Abstract
Occupational health has become an increasingly important dimension of public health, particularly in efforts to prevent chronic disease; however, the application of artificial intelligence (AI) in this area is constrained by the lack of standardized occupational exposure metrics, limited representation of diverse work environments, and fragmented data systems. This study aimed to evaluate the impact of occupational exposure variables on chronic disease risk prediction, assess the performance of Random Forest, XGBoost, and Deep Neural Network (DNN) models across workplace contexts, and propose a framework for interoperable platforms that integrate health and occupational data to strengthen predictive analytics and early diagnostics. Using a cross-sectional dataset of 5,000 workers from the manufacturing, agriculture, healthcare, and service sectors, the study analyzed demographic characteristics, clinical biomarkers, occupational exposure logs, and psychosocial assessments. Model performance was evaluated using ROC-AUC, precision, recall, and F1-score, while feature importance analysis quantified the contribution of occupational variables; in addition, a prototype interoperable platform was developed to demonstrate real-time integration between electronic health records and workplace monitoring systems. The findings showed that the DNN model outperformed the other algorithms, achieving a ROC-AUC of 0.89, precision of 0.85, recall of 0.88, and F1-score of 0.86. Occupational exposure variables contributed 27% to predictive power, with chemical exposure and psychosocial stressors showing the strongest associations with chronic disease markers. Among high-risk individuals, 54% were identified with subclinical conditions, including elevated C-reactive protein and HbA1c levels, while personalized interventions based on model outputs reduced risk scores by 22% and improved biometric indicators. The interoperable platform also successfully synchronized health and exposure data, enabling real-time analytics and targeted alerts. These findings demonstrate that integrating standardized occupational exposure metrics with interoperable data platforms substantially enhances the accuracy and practical utility of AI-driven chronic disease prediction, while supporting more equitable and proactive occupational health surveillance across diverse industries.