Racial Differences in Do-Not-Resuscitate Orders among Pediatric Surgical Patients in the United States.

Very few studies have investigated the racial differences in do-not-resuscitate (DNR) orders in children, and these studies are limited to oncological cases. We aim to characterize the racial difference in DNR orders among U.S. pediatric surgical patients. We retrospectively evaluated the mortality of all children who underwent an inpatient surgery between 2012 and 2017 from the National Surgical Quality Improvement Program. We used log-binomial models to estimate the relative risk (RR) and 95% confidence interval (CI) of DNR use comparing white with African American (AA) children. To estimate the risk-adjusted difference in DNR orders, we controlled the analyses for age, prematurity status, emergent case status, American Society of Anesthesiologists class, year of operation, surgical specialty, and surgical complexity. Between 2012 and 2017, a total of 276,917 children underwent inpatient surgery, of whom 0.8% ( = 1601) died within 30 days of operation. Of the 1601 mortality cases, we retained 1212 children who were of either AA (26.0%,  = 350) or white (63.9%,  = 862) race. Most children were neonates, had an American Society of Anesthesiologists class ≥4 (70.0%,  = 811), and developed one or more postoperative complications (68.7%,  = 833). Overall, AA children were more likely to be neonates at the time of surgery (42.0% vs. 40.3%,  < 0.001), to be premature (66.3% vs. 49.0%,  < 0.001), and develop one or more postoperative complications (73.7% vs. 66.7%,  = 0.017). White children were three times more likely to have a DNR order than their AA peers (adjusted RR: 3.01, 95% CI: 1.09-8.56,  = 0.044). Among pediatric surgical patients in the United States, children of white race were three times more likely to have a DNR order in place than their AA peers despite the latter being "sicker" and more likely to develop postoperative complications. The mechanisms underlying this racial difference deserve further elucidation to improve shared decision making and goal-concordant care.