Deviations From Recommended Drug Dosing in the Prehospital Treatment of Pediatric Patients

Pediatric medication safety is a priority area for research and quality improvement in emergency medical services (EMS). While medication errors may occur due to active failures, latent drivers of deviations from recommended dosages (DRD) are incompletely understood. We sought to understand the prevalence of DRDs and the factors associated with them.

We performed a retrospective cohort study of patients included in a quality improvement registry for prehospital pediatric medication safety. We included nine EMS agencies with uniform patient care protocols providing 9-1-1 services to the metropolitan area of a midsized city in the U.S. Our primary outcome was the prevalence of DRDs (doses ≥20% above or below protocolized doses). We identified DRDs by checking weight-based doses against weight-based and absolute thresholds. We assessed the prevalence of DRDs by medication, age, clinician impression, and administration route. We assessed associations between these factors and DRDs with logistic regression, then performed adjusted analyses with generalized linear mixed models. We assessed the relationship between weight estimation error and DRDs with a linear mixed model. We explored the potential effect of fixed-dosage formulations, full-vial doses, and volumetric limits of intranasal administration on DRDs.

We analyzed 386 medication administrations for 240 patients. The overall prevalence of DRDs was 46.4% (n = 179), and most were underdoses (n = 143, 79.9%). Among common medications, the number and prevalence of DRDs were: ondansetron (n = 29, 85.3%), midazolam (n = 35, 50.0%), fentanyl (n = 55, 45.5%), and epinephrine (n = 26, 44.8%). The DRDs were most common via the intranasal route (n = 62, 69.7%) and among school-aged children (n = 50, 55.6%). Ondansetron (aOR 36.9, 95%CI 6.3-217.6) and the intranasal route (aOR 7.3, 95%CI 2.9-18.6) were associated with higher odds of DRDs. We found no association between weight estimation error and DRDs. Exclusion of potential fixed-dosage formulations and full-vial doses was associated with fewer DRDs (81.6% vs. 42.5%). Only 2% of intranasal fentanyl DRDs could be explained by volumetric limits.

We observed DRDs to be common in the prehospital treatment of children, and they varied by medication and route. Possible latent sources of error include protocol design, medication formularies, and heuristics. The clinical significance of DRDs is uncertain.