Prediction for the Risk of Multiple Chronic Conditions Among Working Population in the United States With Machine...

Area of Science:

• Health Informatics
• Machine Learning in Healthcare
• Public Health

Background:

• Chronic diseases are a leading cause of morbidity, mortality, and healthcare costs.
• Multiple chronic conditions (MCC) pose a significant challenge in healthcare management.
• Limited research exists on predicting MCC risk within the working population.

Purpose of the Study:

• To develop and validate machine learning models for predicting the risk of MCC in working individuals.
• To identify effective algorithms for early detection and risk stratification of MCC.
• To support healthcare practitioners in managing chronic diseases within the workforce.

Main Methods:

• Utilized checkup data from 451,425 working individuals.
• Developed and evaluated seven distinct machine learning algorithms.
• Validated model performance using Area Under the Curve (AUC) metrics.

Main Results:

• All seven machine learning models demonstrated satisfactory predictive performance (AUC range: 0.826–0.850).
• The gradient boosting tree model achieved the highest AUC of 0.850.
• The models effectively identified individuals at risk for MCC.

Conclusions:

• The developed risk prediction model shows potential for automated, real-time diagnosis.
• The model can efficiently support healthcare practitioners in targeting high-risk individuals.
• Proactive strategies for chronic disease prevention and progression delay can be tailored based on these predictions.