Background: Multidimensional Prognostic Index (MPI) is useful as a prognostic tool in hospitalized older patients, but our knowledge is derived from retrospective studies. We therefore aimed to evaluate in a multicenter, longitudinal, cohort study whether the MPI at hospital admission is useful to identify groups with different mortality risk and whether MPI at discharge may predict institutionalization, rehospitalization, and use of home care services during 12 months.
Methods: This longitudinal study, carried out between February 2015 and August 2017, included nine public hospitals in Europe and Australia. A standardized comprehensive geriatric assessment including information on functional, nutritional, cognitive status, risk of pressure sores, comorbidities, medications, and cohabitation status was used to calculate the MPI and to categorize participants in low, moderate, and severe risk of mortality. Data regarding mortality, institutionalization, rehospitalization, and use of home care services were recorded through administrative information.
Results: Altogether, 1,140 hospitalized patients (mean age 84.1 years, women = 60.8%) were included. In the multivariable analysis, compared to patients with low risk group at admission, patients in moderate (odds ratio [OR] = 3.32; 95% CI: 1.79-6.17; p < .001) and severe risk (OR = 10.72, 95% CI: 5.70-20.18, p < .0001) groups were at higher risk of overall mortality. Among the 984 older patients with follow-up data available, those in the severe-risk group experienced a higher risk of overall mortality, institutionalization, rehospitalization, and access to home care services.
Conclusions: In this cohort of hospitalized older adults, higher MPI values are associated with higher mortality and other negative outcomes. Multidimensional assessment of older people admitted to hospital may facilitate appropriate clinical and postdischarge management.
Keywords: Aged; Cohort study; Hospital; Multidimensional Prognostic Index; Prognosis.
© The Author(s) 2018. Published by Oxford University Press on behalf of The Gerontological Society of America. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com.