Predicting Low-Resource-Intensity Emergency Department Visits in Children

Area of Science:

• Health Services Research
• Pediatric Health
• Healthcare Analytics

Background:

• Frequent emergency department (ED) use by children presents a challenge for healthcare resource allocation.
• Predicting high-frequency, low-resource-intensity (HF-LRI) ED visits is difficult, limiting targeted interventions.

Purpose of the Study:

• To develop a population-based model to predict Medicaid-insured children at risk for HF-LRI ED visits.
• To identify key predictors of frequent low-value ED utilization in pediatric populations.

Main Methods:

• Retrospective cohort analysis of Medicaid-insured children (1-18 years) from the MarketScan database (2013).
• Defined LRI visits as ED encounters without hospitalization, procedures, or advanced diagnostics.
• Developed and validated a generalized linear regression model using split-sample validation.

Main Results:

• 5% of children accounted for 21% of all LRI ED visits.
• Prior ED visits (≥2) and ≥3 chronic conditions were strong predictors of future HF-LRI ED use.
• The predictive model achieved an area under the curve of 0.74.

Conclusions:

• Demographic factors and prior ED utilization patterns can identify children at risk for future HF-LRI ED visits.
• Interventions targeting high-risk children can help reduce low-value ED utilization.