Traditionally, hospital perinatal quality and rankings have been based on cesarean rates among nulliparous, term, singleton, vertex patients, and hospitals have recently added unexpected term newborn complication rates as a separate outcome category. The drawbacks of this methodology are two-fold: first, maternal complications are not considered and second, the maternal-newborn outcomes, which may not be aligned with each other, are reported separately.

The objectives were to: 1) evaluate the relationships between cesarean, maternal and neonatal complication rates in nulliparous, term, singleton, vertex patients; 2) develop unified measures incorporating cesarean, maternal, and neonatal complications, utilizing desirability of outcome ranking methodology, to evaluate individual hospital performances; and 3) compare hospital rankings using the most desirable dyadic outcome “vaginal delivery with no maternal and no neonatal complications” to cesarean rate–based rankings for the overall population, as well as for low- and high-risk patients.

This retrospective cross-sectional study included all nulliparous, term, singleton, vertex deliveries at 7 hospitals of the Northwell Health system from January 2019 to December 2024. Maternal complications included “severe obstetric complications” as per the Joint Commission criteria. Neonatal complications included the conditions described by the Joint Commission as “unexpected complications in term newborns.” First, statistical analyses were performed to evaluate correlations among cesarean, maternal, and neonatal complication rates in the 7 hospitals. Second, we employed dyadic maternal-newborn outcomes using a desirability of outcome ranking integrating cesarean, maternal, and neonatal complication rates for each hospital. Third, we used the most desirable outcome, “vaginal delivery with no maternal and no neonatal complications” to derive a new seven-hospital ranking which was then compared to the cesarean rate–based ranking. The same comparisons of rankings were also performed after stratification of the data to low- and high-risk patients based on the obstetric comorbidity index score on admission (0–3 and ≥4, respectively).

A total of 55,841 nulliparous, term, singleton, vertex deliveries during the years 2019 to 2024 were analyzed. There was a significant negative correlation between cesarean and neonatal complication rates (r=−0.79, P =.04), and no correlations between cesarean vs maternal complication rates (r=−0.08, P =.86) or maternal complication vs neonatal complication rates (r=−0.33, P =.47) indicating the need for a combined metric. Based on the desirability of outcome ranking methodology, 4 groups of dyadic outcomes were formed: a) vaginal delivery with no maternal and no neonatal complications; b) cesarean with no maternal and no neonatal complications; c) vaginal delivery with maternal and/or neonatal complications; and d) cesarean with maternal and/or neonatal complications. The rates of dyadic outcomes were recorded for each hospital and the best possible dyadic outcome (vaginal delivery with no maternal and no neonatal complications) was used to create a new hospital ranking which was then compared with the (referent) cesarean rate-based ranking. There were significant changes in the overall ranking based on the new maternal-newborn dyadic measure: 2/7 (29%) hospitals changed ranking in the overall population (Kendall Tau 0.905, P =.002); 3/7 (43%) in the low-risk group (Kendall Tau 0.810, P =.01); and 5/7 (71%) in the high-risk group (Kendall Tau 0.714, P =.03).

The study demonstrates the need for a dyadic maternal-newborn perinatal quality measures that incorporates cesarean rates, maternal and neonatal complication rates. Our findings suggest that separate reliance on cesarean or neonatal complication rates may provide an inaccurate representation of perinatal care quality. We propose that our Northwell composite dyadic measure “vaginal delivery with no maternal and no neonatal complications” allows for a comprehensive assessment of quality of perinatal care.

(13.75 Mo)