Predicting Pneumoconiosis Risk in Coal Workers using Artificial Neural Networks

Area of Science:

• Occupational Medicine
• Data Science
• Public Health

Background:

• Pneumoconiosis poses a significant risk to coal workers' health.
• Accurate risk prediction is crucial for targeted interventions and health program development.

Purpose of the Study:

• To develop and validate an artificial neural network (ANN) model for predicting pneumoconiosis risk in coal workers.
• To identify key factors influencing pneumoconiosis development in this population.

Main Methods:

• An ANN model was constructed using health records from male coal workers.
• Input variables included age, employment history, occupational category, underground work duration, and smoking status.
• Output variables represented the presence or absence of pneumoconiosis.

Main Results:

• The ANN model accurately predicted pneumoconiosis risk with a 95.3% success rate.
• The model demonstrated high sensitivity (90.3%) and specificity (96.5%).
• Age and duration of employment in high-risk (group 1) jobs were the most significant predictors.

Conclusions:

• ANN models offer a powerful tool for estimating pneumoconiosis risk in coal miners.
• Integrating such models into occupational medicine can enhance miner health surveillance and preventive strategies.