Professor Mayo Clinic Rochester Rochester, Minnesota, United States
Introduction: Prior population-level research has suggested there to be disparities in the outcomes experienced by patients of different racial backgrounds undergoing surgical intervention. The role of different clinical and socioeconomic factors in mediating these disparities is unclear.
Objective: To estimate difference in risk of major complications and prolonged hospitalization among black and white patients undergoing brain tumor surgery, and to weigh the contribution of different clinical and socioeconomic factors to the disparities.
Methods: The National Inpatient Sample registry was queried for patients undergoing craniotomy for brain tumor from 2002 to 2019. Outcomes of interest were the occurrence of an in-hospital major complication/death, and prolonged hospitalization (defined as upper decile of the cohort). All risk estimates were adjusted for age and sex. Serial multivariable logistic regressions were performed to identify the attributable risk (AR) of each factor.
Results: Black patients experienced major complication/death at a rate 1.6 times higher than white patients (18.2 vs. 11.3%) and experienced prolonged hospitalization at a rate 2.03 times higher (19.4 vs. 9.5%). Our multivariable models were able to explain 68.4% of the discrepancy in the rate of major complication/death and 56.1% of the difference in the rate of prolonged hospitalization. Differences in comorbidities and admission status between black and white patients explained the majority of the difference in complication/death rates (AR=3.99%). Within the prolonged hospitalization model, they also explained 2.92% of the excessive risk, while hospital characteristics explained 1.11%, and payor and income quartile explained 0.91%.
Conclusion : The present data suggest differences in comorbidities and admission status (elective vs. non-elective) drive much of the disparities in outcomes seen between black and white patients following craniotomy for tumor. Socioeconomic differences between groups may also contribute significantly to prolonged hospitalization. While the present model fails to fully explain observed disparities, it identifies potential drivers that could serve as points for intervention.
