A Study on Multidimensional Predictors of Frailty in Community-Dwelling Older Adults: A Decision Tree-Based Approach

Abstract

This study aims to identify the key multidimensional predictors of frailty in community-dwelling older adults and propose strategies for frailty prevention and management at the community level. To achieve this, a decision tree analysis was conducted to explore the interactions among various frailty predictors based on a multidimensional frailty model and to identify key patterns that increase frailty risk. The study utilized data from the 2023 National Survey of Older Koreans, analyzing a total of 9,951 community-dwelling adults aged 65 and older. The results indicated that instrumental activities of daily living (IADL) were the most critical factor in distinguishing frailty status. Additionally, subjective health status, sleep quality, educational attainment, and nutritional status were identified as significant predictors of frailty among community-dwelling older adults. Based on these findings, this study proposes specific measures for frailty prevention in healthy older adults and early detection and intervention for pre-frail individuals. Furthermore, the findings can serve as foundational data for the development of a community-based integrated care system.