Analysis of Health Insurance Big Data for Early Detection of Disabilities: Algorithm Development and Validation

Area of Science:

• Pediatric Health
• Data Science in Healthcare
• Disability Prevention

Background:

• Early detection of developmental delays is critical for effective disability treatment.
• Timely intervention significantly improves outcomes for children with disabilities.

Purpose of the Study:

• To explore the potential of using big data from health insurance databases for early detection of childhood developmental delays.
• To identify children at risk of developing disabilities before formal clinical diagnosis.

Main Methods:

• Analysis of a large dataset (n=2412) of children up to 13 years from the Korea National Health Insurance Service Sample Cohort 2.0 DB.
• Utilized a tree-based model to select important features across disability categories (physical, brain lesion, visual, hearing, other).
• Employed multiple classification algorithms to identify the optimal predictive model for different age groups.

Main Results:

• A disability detection model achieved significant accuracy in identifying disabilities by age 4.
• This age is approximately one year earlier than the average clinical diagnosis age of 4.99 years.
• The study identified the earliest age range for clinically significant performance in early detection.

Conclusions:

• Big data analysis facilitates earlier identification of childhood disabilities compared to traditional clinical diagnoses.
• This early detection capability enables timely interventions to prevent or mitigate disabilities.
• The findings support the use of health insurance big data for proactive child development monitoring.