Predictors of two-year post-hospitalization mortality among elderly veterans in a study evaluating a geriatric consultation team.

OBJECTIVE: To determine predictors of 2-year post-hospitalization mortality in a cohort of elderly hospitalized patients originally assembled to assess the impact of a Geriatric Consultation Team (GCT). DESIGN: Two-year follow-up of an inception cohort. SETTING: University-affiliated tertiary care VA Medical Center. PATIENTS: One hundred sixty-seven veterans age 75 or older discharged following hospitalization on medical, surgical, or psychiatry services but not intensive care units. INTERVENTION: None specifically studied here though cohort was previously part of randomized control trial of a Geriatric Consultation Team. MEASUREMENT: Mortality during 2 years of post-hospitalization follow-up. RESULTS: Two-year post-hospitalization mortality was 28 percent with no difference between the original GCT and control groups. For the entire sample, age, mental status, admission or discharge ADLs (but not change in ADL status), number of admission problems, number of discharge diagnoses, and discharge site were significant predictors of mortality in univariate analysis. Only discharge ADLs and discharge site remained significant in multivariate analysis. CONCLUSION: Measures of ADLs during hospitalization are stronger predictors of mortality following hospitalization than disease diagnoses. Impaired ADLs and placement other than at home are significant predictors of mortality, suggesting that the decision for nursing home placement contains other independently predictive information within it and/or that the subsequent nursing home period produces excess mortality. As had been indicated in short-term follow-up, there was no survival advantage for the Geriatric Consultation Group.

Duke Scholars

Author Harvey Jay Cohen Medicine, Geriatrics and Palliative Care

Author Gregory P. Samsa Biostatistics & Bioinformatics, Division of Biostatistics