Fig 2 Adjusted annual maternal complication rates (with 95% confidence areas) by physicians’ years of experience for all, cesarean, and vaginal deliveries. Estimates were generated from negative binomial regression models that controlled for state, secular trend, physicians’ annual cesarean delivery rate and total number of deliveries, patients’ comorbidities, demographics, weekend admission, and a physician level random effect. Confidence areas were further adjusted for multiple comparisons with Sidak’s method