Predictors of applicant pool racial and ethnic diversity among physician assistant education programs: a national cross-sectional cohort study

Abstract

Background Numerous studies have demonstrated that the increasing racial and ethnic diversity of the US population benefits from access to healthcare providers from similarly diverse backgrounds. Physician assistant (PA) education programs have striven to increase the diversity of the profession, which is predominantly non-Hispanic white, by focusing on admitting students from historically excluded populations. However, strategies such as holistic admissions are predicated on the existence of racially and ethnically diverse applicant pools. While studies have examined correlates of matriculation into a medical education program, this study looks earlier in the pipeline and investigates whether applicant – not matriculant – pool diversity varies among PA programs with different characteristics. Methods Data were drawn from the 2017–2018 Central Application Service for PAs admissions cycle. Applications to programs with pre-professional tracks and applicants missing race/ethnicity data were excluded, resulting in data from 26,600 individuals who applied to 189 PA programs. We summarized the racial and ethnic diversity of each program’s applicant pools using: [1]the proportion of underrepresented minority (URM) students, [2]the proportion of students with backgrounds underrepresented in medicine (URiM), and [3]Simpson’s diversity index of a 7-category race/ethnicity combination. We used multiple regressions to model each diversity metric as a function of program characteristics including class size, accreditation status, type of institution, and other important features. Results Regardless of the demographic diversity metric examined, we found that applicant diversity was higher among provisionally accredited programs and those receiving more applications. We also identified trends suggesting that programs in more metropolitan areas were able to attract more diverse applicants. Programs that did not require the GRE were also able to attract more diverse applicants when considering the URM and SDI metrics, though results for URiM were not statistically significant. Conclusions Our findings provide insights into modifiable (e.g., GRE requirement) and non-modifiable (e.g., provisionally accredited) program characteristics that are associated with more demographically diverse applicant pools.