P2C134: Hospital Variation in Neonatal Intensive Care Utilization and Healthcare System Impact Among Late Preterm Newborns in Texas

Background: Late preterm infants (born 34-36 weeks’ gestational age [GA]) account for 73% of all liveborn preterm infants in the US. While variation in neonatal intensive care (NIC) for this population has been described, there exists little information on the impact of this variation on the healthcare system.

Methods: Retrospective population-based cohort of Texas Medicaid-insured infants born without congenital anomalies in Texas in 2010-2014. The study objective is to investigate among infants born 33-37 weeks’ gestation: 1) variation in NIC use, pre-discharge length of stay (LOS), and expenditures (Medicaid payments) for birth admission & infants’ 1st year of life; and 2) variation in LOS and expenditures attributable to variation in NIC use. Data are presented for 100 hospitals with largest Medicaid birth volume for each GA week. NICU admission, LOS, and Medicaid payments were determined from claims data linked with vital records and Texas Medicaid enrollment data. Dollar amounts were inflated to 2023 US$ based on the medical component of the consumer price index. We reported mean, median and interquartile ranges (IQR) for hospital rates of all outcomes. We used the intraclass correlation coefficient (ICC) to describe between-hospital variation in NIC use for each GA week. Multilevel mixed-effects linear models were used to estimate the ICCs, adjusting for birth weight, sex, singleton (yes/no), and within-center correlation.

Results: A total of 150,369 Medicaid-insured 33-37 weeks infants were identified. Among the 100 hospitals with the highest GA week-specific Medicaid birth volume, NIC use varied substantially by gestational age (Figure 1, Table 1). Nearly all hospitals provided NIC to nearly all infants born at 33 weeks (median hospital rate 97.8% [IQR: 95.7-100.0%]) and 34 weeks (median hospital rate 93.4% [87.3-97.0%]). There was substantial variation in NIC use at 35 weeks (median hospital rate 53.6% [42.7-68.1%]) and less at 36 weeks (median hospital 25.2% [20.2-32.2%]). At 35 weeks, NIC use explained 14.6% of between-hospital variation in LOS (median hospital 7.7 days [5.0-10.0]), 39.7% of birth admission expenditures (median hospital $10,917 [$8,428-$14,594]) and 33.2% of first-year expenditures ($15,694 [$12,747-$19,913]. Despite little variation in LOS for infants born at 36 and 37 weeks, NIC admission explained 59.8% and 50.9% of birth hospitalization expenditures and 1-year expenditures, respectively, at 36 weeks and 55.3% and 46.4% at 37 weeks.

Conclusion: Variation among hospitals in NIC use for late preterm infants, particularly at 35 and 36 weeks’ gestation, is substantial. Large proportions of observed between-hospital variation in LOS and expenditures were attributable to variations in NIC use. Further evidence is needed to support appropriate NIC use for infants born late preterm.