Artificial Intelligence-Assisted Occupational Disease Surveillance: Development and Validation of a Machine Learning Framework for Nigerian Healthcare Settings

Occupational disease surveillance in Nigeria is constrained by fragmented data systems, inadequate trained personnel, and inconsistent diagnostic coding, representing critical barriers to evidence-based policy. Artificial intelligence and machine learning offer transformative potential for automated occupational disease detection from existing health facility data. This doctoral study developed and validated a machine learning-based occupational disease surveillance framework for Nigerian healthcare settings. A multi-phase design was implemented. Phase One systematically integrated 12 years of electronic health record data from 15 tertiary and secondary hospitals across six geopolitical zones into a harmonized occupational health data warehouse. Phase Two developed natural language processing algorithms for occupational disease identification from clinical narratives, and supervised machine learning classifiers for disease-work exposure attribution. Phase Three validated the framework against expert occupational physician review and against the Nigerian national disease notification registry. The optimized ensemble model achieved a sensitivity of 87.3 percent and specificity of 91.6 percent for occupational disease identification, substantially outperforming physician-only documentation systems. Silicosis, noise-induced hearing loss, and occupational asthma were the most successfully detected conditions. Geospatial dashboard integration enabled real-time cluster detection. The study delivers an original AI surveillance infrastructure prototype adapted to Nigerian health data realities and recommends piloting through a National Centre for Disease Control partnership, integration into health information management systems, and investment in standardized occupational diagnosis coding training. Keywords: artificial intelligence, occupational disease surveillance, machine learning, health informatics, Nigeria.