Although regionalization of hospital health care has occurred, regionalization of perinatal care has declined. Earlier studies suggested that delivery at a high-volume, high-technology hospital reduced neonatal mortality, but little information is available on the effect of delivery hospital on other outcomes. This retrospective population-based cohort study examined all hospital-based deliveries in Pennsylvania, California, and Missouri to obtain unbiased measurements of the impact on mortality of delivering at high-volume, high-level neonatal intensive care units (NICUs) and compared non-NICU hospitals in these states with different systems of regionalization and different patient populations to examine common complications of premature birth.
Birth certificates were obtained for all deliveries occurring in Pennsylvania and California between 1995 and 2005 and Missouri between 1995 and 2003. These birth certificates were linked to death certificates, with greater than 98% of the linked records then matched to maternal and newborn hospital records. The primary infant cohort had a gestational age of 23 to 37 weeks and a birth weight of 400 to 8000 g. A secondary cohort of infants with a birth weight of 500 to 1500 g was used for comparison. The final cohort included 1,328,132 births. The primary outcome was in-hospital mortality, based on neonatal deaths (during the initial birth hospitalization) and fetal deaths with either a gestational age 23 weeks or greater or a birth weight 400 g or greater. The complications of premature birth included bronchopulmonary dysplasia (BPD), necrotizing enterocolitis, fungal sepsis, bacterial sepsis, and retinopathy of prematurity. Covariates were gestational age, birth weight, maternal sociodemographic factors, maternal residential zip code, sociodemographic information, maternal comorbid conditions, and 49 congenital anomalies grouped by affected organ system. A specialty hospital was defined as a level III facility that delivered a minimum of 50 very-low-birth-weight (VLBW) infants, on average, per year. The study used a matched-pair instrumental variables design and an instrumental variables approach to control for measured and unmeasured differences between hospitals.
Women who delivered at a high-level NICU were more likely to have a preexisting comorbid condition or a complication of pregnancy. In Pennsylvania, 79.8% of the pregnancies in the first quartile delivered at a high-level NICU compared with 23.9% in the fourth quartile. Similar instrument strengths were seen in California (79.6% vs 38.3%, respectively) and Missouri (55.7% vs 10.1%, respectively). In the unadjusted analysis, for infants at 23 to 30 weeks’ gestation, delivering at a high-level NICU was associated with higher mortality rates in all 3 states. After adjusting for measured and unmeasured case-mix differences, delivering at a high-level NICU was associated with lower in-hospital mortality rates. The risk differences ranged from 2.7 fewer deaths per 1000 deliveries in California to 12.6 fewer deaths per 1000 deliveries in Missouri. The risk ratios for in-hospital mortality at high-level NICUs ranged from 0.35 in Pennsylvania to 0.82 in California. Higher complication rates occurred at high-level NICUs regardless of state. Delivering at a high-level NICU in Missouri was associated with lower rates of BPD; Pennsylvania and California showed smaller, non–statistically significant changes. Rates of other complications were similar between the high-level NICU and other delivery hospitals, with the exception of infection rates, where the risk difference decreased from 5 to 45 extra infections at high-level NICUs per 1000 deliveries in unadjusted analyses to 0 to 14 cases per 1000 deliveries in adjusted analysis. For infants with birth weights of 500 to 1500 g, delivery at a high-level NICU was associated with lower mortality rates in all 3 states. Complication rates were similar between the 2 types of NICUs, except rates of BPD (lower in Missouri) and rates of bacterial sepsis (higher in Pennsylvania). The relative risk for the improvement in mortality in the 500- to 1500-g cohort was smaller for each state compared with the 23- to 37-week cohort. In the propensity score analysis, neonatal mortality rates at both types of hospitals were statistically similar in Missouri, Pennsylvania, and California. Higher rates were found for all complications at high-level NICUs regardless of state.
Determining the impact of a policy intervention such as perinatal regionalization is critical to accurately weighing the benefits and costs of this intervention. The survival benefit to delivering at a high-level NICU is larger than previously reported and appears to apply to both extremely and moderately preterm infants. The effect of delivery hospital may depend on the organization of perinatal services or the types of populations served. Assessments of perinatal policies that use only variables available in administrative databases may not adequately adjust for actual case-mix differences between hospitals.
Department of Pediatrics, Children’s Hospital of Philadelphia and University of Pennsylvania School of Medicine; Center for Outcomes Research, Children’s Hospital of Philadelphia; and Leonard Davis Institute of Health Economics, and Department of Statistics, Wharton School, University of Pennsylvania, Philadelphia, PA; and Department of Statistics, Stanford University, Stanford, CA