aamc.org does not support this web browser.
  • Viewpoints

    Rooting out implicit bias in admissions


    Editor’s note: The opinions expressed by the author do not necessarily reflect the opinions of the AAMC or its members.

    Jamal (not his real name) was nervous going into the medical school interview, but he had studied this school, reviewed his application, and aced several mock interviews. The interviewer, Professor Smith, MD, (also not his real name) was a veteran of the admissions committee and prided himself on being a compassionate physician and educator.

    After a cordial greeting, Smith began to ask the standard questions and evaluate Jamal’s answers. Good understanding of his research, check. Good answer to what he’d do if he couldn’t be a doctor, check. Things were going fine from the professor’s perspective, pretty typical.

    But Jamal noted that Smith did not look comfortable. Smith seemed to be cutting off Jamal’s answers, and making less eye contact than other interviewers. The usual small talk about sports, music, and hometown was omitted, and the conversation felt mechanical. Then it was over, and they were shaking hands. Both felt it was a standard interview, but … something was missing.

    It’s quite possible that an undetected influence soured this interview. Perhaps the very moment the professor heard the name “Jamal” or the instant he met the dark-skinned African-American male, he unconsciously made certain associations — “danger,” “trouble,” “less intelligent” — and these associations affected his rating of Jamal’s application.

    Despite his best intentions, the professor exhibited a form of implicit bias — also known as unconscious bias — called “implicit white race preference.” People with this bias unwittingly associate a white face with positive words or feelings and a black face with negative words or feelings — and they may act on those associations. According to a 2012 article, Professor Smith has a lot of company: 70% of several million volunteers who took the computer-based implicit association test displayed implicit white race preference, as did a significant proportion of tested physicians.

    Medical school admissions committee members like Smith are gatekeepers to the profession of medicine — and therefore wield powerful influence over the health of the nation. In the face of that power, admissions committees have an obligation to identify and remediate members’ implicit biases, whether based on race, gender, sexual orientation, or other characteristics. At one medical school, we are attempting to do just that.

    Perhaps the very moment the professor heard the name “Jamal” or the instant he met the dark-skinned African-American male, he unconsciously made certain associations.

    In 2012, each member of The Ohio State University College of Medicine admissions committee took various implicit association tests (IATs). When results showed that many committee members exhibited implicit white race preference, implicit bias against homosexuals, and unconscious association of men with “career” and women with “homemaker,” we initiated annual implicit bias mitigation workshops.

    The results were impressive: The very next cycle, we matriculated the most racially diverse class up to that point. Our proportion of underrepresented minorities increased from 17% to 20%, and it now averages around 25%. In addition, two years after implementing the training, women slightly outnumbered men in the entering class for the first time in our history, a trend that has continued for each of the last five years.

    Surprisingly, the increased racial diversity was not due to our extending more offers to minority students. Instead, a higher number of accepted minority students chose to matriculate at Ohio State. This result suggests that interviewers trained in implicit bias mitigation interact differently with minority candidates, which makes candidates feel more comfortable. That isn't surprising given research on how people interact with those against whom we have an unconscious bias: we tend to talk over them, avoid eye contact, and limit extraneous conversation, for example.

    Other studies point to the impact of biases in the selection process. For example, research shows that evaluators grade minority job candidates’ credentials and essays more harshly than those of white applicants, and that people with “ethnic-sounding” names are less likely to be hired than those with “mainstream” names despite having the same credentials. While these findings likely reflect both implicit and explicit biases, if extrapolated to medical school admissions, they suggest that many qualified applicants may be rejected before even getting to the interview stage.

    Diversity among physicians is critically important to reducing health disparities, and reducing implicit bias is key to a diverse medical workforce.

    Implicit biases develop over a lifetime and require ongoing practice to mitigate. Therefore, our admissions committee continues to review the topic through various methods, including AAMC webinars. We have taken the following steps, which we believe other institutions can easily reproduce:

    • Annual, mandatory implicit bias mitigation training sessions: All our application screeners and admissions committee members participate in 45-minute moderated discussions of implicit bias vignettes and evidenced-based strategies to reduce bias. In addition, the admissions dean leads 2 1/2-hour implicit bias workshops throughout the year for the entire medical center community, which admissions committee members often voluntarily attend.
    • Recommended readings on implicit bias: We provide research studies and other readings to our committee and include these in the admissions committee manual.
    • Interview day “cheat sheet”: Before meeting a candidate, interviewers review a bulleted list of strategies to reduce implicit bias. Strategies include “Consider the Opposite,” in which the reader decides about an applicant’s qualifications but then re-reviews the file looking for evidence supporting the opposite conclusion before making a final decision.

    Back to Jamal and Professor Smith. It is important to emphasize that the professor strives for fairness in all of his interactions. Yet in his interview with Jamal, he may have smiled less than usual, talked a great deal, and cut the meeting short — and he may have graded Jamal’s written application and verbal communication more harshly. Unconscious bias could have led to a tremendous loss for Jamal, the medical school, and ultimately, health care in this nation.

    Diversity among physicians is critically important to reducing health disparities, and reducing implicit bias is key to a diverse medical workforce. We therefore challenge key players in medicine — medical school admissions committees, residency and fellowship program selection committees, and others — to take an IAT and then spend time studying and acting on bias reduction strategies.