Racial diversity in abortion education and intention to provide abortion in postresidency practice: data from the Ryan Program.

Background: In 2011, 38% of US reproductive-aged women lived in the 89% of counties with no abortion provider. Physicians from racial and ethnic minority backgrounds (black, Latino, Native American, and Asian American) are more likely than white physicians to practice in underserved areas and serve patients who are poor or minorities. Abortion patients are racially diverse. However, we know little about racial and ethnic makeup of abortion providers and the differences in physicians' interest in providing abortions.

Objective: The objective of the study was to examine racial differences in participation in abortion training and intention to provide abortion in postresidency practice.

Study Design: This is a cross-sectional study of Ryan Program residents after completing a family-planning rotation. The Ryan Program supports obstetrics-gynecology residency programs to incorporate routine abortion care into training. Since 2003 the Ryan Residency Program has administered postrotation resident surveys, and race/ethnicity was added in 2015. We assessed correlates of intention to provide abortion, specifically comparing minorities with whites and whether training participation varied by race. We conducted a modified mediation analysis to assess the role of potential mediators in the relationship between race and intention to provide abortion.

Results: A total of 777 residents (79.0%) responded from September 2015 through August 2018. The proportions were as follows: 64.9% white, 8.5% black, 4.1% Hispanic/Latino, 18.8% Asian, and 3.8% as other. Overall, 56.9% intended to provide abortion for all indications and 82.4% for pregnancy complications. In a univariate analysis, Asian residents were significantly more likely to intend to provide abortions for all indications compared with white residents (68.4% vs 56.0%, odds ratio, 1.69, confidence interval, 1.13-2.53). This difference was not significant when controlling for religiosity and abortion attitudes. Religiosity (odds ratio, 0.60, confidence interval, 0.47-0.77) and abortion attitude (odds ratio, 3.32, confidence interval, 2.48-4.44) were significantly correlated with intention to provide abortion for nonmedical indications after residency. In a modified mediation analysis, the relationship between race and intention to provide was mediated by religiosity for black residents and abortion attitude for Asian residents. There was no difference in participation in abortion training by race/ethnicity.

Conclusion: Racial differences in intention to provide abortion in postresidency practice are mediated by religiosity and abortion attitude. Better understanding the intricate relationships between race, religiosity, participation in training, and future practice will allow us to improve abortion training while paving the way to support a more diverse abortion provider workforce.