al., 2009)

(Blecken

et al., 2010;

Date et al.,

2011;

Moore et al.,

T.K. Dasaklis et al. / Int. J. Production Economics 139 (2012) 393410 399Siddiqui and Ed

Post-event

(Barbera et al., 2001; Brandeau et al., 2008; Craft

et al., 2005; Giovachino et al., 2005; Herrmann

et al., 2009; Hu and Zhao, 2012, 2011; Hui, 2010;

Hupert et al., 2002; Jingjing et al., 2009; Kaplan

et al., 2003; Ke and Zhao, 2008; Lee, 2008; Liu and

(Collin and de R

Hadler, 2005)Table 1Time-frame and context-specic related literature classication.

Time-frame

Context-specic aspects

Bioterrorist attacks Natural outbre

Pre-event (Berman and Gavious, 2007;Huang et al., 2010; Murali et al., 2012)

(Adu et al., 199

2009; Balicer et

et al., 2005; Demiscellaneous. It is worth mentioning that no paper dealing withepidemics control logistics during mass gatherings has beenidentied. Fig. 3 shows a numerical allocation of the literatureaccording to the time frame of the response and the nature of theoutbreak.

5.2. Logistical attributes and methodologies

In the following sub-sections the literature is classied accord-ing to the logistical attributes embedded within the modelingframeworks and, of course, the methodologies applied. For thisclassication, four clusters have been identied. These refer to theepidemics logistics network conguration, stockpiling of medicalsupplies, triage operations and other approaches. In Fig. 4 anallocation of the selected papers is provided regarding the afore-mentioned clustering, their type and source of publication.

Fig. 3. Time-frame and outbreak-type literature classication.

Zhao, 2009, 2011; Liu et al., 2011; Miller et al.,

2006; Patvivatsiri et al., 2007; Porco et al., 2004;

Richter and Khan, 2009; Wein et al., 2003; Zhao and

Han, 2010)

Integrated (Brandeau et al., 2007; Bravata et al., 2006; Manleyand Bravata, 2009; Shen et al., 2009; Whitworth,

2006; Zaric et al., 2008; Zhao and Sun, 2008)

(Arinaminpathy and

2008; Duintjer Tebb

Jennings et al., 2008Miscellaneous Disasteraftermath

rinaminpathy and McLean,

2005; Chick et al., 2008; Cinti

entis et al., 2008, 2009;

0, 2009; Harrington and Hsu,

d Kizu, 2009; Lee et al., 2006b;

2009; Radonovich et al., 2009;

ds, 2008)

(Adida et al., 2011; Jia et al.,

2007a,2007b; Rebmann et al.,

2011)

5.2

(vaetcstotiowh

Figof p

Mc

ens

).1. Epidemics logistics network conguration

During the containment of an epidemic, medical suppliesccines, antiviral drugs etc.) and commodities (food, water.) held in central storage sites must be distributed to regionalckpile centers and then to local PODs (Fig. 5). The aforemen-ned problem refers to the logistics network design problemere a series of decisions should be made with respect to:

the location, number and capacity of both the stockpile centersand PODs (facilities)the assignment of these facilities to serve certain sub-populationsthe selection of modes of distribution and relevant capacities

. 4. Literature classication according to logistical attributes, type and sourceublication.

1990;

Rottkemper

et al.,

2011, 2012)

Lean, 2008; Conn et al.,

et al., 2010; Hessel, 2009;

(Ak et al., in press; Lee et al.,

2006a)

SiddharthaHighlight

SiddharthaHighlight

SiddharthaHighlight

ntro

T.K. Dasaklis et al. / Int. J. Production Economics 139 (2012) 393410400pro the inventory level of medical supplies and commodities heldas well as the replenishment policies followed.

As far as facility location is concerned, research efforts have beenbased on P-median, P-center and maximal covering location model-ing approaches as well as game theory for the formulation of theproblem. Jia et al. (2007a) propose a series of facility locationapproaches based on the covering model, the P-median model andthe P-center model, each suited for different needs according to thecharacteristics of several large-scale emergencies. A demonstration ofthe P-center and P-median approach is presented (for a possibleanthrax and smallpox bioterrorist attack, respectively). In Jia et al.(2007b) the problem is formulated as a maximal covering problemwith multiple facility quantity-of-coverage and quality-of-coveragerequirements and a genetic algorithm heuristic, a locate-allocateheuristic and a Lagrangean relaxation heuristic are developed forthe solution of the problem. An illustrative example (anthrax attack)and performance analysis for the heuristics are also provided. InBerman and Gavious (2007) the problem is formulated as a gamebetween a terrorist and a state where set-up costs along withpreventive ones are considered. Game theory and mathematicalprogramming are used for the formulation and solution of theproblem. Murali et al. (2012) formulate the problem as a special caseof the maximal covering location problem. Demand uncertainty anddistance-sensitive demand are taken into account by using a lossfunction and a locate-allocate heuristic is used for the solution of the

blem. Huang et al. (2010) use a variation of the p-center problem,

Fig. 5. Epidemics cowhere the model presumes that some of the facilities at a node fail torespond to the nodes demand. Dynamic programming is used fornding the optimal location on a path network whereas an efcientalgorithm for optimal locations on a general network is used,respectively.

In the case of logistics network design the majority of themodeling approaches deals with the estimation of the number ofavailable facilities to be utilized, the inventory to be held, relevantreplenishment policies as well as transportation and distributionactivities for the supplies from central points to regional sites and,consequently, local PODs. Differential equations used for the demandestimation (diseases progression) are coupled with logistics networkmodeling approaches. Hu and Zhao (2012) develop a multi-objectiveprogramming model for the selection of emergency centers and thedetermination of the amount of medicine transported from thereplenishment sources to selected emergency centers and from theselected centers to the points of dispensing. Wang et al. (2009)examine the selection of storage points and distribution activities ofmedical supplies to an affected area by using a multi-objectivestochastic programming model, which is solved by a geneticalgorithm based on Monte Carlo simulation. Hu and Zhao (2011)use a systems dynamics approach to study the dynamics behavior ofthe replenishment sources, receiving warehouses and the dispensingsites in the case of an anthrax attack. Liu and Zhao (2011) develop adynamic optimization model with time-varying demand for thereplenishment and the transportation of medical supplies from localhealth departments to affected areas. A heuristic algorithm is used forsolving the optimization problem. The case of food distribution duringan epidemic outbreak and relevant network design aspects have alsobeen examined (Ekici et al., 2008). Other approaches include logisticsnetwork design frameworks with replenishment and non-replenish-ment sources (Zhao and Sun, 2008) or network design frameworkswhere the objective is total costs (inventory and transportation costs)and response time minimization (Jingjing et al., 2009; Li and Jie,2010).

Research has also been directed towards the issues of epidemicsdistribution network design. As in the case of logistics networkdesign, the demand for the medical supplies to be distributed relieson differential equations describing the progression of the disease. Inmost of the cases medical supplies such as vaccines or antiviral drugsand medical supplements are distributed. Ke and Zhao (2008)examine the problem of emergency materials distribution in ananti-bioterrorism system subject to related logistical constraints byusing dynamic programming theory and simulation techniques. Leeet al. (2009b) develop a simulation model for the distribution of reliefsupplies from central distribution centers to local PODs. The modeltakes into consideration transshipment between the PODs. Liu et al.

l logistics network.(2011) propose a mixed-collaborative distribution model based onthe point-to-point delivery mode and the multiple traveling salesmendelivery system. For the solution of the optimization problem agenetic algorithm is presented. In addition, Zhao and Han (2010)study the problem of distributing medical supplies to affected areastaking into account supply base levels, limited response time-framesand several costs (transportation and shortage costs etc.). Finally,vehicle routing problems in the context of epidemic control logisticshave also been studied (Herrmann et al., 2009; Liu and Zhao, 2009;Shen et al., 2009).

Due to the unpredictable nature of epidemic outbreaks a signi-cant issue that has to be addressed in the aftermath of naturaldisasters is the relocation of inventories. During an ongoing humani-tarian crisis attributed primarily to a natural disaster (earthquake,tsunamis etc.) a disease outbreak may occur triggering a surge ofmedical supplies and necessitating the deployment of additionalcommodities. In this case, available supplies should be relocatedand relevant network design transportation decisions should bemadeenabling the rapid supply of regions coming up with new and urgentdemand (Blecken et al., 2010; Rottkemper et al., 2012, 2011). In

abcHighlight

abcHighlight

abcHighlight

abcHighlight

abcHighlight

abcHighlight

abcHighlight

abcHighlight

Table 2Logistical features and methodologies applied in epidemic logistics network conguration literature.

References Literature characteristics

Logistical attribute Objectives Constraints Problem formulation/methodology applied

Optimization(solution)

Stochastic/deterministic

(Berman and Gavious, 2007) Facility location Cost (or disutility)

minimization

Number of facilities Game theory, mathematical

programming

Deterministic

(Blecken et al., 2010) Inventory relocation,

transportation, transshipment

Cost minimization Demand satisfaction, supply Integer programming Heuristic algorithm Stochastic

(Ekici et al., 2008) Facility location, distribution Cost minimization Capacity, demand satisfaction, supply Mixed integer programming,

simulation

Deterministic

(Herrmann et al., 2009) Vehicle routing, distribution of

medical supplies

Time minimization Capacity, demand satisfaction, supply Integer programming Problem decomposition,

two-stage algorithm

Deterministic

(Hu and Zhao, 2012) Logistics network design Time and cost minimization Capacity, demand satisfaction, supply Multi-objective programming Deterministic

(Hu and Zhao, 2011) Distribution, inventory

management

Inventory level, unsatised

demand

capacity, time, supply System dynamics Deterministic

(Huang et al., 2010) Facility location Minimize the maximum

weighted distance

number of facilities Integer programming Dynamic programming

algorithm

Deterministic

(Jia et al., 2007a) Facility location Demand satisfaction Number of facilities, demand

satisfaction

Integer programming Deterministic

(Jia et al., 2007b) Facility location Demand satisfaction Number of facilities, demand

satisfaction

Integer programming Heuristic algorithms Deterministic

(Jingjing et al., 2009) Logistics network design Cost minimization Capacity, demand satisfaction, supply Mathematical programming Heuristic algorithm Deterministic

(Ke and Zhao, 2008) Distribution of medical supplies Cost minimization Time, demand, type of material Dynamic programming Deterministic

(Lee et al., 2009b) Distribution, dispensing, capacity

planning

Demand satisfaction Capacity Simulation, mathematical

modeling

Deterministic

(Li and Jie, 2010) Logistics network design Cost minimization Supply, demand satisfaction Network equilibrium model Numerical analysis,

approximation methods

Stochastic

(Liu and Zhao, 2009) Vehicle routing Time minimization Number of demand points Multiple traveling salesman

approach

Genetic algorithm Deterministic

(Liu and Zhao, 2011) Logistics network design Cost minimization capacity, demand satisfaction, supply Mathematical programming Heuristic algorithm Deterministic

(Liu et al., 2011) Distribution of medical supplies Distance, time minimization Capacity, demand satisfaction, supply,

distance

Mix integer programming Genetic algorithm Deterministic

(Murali et al., 2012) Facility location Demand satisfaction Capacity, demand satisfaction, supply Integer programming, mix

integer programming

Heuristic algorithm Stochastic

(Rottkemper et al., 2011) Inventory relocation,

transportation

Cost minimization Supply, demand satisfaction, capacity Mixed-integer programming Deterministic

(Rottkemper et al., 2012) Inventory relocation,

distribution, transshipment

Cost minimization Supply, demand satisfaction, capacity Mixed-integer programming Deterministic

(Shen et al., 2009) Vehicle routing Unsatised demand, visit time Time, demand satisfaction Mixed integer programming Tabu heuristic,

approximation heuristic

Stochastic

(Wang et al., 2009) Distribution of medical supplies Time and cost minimization Demand satisfaction Multi-objective programming Genetic algorithm Stochastic

(Zhao and Sun, 2008) Distribution, production Cost minimization Economic analysis, numerical

simulation

Deterministic

(Zhao and Han, 2010) Distribution of medical supplies Cost minimization Supply, demand satisfaction Linear programming deterministic

T.K.Dasaklis

etal./Int.J.Productio

nEconomics

139(2012)393410

401

abcHighlight

held in stock and relevant economic evaluations of vaccines in

develops a queuing network model where distribution times

T.K. Dasaklis et al. / Int. J. Production Economics 139 (2012) 393410402accordance with their time of expiration. Vaccine control strate-gies in the case of secondary bacterial infections (especiallypneumococcal infections) during a pandemic inuenza outbreakhave also been examined (Dhankhar et al., 2010). Finally, DuintjerTebbens et al. (2010) develop a mathematical framework fordetermining the optimal management of a vaccine stockpileover time where several logistical constraints are consideredlike capacity constraints, production and lling delays, risksTable 2 a classication of the epidemic logistics network congura-tion can be seen with respect to the methodologies applied, theobjectives of the models and the constraints. We pay special attentionin this sub-section because the literature described here depicts thecore logistics approaches wemanaged to retrieve. As the control of anepidemic outbreak entails a range of operations to take place, thetopic is very suitable for OR/MS research.

5.2.2. Stockpiling of medical supplies

Inventory control in the case of epidemics control may relate tomanaging specic medical supplies like vaccines, antiviral drugs orantibiotics as well as ancillary medical resources like PersonalProtective Equipment (PPE) etc. In some cases the problem ofstockpiling of medical supplies has been treated as a joint inventorystockpiling problem for several groups of hospitals. In this case it isassumed that mutual aid agreements for inventory sharing areestablished among the hospitals. A game theoretical approach isadopted for the formulation of the problem (Adida et al., 2011;DeLaurentis et al., 2008, 2009). Some researchers have also tried todetermine the amounts of medical supplies like PPEs to be held thatcould serve as a means of preventing inuenza pandemics(Hashikura and Kizu, 2009). Additionally, many researchers havetried to estimate the capacity of health care facilities to respond tospreading diseases in terms of materials (Radonovich et al., 2009)and PPEs needed (Rebmann et al., 2011).

As far as antiviral drugs are concerned, logistical constraintssuch as a nite stockpile of drugs and limited distribution ratesduring the deployment of antiviral drugs (Arinaminpathy andMcLean, 2008) or different antiviral coverage strategies using alimited stockpile of drugs (Arinaminpathy and McLean, 2009)have been examined in the case of an inuenza pandemic control.Lee et al. (2006b) utilize cost-benet and cost-effectivenessanalyses with Monte Carlo simulations to compare strategies forstockpiling neuraminidase inhibitors to treat and prevent aninuenza pandemic outbreak. Lugner and Postma (2009) utilizecost-effectiveness considerations when stockpiling antiviral drugsin order to mitigate an inuenza pandemic outbreak. Siddiqui andEdmunds (2008) develop a decision analytical model to investi-gate the cost-effectiveness of stockpiling antiviral drugs for apotential inuenza pandemic in the United Kingdom and thepossible role of near-patient testing in conserving antiviral drugstocks. Balicer et al. (2005) analyze strategies for the utilization ofstockpiled antivirals for a future inuenza pandemic and estimatecost-benet ratios, while Cinti et al. (2005) provide a strategy forstockpiling certain antivirals at a reasonable cost. Finally,Harrington and Hsu (2010) examine the so-called ManufacturerReserve Programs which are used by manufacturers to promotestockpiling of anti-viral drugs in preparation for pandemic inu-enza by non-governmental organizations such as hospitals.

In the case of vaccine inventory control, Liu (2007) considersthe case in which a sudden demand for vaccines attributed to anurgent incident like a natural outbreak or a bioterrorist attackoccurs and develops mathematical models for estimating thenecessary stockpile levels of vaccines in order to meet futureurgent needs. Dhankhar et al. (2009) examine the quantities to beassociated with the stockpile etc. The case of the polio vaccineand serving rates of antibiotics are considered.Decision support systems have also been developed in the case

of triage management operations. In this case, mathematicalstockpile is selected for the application of the proposedframework.

5.2.3. Triage operations

Once the medical supplies have been positioned to local PODs,dispensing operations must take place. In addition, several facil-ities like hospital and primary health-care premises might need tobe converted into dispensing centers. During the control effortwithin the dispensing centers patients will be examined bymedical personnel and will be further offered medical treatment(vaccination, antiviral treatment etc.). Logistical considerations inthis case may include the facility layout of the selected premisesto serve as dispensing centers, capacity planning, patient owlogistics, relevant dispensing or vaccination rates etc. Simulationis the prevalent methodological tool for addressing this kind ofproblems.

For example, Porco et al. (2004) develop a continuous-timeevent-driven network simulation model of smallpox ring vaccina-tion taking into consideration aspects like response logistics andlimited numbers of vaccinators. Whitworth (2006) utilizes simu-lation models for the evaluation of candidate PODs, alternativedispensing processes, stafng plans and trafc-management stra-tegies to help a community develop its plan for responding to ananthrax attack. Wein et al. (2003) develop a mathematical modelto compare various emergency responses in the event of anairborne anthrax attack where a set of spatially distributed two-stage queuing systems consisting of antibiotic distribution andhospital care is considered. Aaby et al. (2006) use discrete-eventsimulation queuing theory models for the clinic planning ofpublic-health services. Some guidelines for the physical designand lay-out of clinics are also proposed. Miller et al. (2006) applycombination strategies for controlling a bioterrorist attack (small-pox) by using discrete-event simulation. The model estimates themedical resources required for each step in the diagnosis andtreatment within public-health facilities and, therefore, possiblebottlenecks in care delivery. Patvivatsiri et al. (2007) simulatevarious possible bioterrorist attack scenarios in order to deter-mine the total time a patient stays in the system and to identifystaff requirements. Hupert et al. (2002) use discrete eventsimulation modeling to determine stafng levels for entry screen-ing, triage, medical evaluation and drug dispensing stations in ahypothetical antibiotic distribution center operating in low,medium, and high disease prevalence bioterrorism responsescenarios.

Apart from simulation, other approaches combine epidemicdiffusion rules and queuing theory and the models developedcould serve as capacity planning tools. Logistical considerations inthis case may relate to limited vaccination or dispensing rates.Craft et al. (2005) analyze a system of differential equations thatincludes among others a set of spatially distributed tandemqueues for distributing antibiotics and providing hospital care.Giovachino et al. (2005) utilize computer modeling techniqueswhere real data obtained by an exercise are used in order tooptimize the capacity of a dispensing center in the case of abioterrorist attack. Kaplan et al. (2003) evaluate existing andalternative proposals for emergency response to a deliberatesmallpox attack by embedding key operational features of suchinterventions into a smallpox disease transmission model wherethe model explicitly incorporates a tracing/vaccination queue, andhence can be used as a capacity planning tool. Hui (2010)modeling, large-scale simulation and optimization techniques

abcHighlight

abcHighlightimportant

abcHighlight

abcHighlight

abcHighlight

abcHighlight

abcHighlight

abcHighlight

abcHighlight

abcHighlight

abcHighlightalready read

abcHighlight

lann

(Dhankhar et al., 2009) Stockpiling of medical supplies

(Duintjer Tebbens et al., 2010) Stockpiling of medical supplies

lann

patie

lann

logis

lann

logis

tics,

T.K. Dasaklis et al. / Int. J. Production Economics 139 (2012) 393410 403(Giovachino et al., 2005) Dispensing of medical supplies,

patient ow logistics, capacity p

(Harrington and Hsu, 2010) Stockpiling of medical supplies

(Hashikura and Kizu, 2009) Stockpiling of medical supplies

(Hui, 2010) Dispensing of medical supplies,

(Hupert et al., 2002) Dispensing of medical supplies,

patient ow logistics, capacity p

(Kaplan et al., 2003) Capacity planning

(Lee et al., 2006a) Capacity planning, patient ow

(Lee et al., 2006b) Stockpiling of medical supplies

(Liu, 2007) Stockpiling of medical supplies

(Lugner and Postma, 2009) Stockpiling of medical supplies

(Miller et al., 2006) Patient ow logistics, capacity p

(Patvivatsiri et al., 2007) Capacity planning, patient ow

(Pietz et al., 2009) Facility layout, patient ow logis

(Porco et al., 2004) Capacity planningTable 3Logistical features of stockpiling and triage management literature.

References Logistical attributes

(Aaby et al., 2006) Dispensing of medical supplies,

patient ow logistics, capacity p

(Adida et al., 2011) Stockpiling of medical supplies

(Arinaminpathy and McLean, 2008) Stockpiling of medical supplies,

capacity planning

(Arinaminpathy and McLean, 2009) Stockpiling of medical supplies

(Balicer et al., 2005) Stockpiling of medical supplies

(Cinti et al., 2005) Stockpiling of medical supplies

(Craft et al., 2005) Dispensing of medical supplies

(DeLaurentis et al., 2008) Stockpiling of medical supplies

(DeLaurentis et al., 2009) Stockpiling of medical supplies

(Dhankhar et al., 2010) Stockpiling of medical suppliesare used for analyzing different dispensing scenarios, recongur-ing PODs during the containment effort, estimating the most cost-effective combination of dispensing strategies and determiningresources utilization along with facilities set-up and facilitieslay-out options (Lee et al., 2006a; Pietz et al., 2009). Comparisonsbetween different dispensing strategies have also been examined.Richter and Khan (2009) use multi-criteria decision analysis forassessing the trade-offs of different dispensing strategies in alarge metropolitan area during a biological or chemical event. InTable 3, the logistical features incorporated in the triage opera-tions as well as stockpiling of medical supplies literature aredepicted along with the relevant methodologies applied.

5.2.4. Other approaches

This sub-section refers to publications that are both of quali-tative and quantitative nature that could not be classied withinany of the aforementioned classication schemes. The qualitativeapproaches deal with a plethora of logistical issues arising in thecontext of vaccine supply chain or when quarantine and vaccina-tion campaigns are implemented. In addition, aspects of coordi-nation as well as information management are also examined inthese theoretical papers. More integrated response frameworksand sourcing aspects of the epidemic control supply chain buildup the remaining quantitative approaches.

Some of the aforementioned approaches embrace a plethora oflogistical considerations from stockpiling of medical supplies totheir distribution and dispensing. Zaric et al. (2008) develop adynamic compartmental model that incorporates anthrax diseaseprogression, prophylaxis and treatment in a population of

(Radonovich et al., 2009) Stockpiling of medical supplies

(Rebmann et al., 2011) Stockpiling of medical supplies

(Richter and Khan, 2009) Dispensing of medical supplies

(Siddiqui and Edmunds, 2008) Stockpiling of medical supplies

(Wein et al., 2003) Dispensing of medical suppliesProblem formulation/methodology applied

ing

Simulation, queuing theory

Game theory

Mathematical modeling (differential equations

and compartmental modeling)

Mathematical modeling (differential equations

and compartmental modeling)

Economic analysis or cost benet analysis

Scenario analysis

Queuing theory

Game theory

Game theory

Cost-effectiveness analysis

Economic analysis

Economic analysis, system dynamics

ing

Simulation

Mathematical modeling

Calculation system

nt ow logistics Queuing theory

ing

Simulation modeling

Queuing theory

tics Simulation, mixed integer programming problem

Economic analysis, simulation

Mathematical programming

Economic analysis

ing Simulation

tics Simulation

capacity planning Integer programming, simulation

Simulationexposed and potentially exposed individuals, where aspects ofdistribution of supplies from national and regional storage facil-ities to local communities, stockpiling of local inventories anddispensing to affected populations are examined. Cost-effective-ness analysis is utilized to evaluate alternative preparednessplans under different attack and response scenarios. Bravataet al. (2006) utilize cost-benet analysis to evaluate alternativestrategies for maintaining and dispensing local and regionalinventories of antibiotics and medical supplies for responses toanthrax bioterrorism. Brandeau et al. (2007) identies ve supplychain strategies that can potentially increase the speed ofresponse to a bioterrorism attack, reduce inventories and savemoney. Among them are the effective supply chain networkdesign, the effective inventory management, the postponementof product customization and modularization of component parts,the coordination of supply chain stakeholders and appropriateuse of incentives and nally the effective information manage-ment. Lee (2008) develops a simulation model for distributingand dispensing of medical supplies. The effectiveness of alter-native dispensing plans is evaluated and several improvementopportunities in the event of a bioterrorist attack are alsoidentied. Lee et al. (2009a) utilize a systems approach to analyzemass dispensing of countermeasures and present a set of power-ful modeling and computational tools to assist in strategic andoperational planning for the rapid establishment of a network ofdispensing sites and health facilities.

Research has also been directed towards the vaccine supplychain as well as the management of cold chain. Generally, in apossible inuenza pandemic outbreak logistical challenges anddifculties such as production, stockpiling and delivering vaccines

Economic analysis

Review

Multi-criteria decision analysis

Economic analysis

Queuing theory

abcHighlight

aran

atio

usto

gem

upp

ntro

s fo

supplies, patient Simulation, mathematical modeling

sup

of

T.K. Dasaklis et al. / Int. J. Production Economics 139 (2012) 393410404will arise (Jennings et al., 2008). For example (Adu et al., 1996),through an exploratory research, examine several issues thataffect the efciency of cold supply chains like mishandling ofvaccines and false storage conditions. Hessel (2009) examinesissues of vaccine allocation and relevant procurement processes,the establishment of critical health systems and infrastructurerequired for vaccine deployment, storage aspects associated withstockpiling pre-pandemic vaccines, and nally mutually agreedcontractual arrangements between manufacturers and govern-ments or international institutions. Other logistics aspects of thevaccine supply chain may concern vaccines production anddistribution (Collin and de Radigues, 2009). Ethical issues regard-ing vaccine distribution during an inuenza pandemic like whowill likely produce and own the vaccine and how vaccinedistribution and administration might be accomplished are stu-

Table 4Logistical features of other epidemics control literature.

References Logistical attributes

(Adu et al., 1996) Efciency of the cold chain

(Ak et al., in press) Sourcing

(Barbera et al., 2001) Logistical barriers for implementing qu

(Brandeau et al., 2007) Emergency supply chain design, inform

coordination, inventory management, c

(Brandeau et al., 2008) Information and communications mana

(Bravata et al., 2006) Stockpiling and dispensing of medical s

(Chick et al., 2008) Sourcing

(Collin and de Radigu es, 2009) Vaccine supply

(Conn et al., 2008) Information management, inventory co

(Date et al., 2011) Logistical and operational consideration

(Hadler, 2005) Vaccine supply

(Hessel, 2009) Vaccine supply

(Jennings et al., 2008) Vaccine supply

(Lee et al., 2009a) Facility location, dispensing of medical

ow logistics, capacity planning

(Lee, 2008) Distribution and dispensing of medical

(Manley and Bravata, 2009) Coordination

(Moore et al., 1990) Vaccine supply

(Zaric et al., 2008) Stockpiling, distribution and dispensingdied in Hadler (2005).Another important aspect of the epidemics control supply

chain is the sourcing decisions to be made. Antiviral drugs aswell as vaccines should be procured in order to be utilized duringthe containment effort. Governments around the world andpublic health institutes may be faced with uncertainty regardingthe exact amount of medical supplies to be purchased. Forinstance, many times huge quantities of vaccines are procuredby governments but they eventually become obsolete as thedemand for vaccinating the population remains low (unwilling-ness of people to get vaccinated etc.). Chick et al. (2008) examineseveral supply contracts that coordinate buyer (governmentalpublic health service) and supplier (vaccine manufacturer) incen-tives and design a variant of the cost-sharing contract thatprovides incentives to both parties which ultimately leads tothe improvement of the supply of vaccines in the case of annualinuenza outbreaks. In Ak et al. (in press) procurement aspects ofthe Pan American Health Organizations vaccine supply chain areexamined and several recommendations for the improvement ofdemand forecast for vaccines are provided. The study alsoexplores issues of transportation cost for vaccines and thepossible implications of bundle bidding. Despite the fact thatthe above papers address issues of seasonal inuenza, they havebeen included in the analysis as they provide strong insightstowards the issues of contract management and sourcing ofepidemics control supply chain.The remaining studies address issues of coordination, informa-tion management and general logistical impediments during thecontainment effort. For example, some research approachesexamine the logistical barriers for implementing vaccinationcampaigns in the case of complex emergencies like in the after-math of natural disasters (Date et al., 2011) or in the case ofdislocated people and refugees (Moore et al., 1990). Barbera et al.(2001) examined the likely effectiveness of quarantine, thelogistic barriers to its implementation, relevant legal issues raisedand possible adverse consequences that might result from quar-antine action implemented as a response to a possible bioterroristattack. Manley and Bravata (2009) developed a detailed frame-work that facilitates coordination of bioterrorism preparednessplanning among military and civilian decision makers. Cross-functional drivers of the epidemics control supply chain like

plies Simulation

Theoretical framework development, simulation

Case-study analysis

medical supplies Mathematical modeling, scenario analysisProblem formulation/methodology applied

Experimental study

Integer programming

tine programs Review

n management,

mization

Theoretical framework development

ent Theoretical framework development,

simulation, cost/benet analysis

lies Economic analysis

Game theory

Experimental study

l Case-study analysis

r vaccine campaigns Case-study analysis

Theoretical study

Theoretical study

Theoretical studyinformation management have also been studied (Conn et al.,2008). In Brandeau et al. (2008) issues of alternative communica-tion strategies for bioterrorism responses are evaluated and thecosts and benets of these strategies are examined throughsimulation. Logistical features of the literature presented in thissub-section as well as methodologies applied (for the quantitativeapproaches) are presented in Table 4.

6. Discussion

From the analysis of the selected literature a series of insightscan be derived with respect to the issues of the outbreaks dealtwith by the researchers, the logistical attributes taken intoconsideration as well as the methodologies applied. In the caseof the nature of the outbreaks, most research approaches addresstopics of bioterrorist response logistics and natural outbreaks(inuenza). As far as bioterrorist response logistics is concerned, itseems that the 2001 anthrax attack in United States triggered aseries of publications in this subject. Despite the fact that duringan ongoing humanitarian crisis logistical activities must takeplace to ensure the maintenance of a high level of sanitaryconditions and the provision of key medical supplies and vac-cines, research approaches addressing epidemics control logisticsaspects in the aftermath of natural disasters are limited. Evenfurther, epidemic outbreaks in the aftermath of natural disasters

abcHighlight

abcHighlight

T.K. Dasaklis et al. / Int. J. Production Economics 139 (2012) 393410 405still remain controversial among researchers. For example, Floretet al. (2006) suggest that the likelihood of epidemic outbreaksfollowing natural disasters remains very low whereas Kouadioet al. (2012) suggest the opposite. Therefore, this lack of con-sensus among the scientic community members could explain,to some extent, the lack of epidemics control logistics approachesin this context. Finally, research approaches regarding epidemicscontrol logistics in the case of mass gatherings were not located.

Interesting insights can also be drawn from the analysis basedon the logistical attributes of the research approaches reviewed.For instance, the research approaches developed so far in the caseof epidemics control logistics network conguration are based ona series of assumptions and simplications that have littlecorrelation with real-world problems and especially problemsarising in the eld. As mentioned in Section 5, most of theproposed methodologies combine differential equations for mod-eling the progression of the disease along with logistics networktechniques for the transportation and/or distribution of medicalsupplies. In the case of disease progression, it is assumed that therates of infection as well as the rates of transition of individualsbetween compartments are constant. In the majority of epide-miologic literature such assumptions are considered rather sim-plistic. Additionally, most of the research approaches considerthat the commodities transported and/or distributed are homo-geneous or that medical supplies are bundled together. Practi-cally, this is a simplication that contradicts with severalguidelines proposed by health-care policy makers (CDC, 2007).These guidelines clearly state that essential medical supplies andespecially vaccines should be transported by using specicvehicles under fully monitored conditions. Therefore, it is evidentthat issues of capacity and eet management or breaks in the coldchain attributed to faulty equipment have been overlooked so far.

Another aspect of the epidemics control supply chain cong-uration that has been paid little attention so far is the inter-relationship between the agent triggering an outbreak and theconguration of the control supply chain. In fact, the logistics ofresponse to a conrmed outbreak should rely heavily on the agenttriggering an outbreak and, therefore, the conguration of theepidemics control supply chain may be different depending onthe very nature of each agent. For instance, anthrax and smallpoxcould be both utilized as bioterrorists weapons but from anepidemiological point of view their control presents huge differ-ences. As mentioned in Whitworth (2006), anthrax, which is notcontagious, can be treated with antibiotics dispensed to heads ofhouseholds (who can then dispense the antibiotics to all thefamily members). On the other hand, smallpox is contagious, iscaused by a virus and can be only prevented by vaccination. Inthis case, each individual should be administered vaccination.Therefore, treatment centers would be more likely to handle alarger number of people in the case of a smallpox attack than inthe case of an anthrax attack with obvious implications for facilitylocation decisions. Response time frames are also different in eachcase, presenting different logistical challenges for the transporta-tion and distribution of medical supplies. Generally, according toAHRQ (2004) there are two conceptual approaches regardingmass prophylaxis in the case of disease outbreaks: push andpull. The push approach considers bringing medicine directlyto affected individuals and households whereas the pullapproach requires that individuals leave their homes or work-places in order to travel to certain treatment centers to receivemedication or get vaccinated. Putting aside any strengths andweaknesses these two approaches may have, it is evident that theadoption of each approach calls for a different supply chainconguration.

In the case of stockpiling of medical supplies most research

approaches include only the cost of actually purchasing theinventory into the problem formulation. Other aspects of inven-tory cost as well as inventory management operations likemaintenance, in-door handling, and activities of picking, packingand preparing for shipment are omitted. For example, in the caseof control of inuenza outbreaks antiviral drugs can be stockpiledin various forms (powder, pills etc.) thus their management maynecessitate additional handling operations during the contain-ment effort. In addition, inventory replenishment policies duringan ongoing epidemic outbreak are scarce. Unfortunately, themajority of the stockpiling approaches do not consider the factthat after a period of time inventories will be depleted andreplenishing activities should take place. Finally, it seems thatthe possible trade-offs between centralized and decentralizedpolicies of inventory management have not been paid attentionso far.

Research approaches addressing issues of triage managementoperations also rely on several assumptions that may contradictwith real-world problems. The most surprising nding is thattriage operations management approaches assume that there isan adequate supply of medical commodities and that the surgecapacity of dispensing centers is, in most of the cases, innite.Nevertheless, anthrax treatment requires a 60-day regimen ofantibiotics and this treatment will likely be implemented inconsecutive waves (Whitworth, 2006). Therefore, replenishmentpolicies of essential medical supplies should be taken intoconsideration. Apart from replenishment strategies, manyapproaches do not consider transporting people to the clinics toget vaccinated or administered medical treatment. Even further, itis assumed that the inux of patients to the treatment centers isconstant and that the prevalence and severity of disease are alsoconstant over certain periods. In addition, issues of handlingvaccines and other medical supplies are not taken into considera-tion. Service rates within dispensing centers are consideredconstant over time without taking into account fatigue of thepersonnel or even personnel illness. Issues of transshipment ofmedical supplies among dispensing centers are also not studied.Finally, possible bottlenecks attributed to post-vaccination com-plications for some people and, generally, severe adverse effectsof vaccination are not modeled.

Other aspects of the epidemics control supply chain poorlystudied in the literature may refer to some cross-functionaldrivers like sourcing and information as well as to some aspectsof coordination. Pricing and sourcing decisions in the case ofessential medical supplies procured for the containment of anepidemic may result inconvenient and costly. Appropriate con-tractual agreements between manufacturers and public healthinstitutions could alleviate any negative effects for the partiesinvolved but so far they have not been studied at a large scale.Issues of information management in epidemics control supplychain have also been paid limited attention. Information manage-ment plays a vital role during an epidemics outbreak and theprimary sources of information remain the surveillance systemsput in place. A prompt detection of a suspected disease outbreakhas a great impact on the progression of the disease and triggerscertain responses regarding the deployment of the availableresources while providing strong insights towards real demandrequirements. The ow of essential medical supplies, transporta-tion activities and demand for medical personnel are some of thelogistics-oriented features that depend on the available informa-tion regarding diseases progression. At the same time themanagement of materials ow during the containment effortnecessitates its own stream of information. Highly sophisticatedsystems in business supply chain and relevant technologies likeRFID could also be adopted in the case of epidemics containment.

Additionally, successful control of an epidemics outbreak

requires strong co-ordination amongst the parties involved. For

point of view (Jia et al., 2007a). The evaluation and validation ofthe existing research approaches could provide important feed-

T.K. Dasaklis et al. / Int. J. Production Economics 139 (2012) 393410406example, in Butler et al. (2002) aspects of coordination betweenpublic health and law enforcement agencies are studied but theydo not relate to the established epidemics control supply chain.Moreover, synergies between private and public health organiza-tions and governments should be explicitly studied and coopera-tion frameworks should be developed. For diseases outbreaks inthe aftermath of natural or man-made disasters the control effortmay be far more difcult as a result of the conict generatedamong the different parties involved (humanitarian organiza-tions, the army, national agencies, etc.) or real-time problemslike damaged or inadequate infrastructures.

Coordination and information management aspects may alsohave various implications for the adoption of several supply chainstrategies. For example, in Brandeau et al. (2007) several strate-gies are proposed that could lead to substantial increases in theresponsiveness and efciency of the epidemics control supplychain. As a matter of fact, strategies like lean, agile or leagileextensively adopted in the case of the business sector could alsobe adopted in the case of epidemics logistics provided that aseries of prerequisites are met. Generally, the adoption of thesestrategies may be feasible but a thorough assessment is necessaryfor deciding the appropriate strategy to be adopted in each case(nature of the outbreak, disease type etc.). In addition, theappropriate mixture of strategies (leagile) should be based on aclear understanding of the several building blocks of the epi-demics control supply chain like information, coordination, inte-gration and alignment. For instance, in the case of the businesssector information sharing and coordination are tied together(Chen, 2003). This close relationship between them enables thebusiness supply chains to adopt the appropriate strategy whichultimately leads them to match supply and demand. On the otherhand, epidemics control supply chains are overwhelmed withuncertainty and present low levels of coordination and informa-tion sharing among the parties engaged. Moreover, inherentuncertainties of the agent triggering the outbreak as well as theuncertainty of the supply chain itself to meet several needs renderthe adoption of these strategies a very difcult task.

As far as methodologies are concerned, economic analysis hasbeen extensively utilized for decisions regarding stockpiling ofmedical supplies (especially in the case of future inuenza out-breaks). Unlike the business sector, we found a small amount ofpapers addressing issues of inventory management in epidemicoutbreaks by using well established operations research techni-ques. In the case of logistics network conguration the problemsaddressed are formulated using prescriptive models. The majorityof these models are of deterministic nature and just a few of themincorporate stochastic parameters. From the analysis provided inTable 2 it is clear that most of the models incorporate costminimization as an objective while in the case of epidemicscontainment time is the crucial parameter. Techniques of decom-posing and/or deconstructing of the initial problems into smallerones have been limited and most approaches deal with small-scale problem. As a consequence, modeling and solving large-scale problems will necessitate the development of more efcientand effective heuristic algorithms. Furthermore, a large-scaleepidemic emergency in the future might challenge the scalabilityof the existing modeling approaches. In the case of managementof the triage operations and patient ow logistics, the majority ofthe developed models are descriptive. Logistical problems in