乌鸦传媒

Several bias reduction tools used by ophthalmology residency selection committees appeared to boost diversity in the programs, whereas others didn't help or even hurt diversity, a cross-sectional survey found.

In 65 ophthalmology residency programs, the use of multiple selection committee bias reduction tools -- such as department and institutional workshops -- were linked to higher rates of residents accepted from underrepresented racial, ethnic, gender, and socioeconomic groups (OR 1.47, 95% CI 1.13-1.92, P=0.01), reported Ambar Faridi, MD, of Oregon Health & Science University in Portland, and colleagues in . But programs with interview rubrics -- a scorecard used to evaluate interview answers -- were actually less diverse (OR 0.72, 95% CI 0.59-0.87, P=0.001).

The findings lined up with lessons from the business community about the importance of multiple strategies to improve diversity, Faridi told 乌鸦传媒. "Just taking an implicit bias training class or a workshop is not going to suddenly make you an unbiased human. You have to introduce your selection committee to multiple bias reduction tools."

The fact that 14% of the programs reported using no bias reduction tools was surprising, University of California Davis vice chair of workforce diversity Tonya Fancher, MD, MPH, said in an interview with 乌鸦传媒. "This is an absolute requirement for all residency selection committees," said Fancher, who wasn't involved with the new study. "The fact that this was the most impactful strategy to improve recruitment of a more diverse class of ophthalmology residents only reinforces that."

Ophthalmology is one of the least diverse fields in medicine. In 2019, only surveyed identified themselves as Black or African American, Hispanic or Latino, Native American, and/or Pacific Islander individuals versus that year. The LGBTQIA+ (lesbian, gay, bisexual, transgender, queer, intersex, and asexual plus) population also compared with other medical fields, the study authors wrote.

"Diversity is a real issue, because it's literally helping improve ," Faridi said.

Application Filters, Standardization May Not Help

For their study, Faridi's group sought to understand the value of various strategies that aim to boost diversity among ophthalmology residents.

The researchers sent an 18-item online questionnaire to 106 ophthalmology residency program directors in 2022, and 65 responded (61.3%). The average number of residents per class was 4.5. In the five classes prior to the survey, 72.3% of the programs had at least one Black or African American resident, 81.5% reported having a Hispanic or Latinx/a/o resident, 13.8% had a Native American or Pacific Islander, and 60.0% had a resident who identified as LGBTQIA+.

"We found that use of application filters, standardized interviews, and post-interview metrics were not associated with diversity," co-author Ogul Uner, MD, of Oregon Health & Science University told 乌鸦传媒. "However, there is evidence that these parameters impact resident diversity in other specialties. In fact, a sub-analysis of our data suggests programs that standardize over 50% of their interviews may have more resident diversity, though this finding did not achieve statistical significance. More studies with larger samples are needed to support these findings."

Application filters allow programs to focus on details about applicants such as test scores, grades, and honor status. Standardized interviews are intended to reduce bias, and post-interview metrics evaluate whether diversity programs are working.

Why might interview rubrics lead to less diversity? Faridi said research suggests that limiting "random chit-chat" can actually limit discussion of "resilience and life experiences."

For example, "when I look at an application and see that someone worked two jobs while in college and received a 3.8 GPA, that to me is more impressive than a 4.0 GPA," she said. In contrast, an interview rubric may focus more on details like class rank.

On the other hand, Fancher argued that "post-interview metrics are key data for a program to measure and improve its desired outcomes. But moving the needle takes time, a commitment to change, and resources."

Fancher also expressed surprise that "only 50% of the program directors felt that diversity of the selection committee was important. Diversifying the composition of the committee is also a key anti-bias strategy."

More Tools, Workshops Appeared Effective

Faridi and Uner highlighted how their study linked use of a greater number of bias reduction tools to more diversity. The average number of tools used was 2.5, with a range up to five. Programs that scored at least 3 on a 5-point resident diversity scale averaged 2.8 such tools, compared with 2.0 among programs with lower diversity scores (P=0.03). More programs that used bias reduction tools had high diversity scores (OR 1.28, 95% CI 1.00-1.86, P=0.05).

At least 49% of the programs utilized resources from the American Academy of Ophthalmology and Association of University Professors of Ophthalmology (AUPO)'s interactive, online, or in-person workshops and online modules. A third offered "journal club" – discussions of anti-bias research. All of these were at least numerically more common for programs with higher diversity scores.

Study limitations included the modest sample size, potential for recall bias about diversity in previous classes, and lack of confirmation of cause-and-effect relationships.

Overall, the study findings suggested "increasing the number of selection committee bias reduction and critically evaluating interview rubric strategies," Uner said.

An by Saras Ramanathan, MD, and colleagues at the University of California San Francisco noted that prioritizing applicants who express more interest in ophthalmology was negatively linked to diversity (OR 0.83, 95% CI 0.75-0.92, P=0.02).

"As AUPO considers the addition of a pilot preference signaling program to next year's residency match, this unintended consequence should not be overlooked and requires close surveillance," the editorialists added. Examples like this "highlight the complexity of efforts to reduce bias and the idea that initiatives designed to promote diversity require real-world evaluation," they added.

Correction: Interview rubrics, not standardized interviews, were linked to less diversity.

Comment