Prognostic Interplay of Functional Status and Multimorbidity Among Older Patients Discharged From Hospital

Objectives: The purpose of this study was to investigate the prognostic weight of multimorbidity and functional impairment over long-term mortality among older patients discharged from acute care hospitals.Design: A prospective multicenter observational study.Setting and Participants: Our series consisted of 1967 adults aged >= 65 years consecutively admitted to acute care wards in Italy, in the context of the Report-AGE project.Methods: After signing a written informed consent, all patients underwent comprehensive geriatric assessment by Inter-RAI Minimum Data Set acute care. The primary endpoint of the present study was long-term mortality. Patients were grouped into 3 functional clusters and 3 disease clusters using the K-medians cluster analysis. The association of functional clusters, disease clusters, and Charlson score categories with long-term mortality was investigated through Cox regression analysis and the inter-cluster classification agreement was further estimated. Finally, the additive effect of either disease clusters or Charlson score on predictive ability of functional clusters was assessed by using changes in Harrell's C-index and categorical Net Reclassification Index (NRI).Results: Functional clusters, disease clusters, and Charlson score were significant predictors of long-term mortality, but the interclassification agreement was poor. Functional clusters predicted mortality with greater accuracy [C-index 0.66, 95% confidence interval (CI) 0.65-0.68] compared with disease clusters (C-index 0.54, 95% CI 0.53-0.56), and Charlson score (C-index 0.58, 95% CI 0.56-0.59). Adding multi-morbidity (NRI 0.23, 95% CI 0.14-0.31) or Charlson score (NRI 0.13, 95% CI 0.03-0.20) to functional cluster model slightly improved the accuracy of prediction.Conclusions and Implications: Functional impairment may better predict prognosis compared with multimorbidity, which may be relevant to optimally address individuals' needs and to design tailored preventive inte