Prediction of Death with the Frail-NH in Institutionalized Older Adults: A Longitudinal Study from a Middle-Income Country

Abstract

Background/Objectives Frailty is common in nursing homes. However, few studies reported longitudinal validation for death prediction or cut-off scores with the FRAIL-NH, which is designed to be used in nursing homes. Moreover, no studies came from Latin America, where frailty is highly prevalent. Our objectives were to evaluate (1) the prevalence of frailty according to the FRAIL-NH scale, and (2) its association to and the best cut-off score for predicting death after 12 months. Design Longitudinal study with 12-month follow-up. Setting 6 nursing homes in southwest of Brazil. Participants 293 residents with 60 years old or more. Methods Frailty was evaluated through the FRAIL-NH scale. Logistic regression was used to estimate the associated between frailty and mortality adjusted for age and sex. ROC curve was used to evaluate the accuracy of the scale for mortality prediction. Results Frailty was prevalent (47.4%) and was associated with death (odds ratio=1.31, 95% confidence interval [CI]=1.18-1.48, p<0.001). The area under the curve was 0.741 (95%CI=0.68-0.79). The sensitivity and specificity of the FRAIL-NH scale according to the best value of the Youden Index was 72.9% and 66.5%, respectively, for a cut-off > 8 points. Conclusions Frailty is prevalent in nursing homes according to the FRAIL-NH and it was associated with one-year prediction of death for a cut-off >8 points.