Hospital Cesarean Delivery Rankings by Overall, Low-Risk, and Reliability-Adjusted Metrics.

Importance: The Society for Maternal-Fetal Medicine (SMFM) low-risk cesarean delivery metric aims to improve comparability, but the extent to which additional patient-level adjustment and reliability profiling alter hospital rankings is not well characterized. Objectives: To evaluate how hospital cesarean delivery rankings differ across overall, SMFM-defined low-risk, and reliability-adjusted low-risk approaches, and to identify hospital characteristics associated with reliability-adjusted low-risk rates. Design, Setting, and Participants: This cross-sectional study used data from the 2022 National Inpatient Sample, a nationally representative all-payer database, on individuals aged 11 to 55 years who were hospitalized for childbirth. Data were analyzed from October 2025 through June 2026. Main Outcomes and Measures: Hospital cesarean delivery rates and rankings based on overall births, SMFM low-risk births (unadjusted), and empirical bayesian reliability-adjusted SMFM low-risk rates (mixed-effects logistic regression adjusting for age and obstetric comorbidities). Concordance was summarized by quartile cross-classification. Hospitals were classified as as-expected, lower-than-expected, or higher-than-expected on the basis of the 95% credible interval for the reliability-adjusted rate relative to the overall mean, and a hospital-level multivariable regression assessed characteristics associated with that rate. Results: Of the 330 977 SMFM low-risk deliveries (mean [SD] maternal age, 29.0 [5.7] years) across 1231 hospitals, 45 471 (13.7%) were cesarean. Mean (SD) hospital cesarean delivery rates were 32.1% (5.4%) for overall cesarean delivery, 13.5% (4.2%) for SMFM low-risk cesarean delivery, and 10.4% (2.4%) for reliability-adjusted SMFM low-risk cesarean delivery. Quartile concordance between overall and unadjusted SMFM low-risk rankings was modest (61.7% agreement in the lowest quartile), whereas concordance between SMFM low-risk and reliability-adjusted rankings was higher (85.7% agreement in the lowest quartile). Based on 95% credible intervals, 1024 hospitals (83.2%) were classified as having as-expected rates of cesarean delivery, 81 (6.6%) as having lower-than-expected rates of cesarean delivery, and 126 (10.2%) as having higher-than-expected rates of cesarean delivery. In multivariable analysis, Midwestern (β = -1.13 percentage points; 95% CI, -1.56 to -0.70 percentage points) and Western (β = -1.26 percentage points; 95% CI, -1.68 to -0.83 percentage points) hospitals had lower reliability-adjusted rates than Northeastern hospitals. Conclusions and Relevance: In this cross-sectional study of US hospitals, cesarean delivery rankings differed substantially between overall and SMFM low-risk denominators, highlighting limitations of overall rates for accountability. The high concordance between unadjusted and reliability-adjusted rankings suggests that the SMFM low-risk metric already captures most of the between-hospital signal. US region was the strongest correlate of reliability-adjusted rates.