RESUMO
INTRODUCTION: To compare the validity of the Barthel (BI) and Charlson comorbidity index (CCI) as well as the age-adjusted CCI for the prognosis of mortality and institutionalization in hospitalized geriatric patients. MATERIAL AND METHODS: A retrospective study was conducted to evaluate patients admitted to an acute care geriatric ward and mid-term care facility in 2006 (n=343). The independent variables were the admission BI (registered in the clinical history in all patients) and the CCI and the age-adjusted CCI, which were registered "a posteriori" using the available information. The outcome variable was mortality or institutionalization when the patient was discharged. The sensitivity and specificity values for each of the cut-off points selected from the scales analyzed were recorded and ROC curves were constructed for each of these points; the area under the curve (AUC) with a confidence interval (CI) of up to 95% was calculated. RESULTS: The mean age was 82.3 years (58.6% women). The AUC was 0.736 (95% CI=0.68-0.79) for the BI, 0.61 (95% CI=0.55-0.67) for the CCI and 0.625 (95% CI=0.56-0.69) for the age-adjusted CCI. Statistically significant differences were found among the AUC of the BI compared with that of the other two indexes (p<0.01). CONCLUSIONS: As predictors of mortality and institutionalization, the BI was superior to the CCI and the age-adjusted CCI. The BI could therefore be more useful than the other two indexes when considering an adequate use of healthcare services.