Accuracy of two prognostic indexes to predict mortality in older adults with advanced dementia

Abstract

ABSTRACT. Dementia is a cause of disability among older adults. Accessing advanced dementia prognosis is a challenge. Objective: The objective of this study was to evaluate the accuracy of the Charlson and Carey indexes in predicting 3-year survival of older adults with advanced dementia. Methods: This is a retrospective cohort study of 238 patients aged ≥60 years with advanced dementia from an outpatient clinic and classified as stage ≥6A by using the Functional Assessment Staging scale. We excluded patients with missing data. We reviewed the semi-structured interview (clinical, sociodemographic, and functional data) from the baseline visit. This information was used to calculate 3-year mortality risks according to the Charlson and Carey indexes. We used Cox proportional hazard models to evaluate the associations of all-cause mortality with both indexes, adjusted for sociodemographic variables. We used Harrell’s C measure to determine the discrimination. We calculated the absolute differences between observed and predicted 3-year mortality risks for each index for calibration. Results: In 238 patients, the average age was 80.5±7.8 years, with 36% being men. The median follow-up time was 1.8 years (0.05–3.0). The 3-year all-cause mortality rate was 50% (119 deaths). The Carey index was associated with mortality, with one point increase related to a 15% increase in the mortality risk (hazard ratio [HR]=1.15, 95% confidence interval (95%CI) 1.06–1.25, p=0.001), even after adjustment. Accuracy for the Charlson index and Carey index was 0.55 (95%CI 0.49–0.60) and 0.60 (95%CI 0.52–0.62), respectively, with no difference between them (p=0.44). Conclusions: Both indexes had poor discrimination and calibration performances in predicting 3-year mortality in patients with advanced dementia.