Unmasking ethnic disparities in Asian representation in U.S. oncology training programs.

Abstract

e18548 Background: Asian Americans are not traditionally considered underrepresented in oncology; however, prior analyses of oncology workforce representation have not disaggregated Asian race by ethnicity. This study assessed Asian representation in U.S. oncology training programs by country of origin. Methods: All trainees with self-reported race/ethnicity data in U.S. Radiation (Rad Onc), Medical (Med Onc), Pediatric (Peds Onc), and Surgical Oncology (Surg Onc) programs were identified using 2013-2022 data from the Association of American Medical Colleges. Complex Surgical, Gynecologic, and Orthopedic Oncology programs were included within the Surg Onc category. Asian race was disaggregated by available subgroups (i.e., Bangladeshi, Cambodian, Chinese, Filipino, Indian, Indonesian, Japanese, Korean, Pakistani, Taiwanese, and Vietnamese). Asians with > 1 or no subgroup classification were excluded. Representation quotients (RQ) were used to measure representation relative to the U.S. census population adjusted by age and year, whereby RQ = 1 denoted proportional representation; RQ > 1 and RQ < 1 denoted representation greater or less than expected; RQ = 0 denoted no representation. Linear regression assessed trends in RQ over time for each subgroup. Results: Overall, 27% of oncology trainees identified as Asian (31% Rad Onc, 30% Med Onc, 30% Peds Onc, 19% Peds Onc, 16% Surg Onc). Apart from Indonesians in Rad Onc (median [IQR] RQ, 1.21 [0-3.54]), RQ = 0 for all Indonesians, Cambodians, and Laotians, while median RQ was < 1 for all Filipinos, regardless of program type. In contrast, all other Asian ethnic subgroups had median RQ values > 1, except for Taiwanese and Vietnamese fellows in Surg Onc (0.84 [0.45-0.96]) and Peds Onc (0.43 [0.30-0.63]), respectively. Among Filipinos, median RQ values for Med Onc (0.39 [0.31-0.48]) and Peds Onc (0.51 [0.45-0.58]) were approximately half that of their respective primary residencies (Internal Medicine, 0.75 [0.75-0.73]; Pediatrics: 0.97 [0.99-91]). From 2013-2021, representation remained largely stable for all ethnic groups, with the exception of increased representation among Taiwanese Med Onc (RQ slope [95%CI], 2.96 [1.81, 4.11]; p< 0.001) and Indonesian Rad Onc (0.94 [0.53 to 1.35]; p= 0.001) trainees; and decreased representation among Indian Med Onc (-0.44 [-0.65, -0.23]; p= 0.002) and Korean Surg Onc (-0.04 [-0.61, -0.20], p= 0.003) trainees. No changes in Filipino representation were observed. Conclusions: In this first workforce analysis of oncology trainees to disaggregate Asian race by ethnicity, Cambodians, Filipinos, Indonesians, and Laotians were consistently underrepresented in oncology training programs. Further investigation is needed to understand these differences to ensure that efforts to promote diversity in oncology avoid overlooking ongoing imbalances in representation and inclusion in the U.S. physician workforce.