Objective: Diagnosis grouping systems are essential for health services and epidemiological and health outcomes research. The Diagnosis Grouping System (DGS) was developed to classify International Classification of Diseases, 10th Revision diagnosis codes for pediatric emergency department (ED) encounters. We applied and compared the DGS to large samples of US ED encounters for children using 3 large health system databases.
Methods: We conducted a retrospective cross-sectional study of children (younger than age 18 years) between 2016 and 2021 using 3 ED datasets: the National Hospital Ambulatory Medical Care Survey (NHAMCS; a complex survey of US ED encounters), COMPdata (a dataset of encounters for Illinois), and the Pediatric Health Information Systems (PHIS; an administrative database of US children's hospitals). We categorized principal diagnoses using the DGS and compared their distribution.
Results: We included 181 607 978 survey-weighted encounters from NHAMCS, 5 352 016 encounters from COMPdata, and 20 859 019 encounters from PHIS. A higher proportion of encounters in the PHIS sample were admitted or transferred (12.7%) compared with those in the NHAMCS (6.1%) or COMPdata (4.2%) samples. Trauma was the most common major group identified among all datasets (22.0% in NHAMCS, 22.2% in COMPdata, and 17.0% in PHIS). Some conditions, such as sickle cell anemia, diabetes mellitus, appendicitis, seizures, and asthma, presented more frequently among children presenting to the PHIS hospitals.
Conclusions: We identified comparable diagnostic distributions across different datasets, with differences that are likely attributable to specific characteristics of the data source. Our findings provide a baseline for pediatric diagnoses to facilitate future research on pediatric ED utilization.
Copyright © 2025 by the American Academy of Pediatrics.