Variation in Long-Term Postoperative Mortality Risk by Race/Ethnicity After Major Non-cardiac Surgeries in the...

Few large sample studies have examined whether disparities, as measured by the proxy of race/ethnicity, are observed in long-term mortality after high-risk operations performed in a United States national health system. We compared operation year-related mortality risk by race/ethnicity after high-risk operative interventions among patients receiving care within the VHA.

From the Veterans Affairs Corporate Data Warehouse and Surgical Quality Improvement Program, data were retrieved for 426,695 patients undergoing high-risk surgical procedures in non-cardiac, general, vascular, thoracic, orthopedic, neurosurgery, and genitourinary specialties between 2000 and 2018. Operation year was used as a surrogate measure of advances in technology and perioperative management. Underrepresented race/ethnicity groups were compared in a binary form with Caucasian/White race, as the reference category. The primary outcome was time to mortality, defined as death occurring at any time, due to any cause, during follow up, and after the initial, eligible surgery.

The median follow-up after 537,448 operations among 426,695 patients was 4.8 years. After adjustment for preoperative risk factors and demographics, long-term mortality risk decreased significantly to a hazard ratio of 0.96 (95% confidence interval, 0.962 to 0.964) over calendar time. Long-term mortality was not significantly higher among African Americans/Blacks compared to Caucasians/Whites (p = 0.22). Among Hispanics, differences in mortality risk favored Caucasians/Whites in the early years under study-a difference that dissipated as time progressed. In the most recent years, no difference in mortality was observed among Asian/Native Americans and Caucasians/Whites.

Risk-adjusted long-term mortality after high-risk operations among Veterans Affairs hospitals did not significantly vary between African Americans/Blacks, Hispanics, and Asian/Native Americans groups.