OBJECTIVE: To estimate the contributions of patient and health care provider-hospital characteristics to the variation in the frequency of nulliparous, term, singleton, vertex cesarean delivery in a multi-institutional U.S. cohort. METHODS: We performed a secondary analysis of the multicenter Assessment of Perinatal Excellence cohort of 115,502 mother and neonatal pairs who were delivered at 25 hospitals between March 2008 and February 2011. Women met inclusion criteria if they were nulliparous and delivered a singleton in vertex presentation at term. Hospital ranks for nulliparous, term, singleton, vertex cesarean delivery frequency were determined after risk adjustment. The fraction of variation in nulliparous, term, singleton, vertex cesarean delivery frequency attributable to patient and health care provider-hospital characteristics was assessed using hierarchical logistic regression. RESULTS: Of the 115,502 deliveries in the initial cohort, 38,275 nulliparous, term, singleton, vertex deliveries met inclusion criteria. The median hospital nulliparous, term, singleton, vertex cesarean delivery frequency was 25.3% with a range from 15.0% to 35.2%. The majority of hospitals (16/25) changed rank quintiles after risk adjustment; overall the changes in rank were not statistically significant (P=.53). Patient characteristics accounted for 24% of the nulliparous, term, singleton, vertex cesarean delivery variation. The analyzed health care provider-hospital characteristics were not significantly associated with cesarean delivery frequency. CONCLUSION: Although patient characteristics accounted for some of the variation in nulliparous, term, singleton, vertex cesarean delivery frequency and accounting for case mix had implications for hospital cesarean delivery rankings, the majority of the variation was not explained by the characteristics evaluated. These findings emphasize the importance of continued efforts to understand aspects of obstetric care, including case mix, that contribute to cesarean delivery variation.