Abstract
OBJECTIVE
Patient complexity is defined in a variety of ways and is associated with health care costs and use. In this study, we evaluate whether pediatric inpatient complexity represents a single construct or multiple domains using multiple measures of diagnostic burden, health care use, and intensity.
METHODS
We analyzed administrative billing data from pediatric hospitals in the Pediatric Health Information System (PHIS) including medical and surgical hospitalizations discharged from 2021 to 2023 for patients aged 0 to 17 years with lengths of stay (LOS) of 30 days or fewer. We extracted measures of hospital use and clinical characteristics including LOS, readmissions, diagnosis and procedure codes, billing codes, and clinical classifications. For each discharge, we calculated measures capturing diagnostic burden, overall use and intensity, including unique diagnosis codes, unique items, services, and medications received, and complex chronic organ systems. We performed a principal components analysis (PCA) to examine patterns of shared variation and underlying dimensions among the included measures.
RESULTS
We included N = 2 090 139 discharges from 48 hospitals. Over three-quarters (79.5%) were medical admissions and 85% stayed less than 1 week. Our PCA identified 9 factors accounting for 74.1% of the variance. The first 3 factors comprised measures of diagnostic complexity (accounting for 20% of variance), use of medications, supply, and other items or services (17%), and procedural and overall use including LOS and intensive care unit use (10%).
CONCLUSIONS
Our findings indicate pediatric inpatient complexity is multidimensional, with 9 distinct factors identified by PCA that may have important implications for health services research, quality measurement, and risk adjustment.