Self-Rated Health Associated with Chronic Diseases and Types of Physical Activity Among Middle-Aged and Older Adults: A Cross-Sectional Study Using Explainable Machine Learning

Background: Self-rated health (SRH) is a widely used indicator reflecting overall physical, psychological, and functional health status. Chronic diseases and physical activity (PA) are associated with SRH in middle-aged and older adults; however, the contribution of chronic disease burden and different types of PA to SRH requires further investigation. This study investigated the associations of chronic diseases and PA with SRH among middle-aged and older adults. Methods: This cross-sectional study included adults aged ≥40 years who participated in the 2024 Korea National Health and Nutrition Examination Survey (KNHANES), the ninth survey cycle. SRH was classified as poor or favorable based on self-reported health status. Sociodemographic characteristics, chronic diseases, and PA were evaluated using multivariable logistic regression analysis. The reported ORs were adjusted for age, sex, BMI, education level, household income, smoking, alcohol consumption, and employment status. Additional adjustment for other chronic diseases was applied in Model 3 of the overall analysis and in the subgroup analyses. Machine learning was additionally applied to evaluate the contribution of individual predictors to SRH. Model performance was evaluated using the Area Under the Receiver Operating Characteristic curve (AUROC) with five-fold cross-validation. Results: A total of 4079 participants were included, of whom approximately 58.1% were women, and 748 (18.3%) reported poor SRH. Participants with poor SRH were older, had lower income and education levels, and were less likely to meet the aerobic and muscle-strengthening PA guidelines than those with favorable SRH (all p < 0.001). All chronic diseases were significantly associated with poor SRH. Stroke (OR = 2.66, 95% CI = 1.74–4.06) and myocardial infarction or angina (OR = 2.69, 95% CI = 1.91–3.79) showed relatively high ORs, and the likelihood of poor SRH increased progressively with the number of chronic diseases. Participants with three or four chronic diseases had a 7.15-fold higher likelihood of reporting poor SRH than those without chronic diseases (95% CI = 4.63–11.04). Muscle-strengthening PA was related to lower odds of poor SRH (OR = 0.548, 95% CI: 0.435–0.690) and showed consistent inverse associations across most chronic disease groups, whereas aerobic PA showed weaker associations. The XGBoost model showed an AUROC of 0.631 based on five-fold cross-validation. SHAP analysis showed relatively greater contributions of chronic disease burden and household income to the prediction of poor SRH, while muscle-strengthening PA showed a greater contribution than the other PA variables. Conclusions: Chronic diseases and PA were associated with SRH among middle-aged and older adults. The likelihood of poor SRH increased with the number of chronic diseases, whereas muscle-strengthening PA was consistently associated with favorable SRH. These findings suggest that muscle-strengthening PA is associated with favorable SRH among middle-aged and older adults; however, longitudinal studies are needed to further examine this association.