Introduction:
Seafarers face unique health challenges due to exposure to infectious diseases and occupational hazards, compounded by their extended time at sea¹. The maritime industry’s critical role in global trade necessitates enhanced health monitoring and disease prevention methods^1,2. This study aims to apply supervised Machine Learning (ML) models³ to predict the incidence of infectious diseases and improve occupational health among seafarers, based on health records and voyage data.

Methods:
We utilized health records and voyage details from the International Radio Medical Centre (C.I.R.M.) in Rome. Supervised ML models were trained on these data sets to predict disease incidence and provide actionable insights for prevention. We compared the performance of four different ML models used to predict the incidence of infectious diseases among seafarers: Random Forest, Support Vector Machine (SVM), Convolutional Neural Network (CNN), and Gradient Boosting.

Results:
Among the models, CNN showed the highest overall performance with an accuracy of 91%, a precision of 92%, and the highest AUC-ROC of 0.94, indicating that it is the most effective at predicting disease incidence. Random Forest also performed well, with slightly lower but competitive scores across all metrics. These results highlight the potential of machine learning in disease prediction, with the Neural Network model emerging as the most robust.

Conclusions:
These ML models demonstrated a significant ability to forecast disease outbreaks among seafarers, allowing for timely intervention⁴. These predictive models hold the potential to integrate seamlessly into maritime health protocols, reducing disease incidence and enhancing overall seafarer well-being. The use of ML models for disease prediction in the maritime industry can revolutionize public health strategies. By integrating these models into existing health protocols, we can reduce infectious disease outbreaks and occupational health issues in seafarers.