Variation and racial/ethnic disparities in Cesarean delivery at New York City hospitals: the contribution of hospital-level factors

Abstract

We aimed to quantify general and specific contextual effects associated with Cesarean delivery at New York City hospitals, overall and by maternal race/ethnicity. Among 127,449 singleton, nulliparous births at New York City hospitals from 2015 to 2017, we used multilevel logistic regression to examine the association of hospital characteristics (public/private ownership, teaching status and delivery caseloads) with Cesarean delivery, overall, and by maternal race/ethnicity. We estimated the intra-class correlation to examine general contextual effects and 80% interval odds ratios (IOR) and percentage of opposed odds ratios (POOR) to examine specific contextual effects. Overall, 27.8% of births were Cesareans. The general contextual (hospital) effect on Cesarean delivery was small (intra-class correlation: 1.8%). Hospital characteristics associated with Cesarean delivery differed by maternal race/ethnicity, with delivery in teaching hospitals reducing the odds of Cesarean delivery among White (IOR: 0.31, 0.86; POOR: 4.7%) and Asian women (IOR: 0.41, 0.95; POOR: 7.3%), but not among Black (IOR: 0.51, 1.34; POOR: 30.7%) or Hispanic women (IOR: 0.44, 1.24; POOR: 22.6%). Hospital ownership and caseloads were not associated with Cesarean delivery for any group. There is little within-hospital clustering of Cesarean delivery, suggesting that Cesarean disparities may not be explained by hospital of delivery.