Multimorbidity measures were poor predictors of adverse events in patients aged ≥80 years: a prospective cohort study

OBJECTIVES: To assess and compare the ability of two measures of multimorbidity and a simple disease count (DC) to predict health outcomes in a population of patients aged ≥80 years. STUDY DESIGN AND SETTING: A prospective, observational, and population-based cohort study including 567 individuals [3.0 years (standard deviation ± 0.25) follow-up]. RESULTS: Of the patients, 37.6% were reported with five or more diseases. Multimorbidity was measured by means of a modified Charlson comorbidity index [mCCI; median score, 5 (range, 4-15)], Cumulative Illness Rating Scale [CIRS; median score, 4 (range, 1-11)], and a simple DC of 22 selected chronic conditions [median score, 4 (range, 0-13)]. All measures were independently related to mortality [adjusted hazard ratio (HR) mCCI, 2.5 (confidence interval {CI}: 1.5, 4.1); CIRS, 2.1 (CI: 1.4, 3.2); DC, 2.1 (CI: 1.4, 3.2)] and hospitalization [adjusted HR DC, 2.3 (CI: 1.7, 3.1); mCCI, 2.1 (CI: 1.5, 3.0), CIRS, 1.9 (CI: 1.5, 2.6)] but not to functional decline. Areas under the curve for mortality and hospitalization were all below 0.70. Net reclassification improvements did not indicate that any one measure provided a significant benefit over the others. CONCLUSION: In this population, the mCCI, CIRS, and unweighted DC predicted mortality and hospitalization but not functional decline. There is no clear advantage of using one measure over another.