(1492) Regional Variability of Drivers of Cost Among Common Conditions Treated in the PICU

Introduction: Pediatric Intensive Care Unit (PICU) costs represent a growing percentage of total hospitalization costs. We sought to determine drivers of cost for the top 5 PICU diagnoses and analyze how region may impact drivers of cost.

Methods: This is a retrospective cohort study using the Pediatric Health Information System (PHIS) database. We included patients < 18 years old admitted to the PICU from January 1, 2021 to December 31, 2023 with pneumonia, bronchiolitis, seizures, asthma, or diabetes mellitus. We excluded ECMO patients. Service lines included Clinical, Images, Lab, Other, Pharmacy, and Supply. Costs were calculated from PHIS charges using each hospital’s ratio of cost to charges and inflated to 2023 dollars. Generalized linear models were used to determine drivers of PICU costs, clustering by hospital to account for intra-group correlations within hospitals, and subsequently stratifying by region (Northeast, Midwest, South, West). Total cost was log-transformed prior to fitting all regression models.

Results: The study included 78,173 patients (pneumonia n=19,126, bronchiolitis n=18,129, seizures n=14,970, asthma n=13,901, diabetes n=12,047). Other costs (largely room costs) were the highest driver of PICU costs nationwide for patients with pneumonia, bronchiolitis, and asthma (p < 0.001). Specifically, every $100 increase in Other costs was associated with a range of 1.3-1.7% increase in total PICU cost. The highest drivers for diabetes and seizures were Pharmacy and Images (p < 0.001), respectively. For each diagnosis, stratification by region highlighted differences in drivers of cost. For pneumonia, Lab and Clinical costs were the highest drivers in the Northeast, Midwest, and West versus Other costs in the South (p < 0.001). For asthma and bronchiolitis, Supply costs were the highest drivers in the Northeast, Midwest, and West compared to Images and Pharmacy costs (p < 0.001) in the South. For seizures, Pharmacy and Lab costs were the highest drivers in the Northeast and South, respectively (p < 0.001).

Conclusions: This study found significant diagnosis related variability in drivers of cost. There was also significant regional variability for each diagnosis. Understanding this variation will uncover hospital-level and disease-specific processes that can be standardized across hospitals.