Probabilidade de internações repetidas: estudo de validação preditiva de um instrumento de rastreamento de fragilidade em idosos

Abstract

The identification of older people at risk of frailty is a strategy for planning actions and optimizing resources according to their intrinsic functional profiles and allows their insertion into specialized programs. Boult et al. developed and validated the Probability of repeated admission (Pra) for the screening of vulnerable community-dwelling elderly. There is already a long history indicating the utility of the Pra, but the currently available evidence suggests that its utility should be reconsidered. The aim of this study was to evaluate the performance of the Pra for predicting healthcare utilization, functional dependence, and death, and to study the factors associated with the risk of repeated admission in a community-based sample of the elderly in Rio de Janeiro City, Brazil. This was a prospective longitudinal study of a sample of community-dwelling elderly who participated in the Frailty in Brazilian Older People Study research network in the Rio de Janeiro section. In this study, data was analyzed from the baseline and the follow-up. At the baseline, 764 elders were stratified by the Pra, and the data regarding health services utilization and the survival status was collected three years later. To analyze the tool s accuracy, the area under the curve was obtained, and a Kaplan-Meier model was used to estimate survival curves according to the risk strata. Cancer, falls, chronic obstructive pulmonary disease, use of medication, receiving a visit from a health professional, being bedridden at home, living alone, and level of activities of daily living were statistically associated with the risk of hospitalization. The accuracy ranged from 68% (CI 95% 63-73%) for death to 58% (CI 95% 50-66%) for home visits by health professionals. The sensitivity and specificity of the Pra ranged from 8.2% and 94.2%, for hospitalization in one year, to 19.2% and 95.5% for functional dependence. Survival curves showed that high-risk individuals were more likely to die when compared to their low-risk counterparts (p < 0.001). It can be considered that the Pra does not present predictive validity for the screening of frail elderly or frequent users of health services, in the northern area of the Rio de Janeiro city, mainly due to its low sensitivity and poor accuracy.