BACKGROUND
Unexpected admissions of term neonates to the neonatal intensive care unit (NICU) and unexpected postnatal complications have been proposed as neonatal-focused quality metrics for intrapartum care. Previous studies have noted significant variation in overall hospital NICU admission rates; however, little is known about the influence of obstetric practices on these rates or whether variation among unanticipated admissions in low risk, term neonates can be attributed to systemic hospital practices.
OBJECTIVE
To examine the relative effects of patient characteristics and intrapartum events on unexpected NICU admissions, and to quantify the between-hospital variation in NICU admission rates among this group of neonates.
METHODS
We performed a retrospective cross-sectional study using data collected as part of the Consortium for Safe Labor study. Women who delivered term (≥37 weeks), singleton, non-anomalous, liveborn infants without an a priori risk for NICU admission were included. The primary outcome was NICU admission among this population. Multilevel mixed effect models were used to calculate adjusted odds ratios (aOR) for demographics (age, race, insurer), pregnancy characteristics (parity, gestational age, tobacco use, birth weight), maternal comorbidities (chronic and pregnancy-induced hypertension), hospital characteristics (delivery volume, hospital and NICU level, academic affiliation), and intrapartum events (prolonged second stage, induction of labor, trial of labor after cesarean delivery, chorioamnionitis, meconium-stained amniotic fluid, and abruption). Intraclass correlation coefficients were used to estimate the between-hospital variance in a series of hierarchical models.
RESULTS
Of the 143,951 infants meeting all patient and hospital inclusion criteria, 7,995 (5.6%) were admitted to the NICU after birth. In the fully adjusted model, the factors associated with the highest odds for NICU admission included: nulliparity (aOR 1.62 (95% confidence interval (CI) 1.53-1.71)), large for gestational age (aOR 1.59 (95% CI 1.47-1.71)), and small for gestational age (aOR 1.60 (95% CI 1.47-1.73)). Induction of labor (aOR 0.95 (95% CI 0.89-1.01)) was not associated with increased odds of NICU admission compared to women who labored spontaneously. The events associated with higher odds of NICU admission included: prolonged second stage (aOR 1.66 (95%CI 1.51-1.83)); chorioamnionitis (aOR 3.89 (95%CI 3.42-4.44)), meconium-stained amniotic fluid (aOR 1.96 (95%CI 1.82-2.10)), and abruption (aOR 2.64 (95%CI 2.16-3.21)). Compared to women who did not labor, the odds of NICU admission were lower for women who labored: aOR 0.48 (95%CI 0.45-0.52) for women with no uterine scar and aOR 0.83 (95%CI 0.73-0.94) for women with a uterine scar. There was significant variation in NICU admission rates by hospital, ranging from 2.9% to 11.2%. After accounting for case mix and hospital characteristics, the between-hospital variance was 1.9%, suggesting that little of the variation was explained by the effect of the hospital.
CONCLUSIONS
This study contributes to the currently limited understanding of term, NICU admission rates as a marker of obstetrical care quality. We demonstrated that significant variation exists in hospital "unexpected" NICU admission rates and that certain intrapartum events are associated with an increased risk for NICU admission after delivery. However, the between hospital variation was low. Unmeasured confounders and extrinsic factors, such as NICU bed availability, may limit the ability of unexpected term NICU admissions to meaningfully reflect obstetrical care quality.