Racial referral bias in cardiac computed tomography: Differences, disparities or discrimination?

Background: Disparities exist in medicine and can affect patient care. We sought to understand influences of racial biases in diagnostic testing within a Cardiac CT (CCT) population.

Methods: Race of CCT patients, referring physicians and the population in the catchment area were captured between February 2006 and November 2021. The frequency of CCT referrals for each race was indexed to the catchment population.

Results: Of 21,241 CCT patients, 17,514 (82.5 ​%) patients were White. The Non-White population was comprised of 467(2.2 ​%) Indigenous, 656(3.1 ​%) Black, 932(4.4 ​%) Asian, 276(1.3 ​%) South Asian, 1100(5.2 ​%) Middle Eastern and 296(1.4 ​%) Latin American races. The catchment population was 907,675, with 619,514 individuals of whom 69.7 ​% identified as White. Compared to the catchment population, there was a disproportionately higher referral rate for Whites than Non-Whites. The referral index for Whites was higher than Non-Whites (1.2 versus 0.6, p ​< ​0.001)). This pattern was consistent across all racial minorities and age categories. A total of 356 physicians (236(66.3 ​%) White, 4(1.2 ​%) Black, 39(12.0 ​%) Asian, 30(9.2 ​%) South Asian, 43(13.2 ​%), Middle Eastern and 4 (1.2 ​%) Latin American) made referrals to CCT. The racial difference in referral patterns was independent of physician race and was independent of their years in practice.

Conclusions: Racial differences exist in CCT referrals. These differences are independent of prevalence of disease, physician race or years in practice. This study supports the need to better understand reasons for disparity and strategies to mitigate potential bias.