Sex Differences in Predictors of Frailty among Community-Dwelling Older Adults: Machine Learning Approach

Abstract

The identification of frail older adults is a critical public health challenge and a growing priority in geriatric research. Nevertheless, the identification of sex-specific predictors of frailty remains an unsolved challenge. The objective of this study is to examine sex differences in the predictors of physical frailty among community-dwelling older adults in Chile. A nationally representative sample of 9,123 individuals aged 60 and older (n = 5,363 women; n = 3,760 men) from the National Disability and Dependency Survey was analyzed. Supervised machine learning regression models were applied to identify the most relevant predictors of physical frailty. In both male and female subjects, the predominant predictor was the number of health conditions present. For men, physical activity emerged as the second most salient factor, while for women, engagement in leisure and recreational activities demonstrated a more pronounced correlation with frailty. The findings of the study demonstrated that age was a significant predictor in both male and female subjects. These findings offer valuable insights for developing gender-responsive strategies to prevent or delay the progression of physical frailty among older adults in Chile.