Using Multidimensional Prognostic Indices to improve cost-effectivenessof interventions in multimorbid, frail older persons

FINAL REPORT

HEALTH CARE PROVIDERSBACKGROUND INFORMATION AND RECOMMENDATIONS

The MPI_AGE project is a research project co-funded by the European

Union through the Health 2007-2013 Programme that has explored ways

to reduce unnecessary use of health care resources in older subjects through the design

of appropriate, tailored, integrated, multi-professional, planned interventions.

This project opted to use a well-validated diagnostic and prognostic tool, the Multidimensional Prognostic Index (MPI), based on a standard Comprehensive Geriatric Assessment (CGA), and analyzed how it can improve the cost-effectiveness of health interventions in complex older individuals with multimorbidity and polypharmacy.

To do so, a review analysis on the use of predictive rules in clinical and management decision making in older residents in different European regions has been performed to identify reference models to allocate resources and lower costs in healthcare.

Analyses of multi-national databases including information on drug prescriptions and older subjects’ characteristics have identified some setting-specific MPI-profiles in which individual interventions are more cost-effective in terms of improved survival.

OBJECTIVES

The aims of the MPI_AGE project were to:

• identify the most cost-effective health interventions according to the individual prognosticmortality-risk profile;

• improve multi-professional interactions and collaboration in performing integrated carepathways of interventions according to the individual MPI-risk profile;

• develop tailored intervention programs based on the individual MPI-profile of older subjects;

• explore ways to reduce health-related costs, linked to a reduction of hospitalization rates, in-hospital length-of stay, institutionalization rates, and inappropriate and unnecessaryprescription drug use.

INTRODUCTION

PARTNERS OF MPI_AGE

The MPI_AGE project involves the following centers and organizations in twelve different countries.

ASSOCIATE PARTNERS

• Italy: AULSS16 - Azienda Unità Locale Socio Sanitaria nº 6 Euganea, Padova (Coordinating center)

• Czech Republic: Univerzita Karlova, Prague

• France: CHU, Poitiers

• Germany: Universitätsklinikum Köln, Cologne

• Netherlands: Erasmus Universitair Medisch Centrum, Rotterdam

• Spain: Fundación Para la Investigación Biomédica del HospitalUniversitario Ramon y Cajal, Madrid

• Sweden: Karolinska Institute, Stockholm

• Europe: European Union Geriatric Medicine Society (EUGMS)

COLLABORATING PARTNERS

• Australia: Flinders University, Adelaide

• Belgium: EUREGHA; AGE platform Europe

• Bulgaria: Bulgaria Association on Ageing

• Estonia: Estonian Association of Gerontology

• Italy: Regione del Veneto - Servizio per le relazioni socio-sanitarie; Centro Regionale Management Progetti Europei – CReMPE; National Research Council – CNR; IRCCS Casa Sollievo della Sofferenza; Università degli Studi di Messina; E.O. Ospedali Galliera di Genova

• Hungary: Semmelweiss University, Budapest

• United States of America: National Institute on Aging; John Hopkins Bloomberg School of Public Health

THE MULTIDIMENSIONAL PROGNOSTIC INDEX (MPI)

MPI is a prognostic tool based on a standard CGA that has been shown to predict short- andlong-term survival in older subjects, as well as other negative health outcomes (hospitalization,nursing home admission, length of hospital stay).

The MPI includes information in 8 domains of CGA as follows:

1. Activities of Daily Living (ADL): 6 items

2. Instrumental Activities of Daily Living (IADL): 8 items

3. Short Portable Mental Status Questionnaire (SPMSQ): 10 items

4. Mini Nutritional Assessment (MNA): 18 items

5. Exton-Smith Scale (ESS): 5 items

6. Cumulative Index Rating Scale – Comorbidity Index (CIRS-CI): 14 items

7. Number of drugs used: 1 item

8. Co-habitation status: 1 item

To obtain the final index of a given individual, a program calculates a MPI score, which rangesfrom 0 to 1. This calculation can be easily performed by a program that can be downloaded at no cost (www.mpiage.eu) or using an IOS free app (iMPI).

The MPI index predicts mortality – lower values are associated with lower mortality. Usually, results are ranked in three levels:

• 0 to 0.33 – low mortality risk, MPI 1

• 0.34 to 0.66 – moderate mortality risk, MPI 2

• 0.67 to 1 – high mortality risk, MPI 3

Studies have shown that MPI values predict short and long-term mortality in older subjects. The prognostic

accuracy of the aggregated MPI is significantlygreater than the prognostic accuracy of its

individual components considered singularly.Thus, this multidimensional approach can

be successfully used in older patients forboth clinical and administrative purposes.

A cluster analysis on CGA data was made to identify independent domains and variables ofCGA that predicted mortality. Following a step-wise method, other domains of the CGA wereprogressively included in the model in order to obtain the best combination of different CGAdomains for the best prognostic accuracy. Clinical results have confirmed that MPI has a goodprognostic effect on short- and long-term mortality, with a close agreement between theestimated and the observed mortality. The MPI has been validated in geriatric complexpatients, and also in older patients affected with several acute and chronic diseases such asgastrointestinal bleeding, pneumonia, heart failure, dementia, liver cirrhosis, chronic kidneydisease and transient ischemic attack.

The MPI has been validated in hospitalized older people, community-dwelling older (toestablish accessibility to nursing homes or homecare services, and healthy older people). Many clinical studies and recent meta-analyses and systematic reviews report that the MPI hashigh accuracy, excellent calibration, and the highest validity, reliability and feasibility comparedto others tools used to identify frail older patients both in clinical practice and in research.

MPI-SVAMA

The MPI-SVAMA index is an evolution of the MPI. This version obtains information from SVAMA(Scheda per la Valutazione Multidimensionale delle persone adulte e Anziane), a standardizedCGA used in the Veneto Region (Italy) to grant access to health and social care.

The MPI-SVAMA index is calculated from the following nine domains:

1. Age (years)

2. Gender

3. Nursing care: 11 items

4. Exton-Smith scale: 5 items

5. Short Portable Mental Status Questionnaire: 10 items

6. Activities of Daily Living: 6 items

7. Functional mobility (mobility items of Barthel scale): 3 items

8. Social network: 1 item

9. Main medical diagnosis

MPI-SVAMA was built using information from SVAMA of a population of more than 7,800 olderpeople aged ≥ 65 years old living at home, and validated on a separate population of over4,000 older subjects, using 1-month and 1-year mortality as outcomes. This index can becalculated with the use of a free program (www.mpiage.eu), and also ranges from 0 to 1.However, cut-off points used to rank MPI-SVAMA are different and time dependent.

For short term mortality (1 month): For long term mortality (1 year):

• 0 to 0.41 – low mortality risk, MPI 1 • 0 to 0.33 – low mortality risk, MPI 1

• 0.42 to 0.53 – moderate mortality risk, MPI 2 • 0.34 to 0.47 – moderate mortality risk, MPI 2

• 0.54 to 1 – high mortality risk, MPI 3 • 0.48 to 1 – high mortality risk, MPI 3

DEVELOPMENT AND VALIDATION OF THE MPI

A. Analysis on the use of predictive rules in clinical decision making in community-dwelling older people.An analysis on the use of predictive rules (MPI) in clinical and management decision-makingin community-dwelling older subjects was performed using clinical, functional andadministrative data included in large databases from different European regions.

Population-based cohort databases containing data on large populations are potentially anaccurate source in predicting multidimensional impairment and frailty in heterogeneousgroups of older persons in terms of short- and long-term mortality and other negative healthoutcomes.

Multidimensional Prognostic Index in community dwelling older healthy people

The application of MPI in 2,472 older residents in Stockholm (Sweden) included in the SwedishNational Study on Aging and Care (SNAC-K) demonstrated a clear and significant associationbetween MPI score and survival time and risk of hospitalization.

Higher MPI risk scores were associated with more days in hospital and with fewer years of survival across a broad and stratified age range (65-74 years, 75-84 years, > 84 years) in this population based setting with a very long-term follow-up (up to 12 years). This is the first study that confirms the prognostic value of MPI in a population-based cohort.

Multidimensional Prognostic Index in community dwelling older people with dementia

Previous studies showed that MPI accurately stratified hospitalized older patients withdementia into groups at varying risk of short- and long-term mortality. Also, MPI was shown to predict the risk of hospitalization in outpatients with cognitiveimpairment and dementia. Moreover, the MPI has been used as a valid outcome measure in an intervention trial in patients with Alzheimer’s disease.

In MPI_AGE, a cohort of 6,800 community-dwelling older persons with dementia wereidentified in a database of almost 23,000 community-dwelling older

residents in Veneto Region (Italy) who applied to homecareservices or nursing home admission. In this population,

MPI strongly predicted 3-year mortality. When stratified according to MPI-defined strata,

after adjustment for covariables, the use of drugs for dementia (anti-cholinesterase

inhibitors and/or memantine) wasassociated with lower mortality in

subjects with MPI-defined low-risk (HR 0.71 95%CI 0.54-0.92, p<0.01) and moderate-risk (HR 0.61, 95%CI0.40-0.91, p<0.01), but not in thosewith high-risk of mortality (HR 1.04,95%CI 0.52-2.06). These results show that MPI may be useful to make treatment choicesin Alzheimer patients.

CORE ACTIVITIES OF MPI_AGE

B. Use of MPI to improve cost-effectiveness of drug treatments in older people with multimorbidity and polypharmacy. MPI was used to identify the effectiveness of different drugs across different levels of frailty and dependence in large multinational databases.

Multidimensional Prognostic Index in a general practice database

MPI_AGE explored the usefulness of data included in the THIN database (this database collectsinformation on almost 1,200,000 subjects followed-up by their General Practitioners in theUnited Kingdom). It was found that multidimensional indicators were not frequently recordedin the THIN database, so a full MPI could not be obtained. However, the accuracy of a modelincorporating age, sex, and available functional and cognitive measures was able to predict 1-month and 1-year mortality among community-dwelling older people. The use of such indicators in GP database reflects a newer approach, which may improvemortality prediction among older persons in clinical practice.

Multidimensional Prognostic Index and the care of frail older people: where evidence is lacking

Frail subjects are usually excluded from clinical trials, so clinical guidelines give little insight onhow to treat frequent problems in geriatric complex patients. The use of MPI to betterunderstand risks and benefits of drug treatments in this population was explored in two areasin which evidence is still lacking and the risk-benefit ratios are unclear: the use of statins andanticoagulation therapies in very old, high-risk, frail subjects.

First, in the database of almost 23,000 community-dwelling older residents in Veneto Region(Italy), 1,712 subjects with diabetes mellitus (mean age=81.1±7.3 years, males = 43%) wereselected. By using the clinical stratification according to the MPI, it was shown that statintreatment was significantly associated with reduced three-year mortality independently of ageand multidimensional impairment in this population.

A second population of 2,597 patients with coronary artery disease was also analyzed (meanage=84±7.3 years, males = 44.5%), again showing that statin treatment was significantlyassociated with reduced three-year mortality independently of age and MPI score, althoughthe frailest were less likely to be treated with statins.

Finally, a study of 1,827 community-dwelling older persons with atrial fibrillation classified bythe MPI in three grades of risk demonstrated a benefit of anticoagulation in terms of lower all-cause mortality over a mean follow-up of 2 years, regardless of poor health and functionalconditions. Interaction tests showed that reduction of mortality with anticoagulants was higherin subjects with severe complexity (higher MPI score).

Multidimensional Prognostic Index and invasive procedures

An additional relevant topic in geriatric medicine is the appropriate selection of older patientswho can benefit from interventions with invasive therapeutic procedures. To explore the use of MPI in this clinical setting, a prospective observational study was carried out in consecutivepatients aged ≥75 years who underwent transcatheter aortic valve implantation (TAVI). MPI was calculated at baseline. A follow-up time of 1-year was performed. Among the116 patients included (mean age 86.2±4.2 years, mean MPI score 0.39±0.13), mortality rate was significantly different between MPI groups at 6 and 12 months (p=0.04 and p=0.02). Kaplan Meier survival estimates at 1-year stratified by MPI groups were significantly different(HR=2.83, 95% CI 1.38-5.82, p=0.004). The study indicated that CGA-based MPI was an accuratetool to predict prognosis and select older patients for TAVI procedures.

CORE ACTIVITIES OF MPI_AGE

C. Use of MPI to improve resource allocation in older hospitalized persons. A large multicenter clinical trial was performed to analyze the role of MPI in theidentification of hospitalized patients with different characteristics and needs, which were prospectively followed for 1 year.

In the last years, great attention has been given to the proper identification of prognosis tohelp clinical decision-making to tailor appropriate diagnostic and treatment interventions forfrail older patients. It is now well known that adding functional, cognitive, nutritional and socialdata improves prognostic tools that are only based on an index disease, so multidimensionaltools should be useful to better quantify the prognosis of frail older subjects.

However, a large systematic review identified, after searching over 20,000 references, a verysmall number of validated prognostic indices for mortality that meet the necessaryrequirements of accuracy and calibration required to be used in a clinical setting, i.e. eightindices for hospitalized older patients, two for those living in nursing homes, and six forcommunity-dwellers. Among the eight indices studied in hospital settings, MPI was the onlyCGA-based predictive tool. MPI is well-calibrated and has a good discrimination properties andaccuracy in predicting mortality both at short-term (1 month) and long-term (1 year). Veryrecently, a systematic review reported that MPI had the highest validity, reliability and feasibility(i.e. score 14 - maximum value - on the QUADAS system), compared to other tools used toidentify frail older patients.

MPI has been used in several multi-center studies in Italy to predict prognosis in hospitalized,frail older individuals. For example, in a prospective multicenter study involving over 2,000hospitalized older patients in 20 different hospitals, MPI was a significantly more accuratepredictor of short- and long-term all-cause mortality than other three frailty indices commonlyused in clinical practice including FI-CGA and SOF. Other prospective multicenter studiesdemonstrated that MPI was an independent predictor of in-hospital mortality and of length ofhospital stay in such populations. Moreover, during hospitalization MPI score changed in mostof patients (overall, the MPI score improved in 35.3% and worsened in 26.6% of patients, thesechanges were gender and age-related) and therefore MPI might be used to objectively trackand monitor the clinical evolution of acutely ill older patients admitted to the hospital.

The MPI_AGE multicenter trial

Based on these findings, the MPI_AGE project evaluated prospectively 1,148 hospitalizedpatients recruited in 9 international centers across Europe and Australia (mean age 84.4±8.8years, females = 61%) and classified them according to the MPI score at admission and atdischarge from the hospital. Patients were followed for one year.

Results of this prospective multicenter international study confirmed that MPI at hospitaladmission significantly predicts in-hospital mortality (MPI-1 HR =1.0 [reference], MPI-2 HR=3.0,95%CI 1.4-6.7, MPI-3 HR=10.4, 95%CI 4.7-23.2) with good accuracy (area under the curve-AUC=0.76, p<0.001) and MPI at hospital discharge significantly predicts 1-year mortality (MPI-1 HR =1.0 [reference], MPI-2 HR=2.4, 95%CI 1.1-5.4, MPI-3 HR=12.1, 95%CI: 5.2-28.1), again with good accuracy (AUC 0.77). Moreover, MPI class on admission was significantlypredictive of length-of-hospital stay (LOS), i.e. MPI 1= 7.9±5.6 days vs MPI 2= 13.5±10.2 days vs MPI 3= 16.5±13.6 days (data adjusted for age, gender, hospital and diagnosis at discharge).

CORE ACTIVITIES OF MPI_AGE

In this population, MPI significantly changed during hospitalization in most of patients,improving in 35% and worsening in 27% of them. Interestingly, an improvement of MPI duringhospitalization was associated with a significant reduction in mortality at 1 year (HR=0.65,95%CI 0.48-0.87, p=0.004). In addition, MPI worsening was significantly associated with a 54%increase in the use of home-services (HR 1.54, 95%CI 1.10-2.15, p=0.01) (data adjusted for MPIat admission, age, gender, hospital, LOS and diagnosis at discharge).

Diagnostic tests used during hospitalization have been classified according to MPI value. Olderpatients with higher MPI values were more often diagnosed using X-Ray tests (MPI-1 54% vsMPI-2 55% vs MPI-3 61%, p=0.006) and less using ultrasonography (MPI-1 18.5% vs MPI-2 16.6%vs MPI-3 12.8%, p=0.01) or endoscopy (MPI-1 5.0% vs MPI-2 3.5% vs MPI-3 3.7%, p=0.02).No differences were observed in the use of CT, MRI or nuclear medicine diagnostic tests.

During the one-year follow-up period, MPI grade was also significantly associated with the useof home-care services (MPI-1 Odds Ratio 1.0 reference; MPI-2 OR 2.47, 95%CI 1.5-4.0, MPI-3 OR 1.82, 95%CI 1.1-3.0, p=0.002) and with admission to nursing homes (MPI-1 OR 1.0 reference;MPI-2 OR 2.2, 95%CI 1.3-3.8, MPI-3 OR 1.7, 95%CI 0.9-2.9, p=0.002).

Along the one-year follow-up, 606 out of 1,140 patients (53.1%) were re-hospitalized. One-year re-hospitalization was significantly associated with all-cause mortality (HR=1.79,95%CI 1.42-2.36, p<0.001). Moreover, one-year re-hospitalization was significantly associatedwith higher MPI grade (MPI-1 Odds Ratio: 1.0 reference; MPI-2 OR 1.69, 95%CI 1.15-2.48, MPI-3 OR 1.60, 95%CI 1.07-2.38), lower age (OR=0.98,95%CI 0.96-0.99), male gender (OR: 1.29, 95%CI 1.0-1.65) and the hospitalcenter (OR: 0.91, 95%CI 0.86-0.96).

CORE ACTIVITIES OF MPI_AGE

• Health interventions should be adapted to individual needs of older patients,especially for those with high disease burden, high complexity or relevant majorphysical and mental impairments.

• Individual needs should be objectively assessed by means of validated instruments. These instruments should be multidimensional in order to capture all relevantdomains.

• Objective assessment of needs may avoid discrimination of older people (ageism) in decision-making.

• The Multidimensional Prognostic Index (MPI) has proved to be the best validated assessment instrument in various healthcare settings (community, hospital andnursing homes) and across a wide range of diseases and conditions.

• MPI also identifies problems in several domains that may benefit from specialistcomprehensive geriatric care.

• Tailored healthcare interventions have the potential to reduce the inappropriate use of resources (hospitalizations, drugs, diagnostic and other procedures) and to allow well-established treatments and interventions to be used in older people who canbenefit from them.

• Tailored healthcare interventions have the potential to reduce inappropriate health-related costs.

• MPI can be adapted for use in population-based (Primary Care) and disease-oriented databases to accurately predict survival and other health outcomes.

• MPI can be used to explore how evidence-based knowledge of drugs and invasiveinterventions applies across different levels of frailty, complexity and lifeexpectancy.

• In hospitalized older persons MPI identifies groups at risk for several hospitaloutcomes (i.e. mortality, length of stay, use of diagnostic tests). Individuals withineach risk group may benefit from the adaptation of interventions to his/herprognosis and needs.

• In hospitalized older persons MPI predicts several post-discharge outcomes: one-year mortality, rehospitalisation, admission to a nursing home, use of home-care services.

• Changes in MPI during hospitalization predict long-term mortality and use of home-care services.

• The role of MPI as a potential outcome measure for interventions needs to beexplored.

MPI_AGE RECOMMENDATIONS AND CONCLUSIONS

• Angleman SB, Santoni G, Pilotto A, Fratiglioni L, Welmer AK on behalf of the MPI_AGE Project Investigators. Multidimensional PrognosticIndex in association with future mortality and number of hospital days in a population-based sample of older adults: results of the EUfunded MPI_AGE Project. PLoS One 2015;10:e0133789.

• Brunello A, Fontana A, Zafferri V et al. Development of an oncological-multidimensional prognostic index (Onco-MPI) for mortalityprediction in older cancer patients. J Cancer Res Clin Oncol. 2016 May;142(5):1069-77.

• Bureau ML, Liuu E, Christiaens L et al. on behalf of the MPI_AGE Project Investigators. Using a multidimensional prognostic index (MPI)based on comprehensive geriatric assessment (CGA) to predict mortality in elderly undergoing transcatheter aortic valve implantation. Int JCardiol. 2017 Jun 1; 236: 381-386.

• D’Onofrio G, Sancarlo D, Addante F et al. A pilot randomized controlled trial evaluating an integrated treatment of rivastigmine transdermalpatch and cognitive stimulation in patients with Alzheimer’s disease. Int J Geriatr Psych 2015;30:965-75.

• Daragjati J, Musacchio C, Polidori MC et al. on behalf of the MPI_AGE Investigators. Multidimensional Prognostic Index in hospitalized elderlypatients across Europe and Australia: a prospective multicenter study of the European MPI_AGE project. Eur Geriatr Med 2016; 7(S1):S16.

• Dent E, Kowal P, Hoogendijk EO. Frailty measurement in research and clinical practice: a review. Eur J Int Med 2016;31:3-10.• Fontana L, Addante F, Copetti M et al. Identification of a metabolic signature for multidimensional impairment and mortality risk in

hospitalized older patients. Aging Cell. 2013 12(3):459-66.• Gallucci M, Battistella G, Bergamelli C, et al. Multidimensional Prognostic Index in a cognitive impairment outpatient setting: mortality and

hospitalizations. The Treviso Dementia (TREDEM) Study. J Alzheimer’s Dis 2014;42:1461-8.• Giantin V, Valentini E, Iasevoli M et al. Does the Multidimensional Prognostic Index (MPI), based on a Comprehensive Geriatric Assessment

(CGA), predict mortality in cancer patients? Results of a prospective observational trial. J Geriatr Oncol 2013; 4: 2018-217. • Giantin V, Falci C, De Luca E et al. Performance of the Multidimensional Geriatric Assessment and Multidimensional Prognostic Index (MPI)

in predicting negative outcome in older adults with cancer. Eur J Cancer Care 2016 Oct 10; doi:10.1111/ecc.12585 [Epub ahead of print].• Gill TM. The central role of prognosis in clinical decision making. JAMA 2012; 307: 199-200.• Mangoni AA, Pilotto A. New drugs and patient-centred end-points in old age: setting the wheels in motion. Expert Rev Clin Pharmacol

2016; 9 (1): 81-89• Pilotto A, Ferrucci L, Franceschi M, et al. Development and validation of a multidimensional prognostic index for 1-year mortality from the

comprehensive geriatric assessment in hospitalized older patients. Rejuvenation Res 2008;11:151-61.• Pilotto A, Addante F, D’Onofrio G, Sancarlo D, Ferrucci L. The Comprehensive Geriatric Assessment and the multidimensional approach. A

new look at the older patient with gastroenterological disorders. Best Pract Res Clin Gastroenterol 2009;23: 829-37.• Pilotto A, Addante F, Ferrucci L et al. The Multidimensional Prognostic Index (MPI) predicts short and long-term mortality in older patients

with community-acquired pneumonia. J Gerontol A Biol Med Sci 2009;64:880-887.• Pilotto A, Sancarlo D, Panza F, et al. The Multidimensional Prognostic Index (MPI) based on a Comprehensive Geriatric Assessment predicts

short- and long-term mortality in hospitalized older patients with dementia. J Alzheimer’s Dis 2009;18:191-9.• Pilotto A, Addante F, Franceschi M et al. A Multidimensional Prognostic Index (MPI) based on a comprehensive geriatric assessment predicts

short-term mortality in older patients with heart failure. Circ Heart Fail 2010; 3: 14-20.• Pilotto A, Sancarlo D, Franceschi M et al. A multidimensional approach to the geriatric patient with chronic kidney disease. J Nephrol 2010;

23: S5-10.• Pilotto A, Rengo F, Marchionni N, Sancarlo D, Fontana A, Panza F, Ferrucci L on behalf of the FIRI-SIGG Study group. Comparing the

prognostic accuracy for all-cause mortality of the Frailty Instruments: a multicentre 1-year follow-up in hospitalized older patients. PLosOne2012; 7(1): e29090 (1-9).

• Pilotto A, Panza F, Ferrucci L. A Multidimensional Prognostic Index in Common Conditions Leading to Death in Older Patients. Arch InternMed 2012; 172 (7): 594-5.

• Pilotto A, Gallina P, Fontana A, et al. Development and validation of a multidimensional prognostic index for mortality based on a standardizedMultidimensional Assessment Schedule (MPI-SVaMA) in community-dwelling older subjects. J Am Med Dir Assoc 2013;14:287-92.

• Pilotto A, Sancarlo D, Polidori MC on behalf of the MPI_ AGE Investigators. The MPI_AGE European Project: Using MultidimensionalPrognostic Indices (MPI) to improve cost-effectiveness of interventions in multimorbid frail older persons. Background, aim and design. EurGeriatr Med 2015;6:184-8.

• Pilotto A, Panza F, Copetti M, et al. on behalf of the MPI_AGE Project Investigators. Statin treatment and mortality in community-dwellingfrail older patients with diabetes mellitus: a retrospective observational study. PLoS One 2015;10:e0130946.

• Pilotto A, Arboretti GR, Panza F et al. on behalf of the MPI_AGE Investigators. Role of anti-dementia drugs and multidimensional impairmenton mortality rates in frail multimorbid older patients with dementia: results from the European MPI_AGE project. 11th Congress of theEUGMS (European Union Geriatric Medicine Society), Oslo, Norway, 16-18 September 2015. Eur Geriatr Med 2015; 6(S1): S20.

• Pilotto A, Sancarlo D, Rengo F, et al. on behalf of the FIRI SIGG Study Group. The Multidimensional Prognostic Index predicts in-hospitallength of stay in older patients: a multicenter prospective study. Age Ageing 2016;45: 90-6.

• Pilotto A, Gallina P, Copetti M et al. on behalf of the Multidimensional Prognostic Index_Age Project Investigators. Warfarin Treatment andAll-Cause Mortality in Community-Dwelling Older Adults with Atrial Fibrillation: A Retrospective Observational Study. J Am Geriatr Soc 2016Jul;64(7):1416-24.

• Pilotto A, Gallina P, Panza F, et al. on behalf of the Multidimensional Prognostic Index_Age Project Investigators. Relation of statin use andmortality in community-dwelling frail older patients with coronary artery disease Am J Cardiol. 2016 Dec;118(11):1624-1630.

• Pilotto A, Cella A, Pilotto A et al. Three decades of Comprehensive Geriatric Assessment: evidence coming from different healthcare settingsand specifric clinical conditions. J Am Med Dir Assoc. 2017; 18(2):192.e1-192.e11.

• Sancarlo D, D’Onofrio G, Franceschi M et al. Validation of a modified Multidimensional Prognotic Index (m-MPI) including the MiniNutritional Assessment Short-form (MNA-SF) for the prediction of one-year mortality in hospitalized elderly patients. J Nutr Health Aging2011; 15: 169-173.

• Sancarlo D, Pilotto A, Panza F et al. A Multidimensional Prognostic Index (MPI) based on a Comprehensive Geriatric Assessment predictsshort- and long-term all-cause mortality in hospitalized older patients with Transient Ischemic Attach. J Neurol 2012; 259:670-8.

• Siontis GC, Tzoulaki I, Ioannidis JP. Predicting death: an empirical evaluation of predictive tools for mortality. Arch Intern Med 2011; 171: 1721-6.• Sultana J, Basile G, Trifirò G. Measuring frailty in population-based healthcare databases: multi-dimensional prognostic indices for the

improvement of geriatric care. Geriatric Care 2016; 2: 5596. • Sultana J, Fontana A, Giorgianni F, et al. Multidimensional frailty indicators in a nationwide GP database predict mortality in the elderly:

MPI_AGE results. 11th Congress of the EUGMS (European Union Geriatric Medicine Society), Oslo, Norway, 16-18 September 2015. EurGeriatr Med 2015;6(S1):S23.

• Volpato S, Bazzano S, Fontana A, Ferrucci L, Pilotto A on behalf of the MPI-TriVeneto Study Group. Multidimensional Prognostic Indexpredicts mortality and length of stay during hospitalization in the older patients: a multicenter prospective study. J Gerontol A Biol Sci MedSci 2015; 70 (3): 323-9.

• Volpato S, Daragjati J, Simonato M, et al. Change in the multidimensional prognostic index score during hospitalization in older patients.Rejuvenation Res 2016;19:244-51.

• Warnier RMJ, Van Rossum E, Van Velthuijsen E, et al. Validity, reliability and feasibility of tools to identify frail older patients in inpatienthospital care: a systematic review. J Nutr Health Aging 2016;20:218-3

• Yourman LC, Lee SJ, Schonberg MA, Widera EW, Smith AK. Prognostic indices for older adults: a systematic review. JAMA 2012; 307:182-192.

REFERENCES

LEAD PARTNERAULSS 6 Euganea Padova

Via E. degli Scrovegni, 14 • 35131 Padova • ItalyWebsite: www.aulss6.veneto.it

E-mail: geriatria.osa@aulss6.veneto.it

COMMUNICATION LEADEREUGMS - European Geriatric Medicine Society

Secretariat Vienna - Laudongasse, 21/10 • 1080 Vienna • AustriaE-mail: secretariat@eugms.org

PROJECT LEADERDr. Alberto Pilotto - E.O. Galliera Hospital

Mura delle Cappuccine, 14 • 16128 Genova • ItalyE-mail: alberto.pilotto@galliera.it

WEB & MEDIA

www.mpiage.eu

www.facebook.com/MPIAGE

@MPI_AGE