Comprehensive Health Assessment Using Risk Prediction for Multiple Diseases Based on Health Checkup Data

Area of Science:

• Preventive Medicine
• Health Informatics
• Biostatistics

Background:

• Comprehensive health status assessment is crucial for personalized prevention and treatment strategies.
• Existing tools may not adequately integrate multiple disease risk factors from routine checkups.

Purpose of the Study:

• To develop a suite of risk prediction models for multiple diseases using only health checkup data.
• To enable estimation of individual risk for cardiovascular, cerebrovascular, metabolic, liver, and kidney diseases.

Main Methods:

• Retrospective study utilizing health checkup data and electronic health records.
• Cox proportional hazard regression applied to develop multi-disease risk prediction models.
• Model performance evaluated using Area Under the Curve (AUC) with 4-year risk assessments.

Main Results:

• Models demonstrated strong predictive capabilities across various diseases.
• AUC values ranged from 0.50 (subarachnoid hemorrhage) to 0.92 (liver fibrosis).
• High AUCs observed for alcoholic liver disease (0.91), liver fibrosis (0.92), and type-2 diabetes mellitus (0.82).

Conclusions:

• A set of multi-disease risk prediction models can be developed using readily available health checkup data.
• These models support comprehensive individual health assessment.
• Facilitates personalized prevention strategies and early disease diagnosis.