Observational study reveals triage acuity and trauma mechanisms drive length of stay in pediatric forensic emergency cases, highlighting targets for improved resource allocation.
Key Points
To identify operational and clinical predictors of increased emergency department length of stay and characterize clinical profiles in pediatric forensic cases.
Analyzed 1,113 pediatric forensic presentations in a retrospective observational cohort.
Identified time-to-discharge predictors using Cox proportional hazards regression as the primary model.
Applied Firth's penalized logistic regression and Random Forest algorithms to evaluate binary predictors for a data-derived 45-minute high-acuity stay threshold.
Lower clinical acuity reflected by higher Emergency Severity Index scores expedited discharge (HR: 1.954), whereas intoxication (HR: 0.126), burns (HR: 0.127), and out-of-vehicle traffic accidents (HR: 0.237) caused substantial delays.
Arrival accompanied by a parent was independently associated with faster discharge (HR: 1.303), while ambulance arrival did not significantly influence time-to-discharge (p = 0.176).
Random Forest identified Emergency Severity Index as the most important operational feature (37.0% importance; test AUC = 0.906), achieving a negative predictive value of 96.9%.