Race and ethnicity data missingness in hepatobiliary and pancreatic surgery: prevalence and outcomes.

BACKGROUND

Racial and ethnic disparities in surgical outcomes are well-documented across various specialties, including hepatopancreatobiliary (HPB) surgery. Although efforts to address these inequities have expanded, an underexplored issue is the growing prevalence of missing race and ethnicity data in national surgical registries, such as the American College of Surgeons (ACS) National Surgical Quality Improvement Program (NSQIP). This study aimed to evaluate the prevalence of missing race and ethnicity data in HPB procedures within the NSQIP and to determine whether such missingness is associated with adverse postoperative outcomes.

METHODS

A retrospective analysis of the ACS NSQIP database of patients who underwent HPB procedures from 2019 to 2023 was conducted. Patients were stratified into 3 groups based on the completeness of their demographic data: (i) race known vs race missing (RM), (ii) ethnicity known vs ethnicity missing (EM), and (iii) race or ethnicity known vs both missing (BM). The primary outcomes included any postoperative complication, unplanned reoperation, and 30-day mortality. Multivariate logistic regression models were used to assess the associations between BM status and adverse outcomes, adjusting for relevant demographic and procedural covariates.

RESULTS

Among 72,069 patients who underwent HPB procedures, pancreatic procedures were most common (50.3%), followed by liver procedures (47.7%) and bile duct procedures (2.0%). Of the total cohort, 13,106 (18.2%) were classified as RM, 11,256 (15.6%) were classified as EM, and 10,256 (14.2%) were classified as BM. Over the study period, the proportion of patients in the BM group remained relatively stable, ranging from 18.5% to 23.5%. Compared with patients with complete or partial data, those with BM data were slightly older (mean age: 62.3 vs 63.5 years, respectively; P <.001) and had a lower body mass index (28.2 vs 27.5 kg/m2, respectively; P <.001). In addition, compared with patients with complete or partial data, those with BM data had greater unadjusted rates of complications (38.1% vs 42.9%, respectively; P <.001), unplanned reoperation (3.8% vs 5.0%, respectively; P <.001), or mortality (1.7% vs 2.0%, respectively; P =.053). In multivariate analysis, BM status was associated with increased odds of postoperative complications (odds ratio [OR], 1.341 [95% CI, 1.283-1.402]; P <.001), unplanned reoperation (OR, 1.403 [95% CI, 1.268-1.522]; P <.001), and mortality (OR, 1.194 [95% CI, 1.019-1.400]; P =.029).

CONCLUSION

A substantial proportion of HPB procedures in the NSQIP database are affected by missing race and ethnicity data, with the prevalence of BM status remaining consistent over the study period. The BM group demonstrated distinct clinical and demographic characteristics. Although unadjusted outcomes seemed favorable, BM status was independently associated with higher odds of complications, unplanned reoperation, and mortality. These findings suggest the need for further investigation into the underlying causes of missing demographic data and emphasize the importance of addressing data missingness in efforts to promote health equity in surgical care.