Multidimensional Prognostic Index for Predicting Mortality in Hospitalized Older Adults: A Systematic Review

Abstract

Introduction: Traditional frailty tools often require performance-based testing, limiting use in acutely ill geriatric patients. The Multidimensional Prognostic Index (MPI), derived from Comprehensive Geriatric Assessment (CGA), predicts short- and long-term mortality. This review synthesizes validation studies across settings.
Methods: Systematic search of MEDLINE, CINAHL (2018–2024) for prospective studies validating MPI in patients ≥70 years hospitalized for acute illness. Primary outcome: 1-month and 12-month mortality. Meta-analysis of c-statistics.
Results: Thirty-one studies (n=27,603) included. Pooled 1-month mortality c-statistic 0.78 (95% CI 0.74–0.82); 12-month mortality 0.76 (0.72–0.80). High-risk MPI (>0.66) conferred 12-month mortality risk of 68% vs low-risk (MPI<0.33) mortality 12% (p<0.001). MPI outperformed Clinical Frailty Scale (difference in c-statistic +0.07, p=0.03). Integration with electronic health records feasible in 12 studies.
Discussion: MPI predicts mortality independent of diagnosis, but heterogeneity in CGA components (some used 8 domains, others 6) limits direct comparability. Most studies conducted in Europe, generalizability to other populations unknown.
Conclusions: MPI is a valid, mortality-predicting tool for hospitalized older adults. Routine CGA-based calculation should be implemented in geriatric units.