Low-Value Computed Tomography for Children in the Emergency Department: A Repeated Cross-Sectional Study

Objectives: To compare low-value computed tomography (CT) use during pediatric emergency department (ED) visits by hospital type and physician specialty. Methods: Repeated cross-sectional study using linked databases from Ontario, Canada. We reviewed pediatric ED discharges from 2010 to 2019 for 5 diagnoses with recommendations against routine CT use: abdominal pain, constipation, concussion, seizure, and headache. We evaluated CT use by hospital type (pediatric academic, adult academic, community with and without pediatric consultation) and provider specialty pediatric emergency medicine (PEM), emergency medicine (EM), family medicine + EM, family medicine, pediatrician, using multivariable logistic regression, adjusting for patient, ED, and physician characteristics. Results: We included 599,948 pediatric ED discharges mean (SD) age 10.8 y (5.3); 55.4% females: 5000 (1.2%) discharges for abdominal diagnoses included a CT, and 21,398 (11.4%) discharges for neurological diagnoses included a CT. Children had an increased adjusted odds ratio aOR (95% CI) of receiving a CT at all hospital types compared with pediatric academic hospitals: adult academic hospitals ranging from 1.10 (1.01 to 1.21) for headache to 3.46 (1.89 to 6.36) for constipation, community hospitals with pediatric consultation ranging from 1.54 (1.45 to 1.63) for concussion to 3.74 (2.38-5.90) for constipation, and community hospitals without pediatric consultation ranging from 1.24 (1.15 to 1.33) for concussion to 2.29 (1.36 to 3.87) for constipation. Those patients seen by nonpediatric providers (EM, family medicine + EM, family medicine) were more likely to receive CT scans than PEM physicians for all diagnoses. Conclusions: Low-value CT use was higher among children treated in nonpediatric EDs and by nonpediatric providers. Improvement initiatives should target specific hospital types and specialties.