Predicting frequent emergency department visits among children with asthma using EHR data

Area of Science:

• Pediatric Asthma Management
• Health Informatics
• Predictive Analytics in Healthcare

Background:

• Emergency department (ED) visits for pediatric asthma are frequent, costly, and often preventable.
• Predicting these ED visits using electronic health record (EHR) data is an understudied area.

Purpose of the Study:

• To investigate the predictability of frequent ED use in children with asthma.
• To identify key patient factors from EHR data that predict future ED visits.

Main Methods:

• Retrospective cohort study using EHR data from 2691 children with asthma.
• Bivariate and multivariable statistical analyses were performed.
• Machine learning algorithms were applied and compared for predictive accuracy.

Main Results:

• Prior ED use and insurance status were strong predictors of future frequent ED visits.
• Publicly insured children had 3.1 times higher odds of frequent ED use compared to privately insured children.
• A logistic regression model incorporating prior ED visits and insurance status achieved an AUC of 0.86.

Conclusions:

• Prior frequent ED use and insurance status are significant predictors of future ED utilization in children with asthma.
• These findings can inform targeted interventions to reduce avoidable ED visits.