Clinical Fellow Boston Children's Hospital Brookline, Massachusetts, United States
Background: It is well-established that delivering very low birth weight (VLBW, < 1500g) infants at hospitals with high-level, high-volume (regionalized) neonatal intensive care units (NICUs) improves survival while increasing risk of complications of prematurity. Our past work found that minoritized neonates saw greater benefit from delivering at regionalized hospitals than non-Hispanic white neonates. We hypothesize that this is because of a greater difference in NICU quality between non-regionalized and regionalized centers at which minoritized neonates deliver. Our preliminary results comparing minoritized and non-Hispanic white neonates born in CA, MO, PA, and SC between 1996−2012 support this hypothesis. Objective: Using an expanded timeline and state cohort, we will describe differences in overall, non-regionalized, and regionalized delivery hospital NICU quality for non-Hispanic Black, Hispanic, AANHPI, and other race neonates versus non-Hispanic white neonates, using risk-adjusted pre-discharge mortality and morbidity rates as hospital quality metrics. Design/Methods: This study entails secondary analysis of existing vital statistics and patient discharge data. It has been deemed non-human subjects research by the IRBs of Children’s Hospital of Philadelphia and Beth Israel Deaconess Medical Center. Cohort includes VLBW infants born in CA, MI, OR, PA, and SC between 1996−2020 (anticipated n >100,000). Risk-adjusted pre-discharge mortality and morbidity rates—including bronchopulmonary dysplasia, retinopathy of prematurity, necrotizing enterocolitis, intraventricular hemorrhage, and infection—will be calculated for each hospital-year for all hospitals with ≥50 deliveries annually. Kruskal-Wallis equality-of-populations rank, Dunn’s multiple comparison tests, and regression analysis will be used to describe differences in distribution of delivery hospital NICU quality by maternal race/ethnicity, stratified by NICU regionalization status. Data cleaning is ongoing with plans to complete analysis by February 2024 and compile results for presentation and publication by March 2024.
