April 11, 2018

Medical students share results of performance evaluation gender bias study nationally

Is there gender bias in medical student performance evaluations at the Wayne State University School of Medicine? No. And the nine first- and second-year medical students who came to that "surprising" conclusion earned a spot at a national meeting last month to elaborate.

The students, members of the School of Medicine chapter of the American Medical Women's Association, presented a poster of their project - "Analysis of Gender-Based Linguistic Differences in Clinical Performance Evaluations of Third-Year Medical Students" - at the 103rd AMWA Anniversary Meeting, held March 22-25 in Philadelphia.

The idea for the research project was born at one of the organization's monthly journal clubs, where a shared article focused on an Iowa medical school that examined its Medical Student Performance Evaluations for patterns of gender bias and the language of bias in narrative evaluations. The Iowa school found that females were more likely to be described as compassionate, sensitive and enthusiastic, while males were lauded for being "quick-learners."

"The paper we read found that implicit gender bias can influence the word choices used in evaluations and those evaluations can influence applicants' chances of getting the job they are applying for, whether they're a medical student applying to residency or a professor applying for tenure," said Emma Hickman, a second-year student and poster co-author. "Often, jobs or tenure committees are looking for certain qualities in their applicants and if those qualities are, for example, 'independent' or 'leader,' then it becomes a problem if those words are more often being associated with the male gender."

They worked on the project from October 2017 through March 2018, retrospectively analyzing 1,163 Internal Medicine Clinical Performance Evaluations of WSU third-year medical students, including 576 males and 587 female, from five hospital systems. A CPE is a narrative assessment of a medical student's performance in a clerkship as determined by the supervising physician or resident, and can greatly affect their application to residency programs since it determines clerkship grades and clinical honors. It is also a principal component of the MSPE, also known as the Dean's Letter.

Adjectives used in the evaluations were categorized to include up to 23 different descriptions in five categories - grindstone, standout, intelligence, team player, friendly and compassion. They found minimal linguistic differences between male and female students and concluded that there was no obvious gender bias in the evaluations, a surprising improvement compared to previously published studies. They suggested, however, an investigation of WSU School of Medicine evaluation templates provided to evaluators, as well as a review of possible instructions or seminars provided to evaluators specifically aimed to minimize gender bias.

"The feedback at the meeting was all very positive. Everyone was glad to see that we did not find any bias at Wayne State, and they all thought it was very relevant to them as medical students, especially those about to enter their third-year clerkships," Hickman said.

"With regard to our results, we were pleasantly surprised to see that there were no differences, although in the results section we do talk about a few caveats to our research, mainly that we were looking at the evaluations of students on their Internal Medicine rotation, and Internal Medicine, as opposed to Surgery, tends to be a more equal field in terms of gender distribution. We also did not account for the gender of the evaluators themselves, which would have been interesting to incorporate," Hickman added.

The journal club is a monthly event hosted by the School of Medicine's AMWA chapter. Members of the Women in Medicine and Science chapter, many practicing physicians, will often join the discussions. Professor of Internal Medicine and Vice Chair of Education Diane Levine, M.D., also served as a poster co-author.