Predicting Care Needs Among Older Adults Using Explainable Machine Learning: A Multidimensional Approach Integrating Health, Social, and Environmental Factors (Preprint)

BACKGROUND Rapid population aging and a worsening shortage of care workers necessitates the identification of older adults who require proactive interventions. Although machine learning (ML) has been increasingly applied in gerontology, existing studies have predominantly focused on social isolation. OBJECTIVE This study aims to develop a robust predictive model for care needs using ML. Beyond physical health indicators such as disease and functional status, this study adopted a comprehensive approach including mental health, cognitive function, health behaviors, and socio–environmental determinants—such as social participation, social support, and housing environment—to present an integrated model encompassing both health and social care needs. METHODS Data were obtained from the 2023 Korea Senior Survey, a nationally representative sample of adults aged 60 and older. A set of candidate predictors was initially screened using the Light Gradient Boosting Machine (LightGBM) algorithm. Subsequently, we trained and compared 7 predictive models—logistic regression, decision tree, support vector machine, random forest, gradient boosting decision tree, XGBoost, and LightGBM. Model performance was evaluated using accuracy, sensitivity, specificity, F1 score, area under the receiver operating characteristic curve (AUC), and precision-recall AUC. Model interpretability was enhanced using Shapley Additive Explanations (SHAP) to quantify the contributions of individual predictors. RESULTS Among the evaluated models, XGBoost demonstrated the highest predictive performance. SHAP analysis revealed that limitations in instrumental activities of daily living (IADL) were the most influential predictors. Age, nutritional risk, depression, subjective health, work, cognitive function, home modification, and income emerged as the top predictors when functional status was excluded. These findings indicate that care needs are driven not only by functional impairment but also by distinct sociodemographic factors. CONCLUSIONS Explainable ML models offer high predictive accuracy and strong interpretability for identifying care needs among older adults. By highlighting the significant roles of health, social, and environmental factors, this study provides empirical evidence to support evidence-based decision-making for the Long-Term Care Insurance system and integrated community care policies.