Background: Surgical care is delivered around the clock. Elective cases within the Veterans Affairs health system starting after 4 pm appear to have an elevated risk of morbidity, but not mortality, compared with earlier cases. The relationship between operation start time and patient outcomes is not described in private-sector patients or for emergency cases. Study Design: We performed a retrospective cohort study of 56,920 general and vascular surgical procedures performed from October 2001 through September 2004, and entered into the National Surgical Quality Improvement Program database. Operation start time was the independent variable of interest. Random effects, hierarchical logistic regression models adjusted for patient, operative, and facility characteristics. Two independent models determined associations between start time and morbidity or mortality. Subset analysis was performed for emergency and nonemergency cases. Results: After adjustment for patient and procedure characteristics, mortality had a moderately strong association with start time, but only for nonemergency cases starting 9:30 pm to 7:30 am (odds ratio = 1.752; p = 0.028; reference 7:30 am to 9:30 am). As for morbidity, after adjustment, operations starting 9:30 am to 1:30 pm and 5:30 pm to 9:30 pm were associated with a weakly elevated risk of morbidity, but those starting 9:30 pm to 7:30 am demonstrated a strong effect on morbidity (odds ratio = 1.32; p < 0.0001). Subgroup analysis showed this effect was largely a result of elevated risk of morbidity in emergency cases from this overnight time period (odds ratio = 1.48; p = 0.001). Conclusions: Surgical start times are associated with risk-adjusted patient outcomes. In terms of facility operations management and resource allocation, consideration should be given to the capacity to accommodate cases with differences in risk during different time periods.