Evaluating obstetrical residency programs using patient outcomes.

CONTEXT Patient outcomes have been used to assess the performance of hospitals and physicians; in contrast, residency programs have been compared based on nonclinical measures. OBJECTIVE To assess whether obstetrics and gynecology residency programs can be evaluated by the quality of care their alumni deliver. DESIGN, SETTING, AND PATIENTS A retrospective analysis of all Florida and New York obstetrical hospital discharges between 1992 and 2007, representing 4 906 169 deliveries performed by 4124 obstetricians from 107 US residency programs. MAIN OUTCOME MEASURES Nine measures of maternal complications from vaginal and cesarean births reflecting laceration, hemorrhage, and all other complications after vaginal delivery; hemorrhage, infection, and all other complications after cesarean delivery; and composites for vaginal and cesarean deliveries and for all deliveries regardless of mode. RESULTS Obstetricians' residency program was associated with substantial variation in maternal complication rates. Women treated by obstetricians trained in residency programs in the bottom quintile for risk-standardized major maternal complication rates had an adjusted complication rate of 13.6%, approximately one-third higher than the 10.3% adjusted rate for women treated by obstetricians from programs in the top quintile (absolute difference, 3.3%; 95% confidence interval, 2.8%-3.8%). The rankings of residency programs based on each of the 9 measures were similar. Adjustment for medical licensure examination scores did not substantially alter the program ranking. CONCLUSIONS Obstetrics and gynecology training programs can be ranked by the maternal complication rates of their graduates' patients. These rankings are stable across individual types of complications and are not associated with residents' licensing examination scores.