KMAN Publication Inc. (KMANPUB) Psychology of Woman Journal 3041-8615 7 Serial Number 26 2026 04 01 Predicting Adherence to Complex Health Regimens in Women with Chronic Illness: A Random Forest Approach Integrating Health Literacy and Self-Efficacy Predicting Adherence to Complex Health Regimens in Women with Chronic Illness: A Random Forest Approach Integrating Health Literacy and Self-Efficacy 1 11 10.61838/kman.pwj.5319 EN Journal Article 2026 01 12 <p><strong>Objective:</strong> This study aimed to predict adherence to complex health regimens in women with chronic illnesses by evaluating the integrated predictive value of health literacy, self-efficacy, and sociodemographic factors using a Random Forest machine learning approach.</p> <p><strong>Methods and Materials:</strong> A multicenter cross-sectional study was conducted among 482 women with chronic illnesses recruited from referral hospitals in Jakarta and Surabaya, Indonesia. Data were collected via structured interviews utilizing validated tools, including the General Medication Adherence Scale ( ), the Health Literacy Questionnaire ( ), and the Chronic Disease Self-Efficacy Scale ( ). A Random Forest regression algorithm, optimized via grid search hyperparameter tuning on a 70/30 train-test split, was utilized to evaluate predictive accuracy and rank feature importance.</p> <p><strong>Findings:</strong> The optimized Random Forest model (250 trees, maximum depth of 15) successfully explained 74% of the variance in regimen adherence ( ), demonstrating high accuracy with a Mean Absolute Error of 1.85 and a Root Mean Squared Error of 2.42. Feature importance analysis (measured by Gini impurity) revealed that the strongest predictors were self-efficacy in symptom control (0.185), health literacy in understanding information (0.142), self-efficacy in communicating with physicians (0.128), and health literacy in actively engaging with providers (0.115), heavily outperforming static sociodemographic factors such as age, income, and educational attainment.</p> <p><strong>Conclusion:</strong> Adherence to complex health regimens is predominantly driven by modifiable psychological and cognitive assets—specifically symptom-management self-efficacy and health information comprehension—rather than static sociodemographic characteristics.</p> https://journals.kmanpub.com/index.php/psywoman/article/download/5319/9679