Bridging Clinical Prediction and Workplace Exposure: A Data-Driven Risk Stratification and Prediction Model for Occupational Respiratory Diseases.

Abstract

Abstract Occupational respiratory disease (ORD) surveillance remains largely reactive, with diagnosis often occurring only after clinically significant symptoms develop. Existing risk stratification tools are primarily clinically derived and rarely incorporate occupational exposure variables such as exposure intensity, duration, and workplace hazards, limiting early identification of at-risk workers. This review synthesizes current evidence on predictive modelling and risk stratification tools for respiratory diseases, evaluates their applicability to occupational health, and proposes a data-driven framework for proactive risk stratification. A structured literature search was conducted in PubMed, Scopus, and Web of Science for studies published between 2021 and 2025. Medical Subject Headings (MeSH) and free-text terms related to occupational respiratory diseases, predictive modelling, prognostic tools, and occupational exposure assessment were combined using Boolean operators. Extracted data included study characteristics, modelling approaches, predictors, and model performance metrics, which were narratively synthesized using thematic mapping. Of 618 records identified, six representative evidence approaches were retained for detailed comparative synthesis. Clinical models such as BODE and DOSE provided useful prognostic information but lacked occupational exposure parameters. Exposure-based approaches showed associations between cumulative exposure and respiratory outcomes, while a meta-analysis of machine-learning models for acute respiratory distress syndrome reported a pooled AUC of 0.7407. Key predictors included age, smoking status, lung function, and exposure-related factors. Current reactive and clinically focused approaches limit effective ORD prevention. Integrating occupational exposure variables into predictive, data-driven risk stratification frameworks is essential for early identification of high-risk workers. The proposed ORDiRS framework represents a step toward proactive occupational health surveillance in industrial settings.