Abstract
Implicit bias in medical education has been documented, but its influence on standardized assessments remains underexplored. This study examines whether patient demographic descriptors in clinical vignettes impact examinee multiple-choice question answer choice selection and clinical decision-making. A randomized, controlled survey was conducted among third- and fourth-year medical students and first-year residents. Participants (N = 376) answered clinical vignette multiple-choice questions in either a “standard group” (without demographic descriptors) or a “demographic group” (identical questions incorporating race, gender identity, or sexual orientation). Questions were grouped by question type (bias-dependent, bias-obscuring, and diagnostic-agnostic) based on whether the question stem enforced canonical pathologic associations with race and/or sociodemographic data (bias-dependent) or presented a noncanonical racial description (bias-obscuring) that could potentially lead medical trainees to select an incorrect answer choice. Performance was analyzed using Wilcoxon rank-sum tests (α < 0.0025). Participants in the standard group correctly answered more often bias-obscuring questions than those in the demographic group (66.5% vs. 58.8%, P < 0.001). Significant performance discrepancies were noted in questions related to homosexuality and sexually transmitted infection testing, sarcoidosis in Black females, Takayasu arteritis in individuals of Asian descent, and anemia in transgender individuals. Additionally, 52.4% of participants reported that race influenced their exam decision-making, while 64.1% and 55.9% indicated that gender identity and sexual orientation played roles. The majority (54.8%) attributed pro-White bias in medical decision-making to preclinical training. The inclusion of demographic descriptors, including race, sexual orientation, and gender identity, in standardized exam questions affects medical trainees' performance, potentially reinforcing bias.
Keywords: Cognitive bias, Health equity, Implicit bias, Medical education assessment, Standardized testing
Introduction
The Liaison Committee on Medical Education mandates that “medical curriculum provides opportunities for medical students to learn to recognize and appropriately address biases in themselves, in others, and in the health care delivery process.”1 Developing culturally competent physicians improves the quality of care of individuals across different minority groups and may help reduce health disparities related to various racial and ethnic groups. Racial and ethnic health disparities have been demonstrated in numerous diseases, notably in cardiovascular disease. Non-Hispanic Black individuals are more than twice as likely to die from heart disease than non-Hispanic Asian and Pacific Islanders. Hispanic and Black individuals are also disproportionately affected by significant risk factors for the development of cardiovascular disease, including hypertension, obesity, and diabetes.2
There is no better time to address health disparities through implicit bias in medical education to develop culturally competent physicians than from the start. One example of success in this area is the recognition of the characteristic rash of early-stage Lyme disease, erythema migrans. Erythema migrans is frequently presented to medical trainees on white skin, making it difficult for Black patients to receive a timely and accurate diagnosis of Lyme disease.3 A clinical handbook compiled by Mukwende et al. and referenced by Rimmer et al. was released to help medical trainees and clinicians gain exposure to dermatologic findings on darker skin tones. This handbook was made freely downloadable online and used globally, which has helped medical trainees and practicing physicians gain more awareness of how differently skin pathologies can appear on various skin tones.4
Medical education related to health disparities needs to be thoughtfully and longitudinally integrated into medical curricula and assessed throughout the preclinical and clinical years.5 There are a multitude of ways to assess medical trainees’ knowledge of health disparities, including observed structured clinical exams, written assessments, reflective practices, and questionnaires.6 The most common type of assessment during the preclinical years and the most numerous throughout the lifetime of a medical trainee and physician is the written assessment. Inclusion of Black, Asian, and minority ethnic patients in clinical vignettes is believed by medical students to be an integral aspect of achieving health equity.7 Yet, clinical vignettes typically take the form of multiple-choice-questions (MCQs), posing the risk of reinforcing categorical or stereotypical information.
Medical licensing exams assess clinical competency yet their role in reinforcing bias remains under examined. The United States Medical Licensing Examination (USMLE) and similar assessments are intended to be objective but the inclusion of demographic descriptors in question vignettes may introduce unintended cognitive biases that influence test-takers’ clinical reasoning. Previous research has documented implicit bias in medical training, particularly regarding race, gender identity, and sexual orientation.8 However, whether these biases affect standardized test performance has been largely unexplored. Given the implications of licensing exams for medical careers and patient care, identifying and mitigating bias in testing is critical.
This randomized controlled trial provides insight into the relationship between demographic identifiers and specific pathologies as well as the role that medical education plays in the development of biases among medical trainees. We investigate whether demographic descriptors in exam questions influence answer selection and discuss the broader implications for medical education reform.
Materials and methods
A survey consisting of clinical vignette MCQs was developed via Qualtrics (Qualtrics, Provo, UT). The study design, including methods of recruitment, was evaluated by the Institutional Review Board at Loyola University Chicago and found to be exempt.
The survey was distributed via email and social media to voluntary participants. Recruitment materials purposefully omitted the study’s subject matter of bias (email text available in Supplemental Material 1). Links to the survey were shared on X via the following accounts: @KMirza, @cullen_lilley, @PathElective, and @Inside_TheMatch. Participants were eligible to engage in the survey if they were at least 18 years old and were currently in their M3 through PGY1 years of medical training. The survey internally randomized participants to either a standard or demographic set of questions, with questions being identical except for the addition of a key patient demographic descriptor–race, sexual orientation, or gender identity–in the demographic questions (survey available in Supplemental Material 1). Of note, familial descent was retained in select questions from the standard group. Additionally, surveys included questions regarding participant demographics and education level, as well as perceptions of how bias in medical training may develop and influence clinical decision-making.
When opening the survey link, participants were first presented with an informed consent statement that they had to agree to in order to move onto the first section. The first section elicited several pieces of demographic data, including age, ethnicity, gender identity, location of medical school, type of medical degree, current level of training, completion of USMLE and Comprehensive Osteopathic Medical Licensing Examination (COMLEX) examinations, and Medical College Admission Test (MCAT) score. Participants were presented with one clinical vignette MCQ at a time upon advancing to the second section. They were able to toggle back and forth between the questions in section 2. After question 12 in section 2, participants advanced to the third and final section of the survey which presented questions about how trainees experienced bias in medical school. Once advancing to section 3, participants were unable to return to section 2.
With a proposed total sample size of 786 respondents randomly assigned to either standard or demographic question blocks with a 1:1 allocation, the study would have a power of 80.0% to yield a statistically significant result. This computation assumed a two-sided alpha level of 0.05, indicating that the mean difference between the two groups was at least 0.10 with a common SD of 0.5. This difference was selected based on pilot data. At the end of 14 months, the study was halted due to slow recruitment. A significant decline in survey responses in the months leading up to survey termination was observed despite continued recruitment efforts. During this time, a total of 432 survey responses were recorded. Surveys were included for analysis if the participant answered all the questions in their randomly assigned question block (n = 376).
To compare clinical vignette MCQs, the average percentage answered correctly was compared across question blocks using the Wilcoxon rank-sum test. Questions were compared individually and by question type (bias-dependent, bias-obscuring, and diagnostic-agnostic). The question type was determined based on how race was presented in boards study materials. If the question stem included the canonical racial descriptor, these would be considered “bias-dependent,” but if the questions stem included a noncanonical racial descriptor, these would be considered “bias-obscuring.” In bias-dependent questions (questions 4, 5, and 7), race and/or sociodemographic data in the clinical vignette with a classical racial description are introduced in a way that may influence medical trainees to select the correct answer. In bias-obscuring questions (questions 1, 2, 3, 9, and 10), race and/or sociodemographic data with a nonclassical racial description are included, potentially leading medical trainees to select an incorrect answer choice. In diagnosis-agnostic bias questions (question 12), the diagnosis itself is unaffected, but race and/or sociodemographic data are introduced to assess whether medical trainees hold outdated misconceptions regarding associations between demographic descriptors and the correct answer. Only participants in the demographic group were exposed to bias-dependent, bias-obscuring, and diagnosis-agnostic bias. Questions 6, 8, and 11 were not grouped in the aforementioned question types as they reflect an assessment of the medical trainee’s knowledge of more recent developments in medicine pertaining to race and sociodemographics including estimated glomerular filtration rate (EGFR) calculations, hemoglobin reference ranges in transgender patients, keloid pathophysiology, and first-line blood pressure medications.
Lastly, perceptions of patient demographic influence on medical training and clinical decision-making are summarized as counts and percentages among each question block. Due to multiple hypotheses testing, the corrected threshold for significance is set to α < 0.0025 (0.0025 = 0.05/20). All tests are two-sided and analyses were conducted via Stata v.18 (StataCorp, 2023).
Results
The respondents had an average age of 28 years (SD: 4.2; Table 1). Of the 376 respondents, 47% (n = 201) were female, 47.1% (n = 177) were White, and 8.8% (n = 33) were Black or African American (Table 1). Of the 376 respondents, 32.7% (n = 123) attended medical school in the Midwest (definition available within Supplemental Material 1), while 27.9% (n = 82) attended medical school outside the United States. 72.3% (n = 272) were pursuing an MD degree, while 15.4% (n = 58) were pursuing a DO degree. Of all, 29.3% (n = 110) were M3, 34.6% (n = 130) were M4, and 29.0% (n = 109) were PGY1. Of the 376 respondents, 87.3% (n = 328) had taken at least one USMLE or COMLEX test before. All participant demographics and education levels were similar across question blocks.
Table 1.
Participant demographics and medical education history.
| Total (N = 376) |
Standard (N = 186) |
Demographic (N = 190) |
||
|---|---|---|---|---|
| Mean (SD) | Mean (SD) | Mean (SD) | ||
| Age (years) | 28.2 (4.2) | 28.3 (4.6) | 28.1 (3.8) | |
| N (%) | N (%) | N (%) | ||
| Race | White | 177 (47.1%) | 86 (46.2%) | 91 (47.9%) |
| Black/AA | 33 (8.8%) | 18 (9.7%) | 15 (7.9%) | |
| Other | 126 (33.5%) | 63 (33.9%) | 63 (33.2%) | |
| MultiRacial | 30 (8.0%) | 14 (7.5%) | 16 (8.4%) | |
| Unknown/not answered | 10 (2.7%) | 5 (2.7%) | 5 (2.6%) | |
| Gender identity | Male | 163 (43.4%) | 85 (45.7%) | 78 (41.1%) |
| Female | 201 (53.5%) | 93 (50.0%) | 108 (56.8%) | |
| Other | 6 (1.6%) | 4 (2.2%) | 2 (1.1%) | |
| Unknown | 6 (1.6%) | 4 (2.2%) | 2 (1.1%) | |
| Medical school location | Midwest | 123 (32.7%) | 62 (33.3%) | 61 (32.1%) |
| Northwest | 28 (7.4%) | 13 (7.0%) | 15 (7.9%) | |
| South | 82 (21.8%) | 40 (21.5%) | 42 (22.1%) | |
| West | 37 (9.8%) | 23 (12.4%) | 14 (7.4%) | |
| US territory | 1 (0.3%) | 0 (0.0%) | 1 (0.5%) | |
| Outside US | 105 (27.9%) | 48 (25.8%) | 57 (30.0%) | |
| Degree type | MD | 272 (72.3%) | 139 (74.7%) | 133 (70.0%) |
| DO | 58 (15.4%) | 25 (13.4%) | 33 (17.4%) | |
| MBBS/MBChB/Other | 46 (12.2%) | 22 (11.8%) | 24 (12.6%) | |
| Current training | MS3/Year 3 | 110 (29.3%) | 57 (30.6%) | 53 (27.9%) |
| MS4/Year 4 | 130 (34.6%) | 64 (34.4%) | 66 (34.7%) | |
| Year 5 (outside US) | 14 (3.7%) | 6 (3.2%) | 8 (4.2%) | |
| Year 6 (outside US) | 13 (3.5%) | 7 (3.8%) | 6 (3.2%) | |
| PGY1 | 109 (29.0%) | 52 (28.0%) | 57 (30.0%) | |
| Have you taken any United States USMLE or COMLEX tests? | Yes, USMLE only | 271 (72.1%) | 138 (74.2%) | 133 (70.0%) |
| Yes, COMLEX only | 22 (5.9%) | 10 (5.4%) | 12 (6.3%) | |
| Yes, both USMLE & COMLEX | 35 (9.3%) | 15 (8.1%) | 20 (10.5%) | |
| No, neither | 48 (12.8%) | 23 (12.4%) | 25 (13.2%) | |
AA, African American; USMLE, United States Medical Licensing Examination; COMLEX, Comprehensive Osteopathic Medical Licensing Examination.
When comparing the average correct response among individual questions (Fig. 1) among question blocks, the standard group answered correctly more often than the demographic group on questions 1, 2, and 8, while the demographic group answered correctly more often than the standard group on question 7 (all P < 0.001).
Fig. 1.
Assessment Performance by Question. Percentage of correct responses for each survey question (1–12). Statistically significant differences were observed in questions 1, 2, 7, and 8 (∗P < 0.001). The standard group outperformed the demographic group on questions 1, 2, and 8, while the demographic group answered question 7 correctly more often than the standard group.
In question 1, the demographic identifier was “homosexual,” the correct answer was “blood pressure,” and the distractor was “chlamydia/gonorrhea testing.” Of the 186 respondents in the standard group, 82.8% (n = 154) selected blood pressure, and 10.2% (n = 19) selected chlamydia/gonorrhea testing. Of the 190 respondents in the demographic group, 62.6% (n = 119) selected blood pressure and 28.4% (n = 54) selected chlamydia/gonorrhea testing.
In question 2, the demographic identifier was “Black,” the correct answer was “upper airway cough syndrome (postnasal drip),” and the distractor was “sarcoidosis.” Of the 186 respondents in the standard group, 62.4% (n = 116) selected upper airway cough syndrome and 10.2% (n = 19) selected sarcoidosis. Of the 190 respondents in the demographic group, 40.5% (n = 77) selected upper airway cough syndrome and 45.3% (n = 86) selected sarcoidosis.
In question 7, the demographic identifier was “Asian,” and the correct answer was “Takayasu arteritis.” Of the 186 respondents in the standard group, 58.6% (n = 109) selected Takayasu arteritis. Of the 190 respondents in the demographic group, 75.3% (n = 143) selected Takayasu arteritis.
In question 8, the demographic identifier was “trans,” the correct answer indicates that the hemoglobin level was normal and the incorrect answers indicate otherwise. Of the 186 respondents in the standard group, 75.8% (n = 141) correctly identified the hemoglobin level as normal. Of the 190 respondents in the demographic group, only 25.3% (n = 48) correctly identified the hemoglobin level as normal.
When examining groupings of questions based on bias type (Fig. 2), the bias-obscuring questions (questions 1, 2, 3, 9, and 10) were answered correctly significantly more often by the standard group (66.5%) than the demographic group (58.8%) (P-value: <0.001). There was no significant difference between standard and demographic groups observed for bias-dependent and diagnosis-agostic question types.
Fig. 2.
Assessment Performance by Category. Percentage of correct responses by bias type. The standard group answered bias-obscuring questions (questions 1, 2, 3, 9, and 10) correctly significantly more often (66.5%) than the demographic group (58.8%; ∗P < 0.001).
Among all respondents, 89.4% (n = 336) reported that race was used in their medical training to highlight common demographics of specific pathologies (Table 2). Of the 376 respondents, 64.1% (n = 241) and 55.9% (n = 210) reported that gender identity and sexual orientation, respectively, were used in similar manners. Of the 376 respondents, 52.4% (n = 197) indicated that race strongly influences their clinical decision-making in the exam setting, while 40.7% (n = 153) and 21.3% (n = 80) indicated that gender identity and sexual orientation, respectively, influence their clinical decision-making. Of the 376 respondents, 54.8% (n = 206) believed that preclinical years and training contribute the most to pro-White bias among physicians, while 26.9% (n = 101) and 18.8% (n = 35) believe that clinical years and patient interactions, respectively, contribute the most.
Table 2.
Perceptions of patient demographics and clinical decision-making.
| Total |
Standard |
Demographic |
||
|---|---|---|---|---|
| N (%) | N (%) | N (%) | ||
| Which of the following were used in your medical education training to highlight common demographics of certain pathologies? Select all that apply | Race | 336 (89.4) | 162 (87.1) | 174 (91.6) |
| Gender identity | 241 (64.1) | 117 (62.9) | 124 (65.3) | |
| Sexual orientation | 210 (55.9) | 99 (53.2) | 111 (58.4) | |
| None of the above | 26 (6.9) | 18 (9.7) | 8 (4.2) | |
| Which of the following do you feel most strongly influences your clinical decision-making in the exam setting? Select all that apply | Race | 197 (52.4) | 72 (42.5) | 118 (62.1) |
| Gender identity | 153 (40.7) | 83 (44.6) | 70 (36.8) | |
| Sexual orientation | 80 (21.3) | 37 (19.9) | 43 (22.6) | |
| None of the above | 106 (28.2) | 58 (31.2) | 48 (25.3) | |
| Pro-White bias has been demonstrated time and time again among physicians using tools like the implicit association test (IAT). This bias is evidenced even in the absence of explicitly stated preference for whites. What do you think most contributes to this pro-White bias among physicians? |
Preclinical years/Training | 206 (54.8) | 105 (56.5) | 101 (53.2) |
| Clinical years/training | 101 (26.9) | 43 (23.1) | 58 (30.5) | |
| Patient interactions | 35 (18.8) | 35 (18.8) | 30 (15.8) | |
| Unknown/not answered | 4 (1.1) | 3 (1.6) | 1 (0.5) | |
Discussion
In this study, we demonstrate statistically significant differences in participants’ answer choices on mock exam questions when only demographic descriptors are added. Participants in the standard group performed better on bias-obscuring questions compared to the demographic group (66.5% vs. 58.8%), supporting how demographic descriptors affect medical trainees’ answer choices on exams. The discrepancies were most pronounced in conditions historically thought to be associated with specific racial or ethnic groups, such as sarcoidosis in Black females and Takayasu arteritis in Asian females. Despite the lack of desired power, these data suggest that including patient demographic descriptors in exam questions alters response patterns, potentially reflecting underlying biases during medical training. These biases were endorsed by more than half of respondents (52.4%), acknowledging race’s influence on clinical decision-making on exams. Our results align with prior research demonstrating that implicit biases shape medical judgment; however, it is the first of its kind demonstrating its impact on medical assessment.8 On licensing exams, these biases may translate into disparities in medical knowledge assessment, influencing career trajectories and patient outcomes. Given the high-stakes nature of these exams and how preparation for them could potentially influence future medical practice, addressing potential bias in question design is critical.
Bias is a well-documented, pervasive issue in the medical field. Prior studies have shown that implicit and explicit biases affect diagnosis,9,10 treatment,11, 12, 13 and even inclusion in clinical trials.14 These studies have demonstrated that implicit bias has a measurable impact on patient care. However, when it comes to medical training, there are explicit biases when it comes to the diagnosis of specific disease entities, and there has been little work examining what impact training has on medical trainees’ selection of diagnoses on exam questions. Race-based disease descriptors are usually an attempt to account for higher disease prevalence based on ancestry, but biologically, race does not often reflect the nuances needed to understand the origins of diseases with a genetic influence.15 The recognition of this fallacious logic by members of the medical education community has highlighted the need for revising sentinel textbooks in medical education.15,16 This study suggests that there is a measurable impact of these historical biases in medical decision-making on training assessments.
As trainees prepare for medical licensing exams, the vast amount of material needed to learn necessitates the condensation of complex pathobiology into a simplistic “model patient” defined by race or other demographic descriptors, which aids them in arriving at the ‘correct’ answer. However, such simplification could lead to biases in patient care as these trainees progress through their careers. These trained biases could potentially influence how trainees interact with patients, affecting the quality of care and patient outcomes. Thus, it is crucial to critically evaluate and revise medical curricula and exam content to prevent the reinforcement of these biases. Ensuring that the assessment of clinical reasoning does not rely on oversimplified demographic associations is essential for the fair evaluation of trainees and the delivery of equitable patient care.
Our study suggests that participants perceive preclinical training is a significant source of pro-White bias in medical education, with 54.8% of respondents acknowledging this influence (Table 2). Additionally, 64.1% and 55.9% reported that gender identity and sexual orientation, respectively, also played a role in shaping biases. Introducing explicit biases so early in training reinforces outdated or inaccurate associations in disease diagnosis and treatment while simultaneously shaping the implicit biases that guide clinical reasoning. Integrating such training into medical curricula from the preclinical years onward is crucial for mitigating bias and fostering a more nuanced, evidence-based understanding of health disparities.
To address the impact of bias in medical training and assessment, we recommend several key policy considerations (Table 3). Licensing exam content should minimize the use of demographic descriptors unless absolutely necessary, ensuring that test takers do not rely on demographic heuristics when formulating diagnoses. However, one may argue that the use of demographic descriptors in limited settings may continue to propagate biases related to those disease processes. Alternatively, it may be advantageous to include demographic information including race and gender in all questions as to not highlight the association present in limited settings. This information could be listed separately from the clinical vignette to mirror an actual patient encounter in which trainees will never be blinded to certain information including race and gender. By listing the demographic information separately, it may read more as information in the background as opposed to a factor that directly impacts the clinical presentation and course. To further reduce deeply engrained historical biases, we suggest regularly exposing trainees to bias-obscuring questions throughout medical school and postgraduate training to both lessen the association between a classical racial description and disease process and broaden the differential diagnosis for a non-classical presentation. We may even come to a time where demographic information varies from medical trainee to medical trainee within the clinical vignette with the correct answer remaining unchanged. This may more accurately prepare trainees for clinical rotations during which they will be exposed to the same disease processes in different individuals.
Table 3.
Proposed policy recommendations.
| Policy Change | Description |
|---|---|
| Revise licensing exam content | Minimize the use of demographic descriptors unless clinically relevant to prevent reliance on heuristics. |
| Integrate structured bias awareness programs | Implement workshops and case-based discussions in both preclinical and clinical training to mitigate implicit biases. |
| Evolve assessment methods beyond multiple-choice question exams | Incorporate competency-based evaluations, such as standardized patient encounters and clinical simulations free from demographic bias. |
| Expand research on bias in testing | Conduct longitudinal studies on how bias in medical education affects career progression, diagnostic accuracy, and patient outcomes. |
Additionally, medical curricula should integrate structured bias awareness programs, including workshops and case-based discussions, throughout preclinical and clinical training to help trainees recognize and mitigate implicit biases. Although we did not elicit data from study participants related to their exposure to bias training, we believe such programs, ideally standardized across medical schools, would benefit the future clinician by educating them about historical biases and how they can impact clinical decision-making and patient outcomes. Awareness is the first step to reducing these biases on an individual level. Moreover, we have the potential to mitigate the development of biases through teaching methods. We suggest that demographic information related to disease processes be based on current epidemiology and include quantitative values such as prevalence for any association that is to be highlighted. It would also be beneficial to incorporate related health disparities when applicable, so trainees may be aware of such disparities as early in their medical training as possible.
Assessment methods must also evolve beyond MCQ exams to incorporate competency-based evaluations, such as standardized patient encounters and clinical simulations designed to be free from demographic bias, allowing for a more equitable evaluation of clinical reasoning skills. Furthermore, expanding research on bias in testing is critical. By implementing these strategies, medical education can foster a more equitable training environment that prioritizes accuracy, fairness, and culturally competent patient care.
While this study had a sizable sample, due to the long recruitment, we were unable to achieve our desired power level. However significant findings were discovered, other bias-types may have been affected, and we are unable to detect their significance. The demographics of the participants may not be fully representative of the broader medical trainee population as the study had limited diversity in both race and geographic location. This lack of diversity could also reflect the current underrepresentation of certain racial and ethnic groups within the medical trainee pool.17, 18, 19, 20 Additionally, the generalizability of these findings to other medical settings or countries may be limited due to cultural and healthcare system differences. Future research should focus on exploring the long-term effects of bias on medical assessments on clinical practice and patient outcomes. Moreover, it is essential to investigate how implicit and explicit bias training interventions can influence clinical practice and enhance the development of diagnostic acumen in real-world medical settings to better understand how these programs can improve the quality of medical education and patient care.21
Despite these limitations, this study still has a significant impact by shedding light on how bias in medical assessments may influence clinical decision-making. It highlights the need for urgent policy changes to reduce bias in exam content and underscores the importance of incorporating bias awareness into medical training. This research reinforces the necessity of evidence-based approaches to mitigating bias, ensuring that efforts to improve healthcare equity remain rooted in data and patient outcomes. By revealing the influence of demographic descriptors on medical decision-making, the study provides a foundation for future research and interventions aimed at promoting a more equitable and culturally competent healthcare system. As the medical community works toward addressing these issues, this study serves as a crucial step in advocating for reforms that will foster a more inclusive and fair approach to medical education and patient care.
Conclusion
Bias in medical education extends to licensing exams, where demographic descriptors may influence performance. Our findings highlight the need for systemic reforms to reduce bias in test questions and enhance fairness in medical assessments. Implementing bias-aware test design and training initiatives will help ensure an equitable evaluation process, ultimately improving patient care.
Disclosures
Kamran Mirza reports financial support was provided by Loyola University Chicago. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Funding
The authors gratefully acknowledge support from the Leischner Department of Medical Education’s York Educational Research Grant Program, which contributed to the development of this work.
Declaration of competing interest
The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Kamran Mirza reports financial support was provided by Loyola University Chicago. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Footnotes
Supplementary data to this article can be found online at https://doi.org/10.1016/j.acpath.2026.100280.
Supplementary data
The following is the Supplementary data to this article:
References
- 1.Functions and Structure of a Medical School. Liaison Committee on Medical Education; 2025. https://lcme.org/wp-content/uploads/2025/10/2025-26-Functions-and-Structure_2025-05-21.docx [Google Scholar]
- 2.Wall H.K., Ritchey M.D., Gillespie C., Omura J.D., Jamal A., George M.G. Vital signs: prevalence of key cardiovascular disease risk factors for million hearts 2022 - United States, 2011–2016. MMWR Morb Mortal Wkly Rep. 2018;67(35):983–991. doi: 10.15585/mmwr.mm6735a4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Khan S., Mian A. Racism and medical education. Lancet Infect Dis. 2020;20(9):1009. doi: 10.1016/S1473-3099(20)30639-3. [DOI] [PubMed] [Google Scholar]
- 4.Rimmer A. Presenting clinical features on darker skin: five minutes with… Malone Mukwende. BMJ. 2020;369:m2578. doi: 10.1136/bmj.m2578. [DOI] [PubMed] [Google Scholar]
- 5.Zanting A., Meershoek A., Frambach J.M., Krumeich A. The “exotic other” in medical curricula: rethinking cultural diversity in course manuals. Med Teach. 2020;42(7):791–798. doi: 10.1080/0142159X.2020.1736534. [DOI] [PubMed] [Google Scholar]
- 6.Dogra N., Bhatti F., Ertubey C., et al. Teaching diversity to medical undergraduates: curriculum development, delivery and assessment. AMEE guide no. 103. Med Teach. 2016;38(4):323–337. doi: 10.3109/0142159X.2015.1105944. [DOI] [PubMed] [Google Scholar]
- 7.Hariharan B., Quarshie L.S., Amdahl C., Winterburn S., Offiah G. Experiencing racism within medical school curriculum: 2020 ICCH student symposium. Patient Educ Couns. 2022;105(7):2599–2602. doi: 10.1016/j.pec.2021.12.018. [DOI] [PubMed] [Google Scholar]
- 8.Mouhab A., Radjack R., Moro M.R., Lambert M. Racial biases in clinical practice and medical education: a scoping review. BMC Med Educ. 2024;24(1):1196. doi: 10.1186/s12909-024-06119-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Dehon E., Weiss N., Jones J., Faulconer W., Hinton E., Sterling S. A systematic review of the impact of physician implicit racial bias on clinical decision making. Acad Emerg Med. 2017;24(8):895–904. doi: 10.1111/acem.13214. [DOI] [PubMed] [Google Scholar]
- 10.Sabin J., Nagasawa P., Guenther G., et al. Implicit and explicit race and weight biases among physician assistant preceptors and trainees. J Phys Assist Educ. 2025;36(2):167–175. doi: 10.1097/JPA.0000000000000657. [DOI] [PubMed] [Google Scholar]
- 11.Dickason R.M., Chauhan V., Mor A., et al. Racial differences in opiate administration for pain relief at an academic emergency department. West J Emerg Med. 2015;16(3):372–380. doi: 10.5811/westjem.2015.3.23893. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Sabin J.A., Greenwald A.G. The influence of implicit bias on treatment recommendations for 4 common pediatric conditions: pain, urinary tract infection, attention deficit hyperactivity disorder, and asthma. Am J Public Health. 2012;102(5):988–995. doi: 10.2105/AJPH.2011.300621. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Hoffman K.M., Trawalter S., Axt J.R., Oliver M.N. Racial bias in pain assessment and treatment recommendations, and false beliefs about biological differences between blacks and whites. Proc Natl Acad Sci USA. 2016;113(16):4296–4301. doi: 10.1073/pnas.1516047113. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Graetz D.E., Madni A., Gossett J., et al. Role of implicit bias in pediatric cancer clinical trials and enrollment recommendations among pediatric oncology providers. Cancer. 2021;127(2):284–290. doi: 10.1002/cncr.33268. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Deyrup A., Graves J.L. Racial biology and medical misconceptions. N Engl J Med. 2022;386(6):501–503. doi: 10.1056/NEJMp2116224. [DOI] [PubMed] [Google Scholar]
- 16.Li A., Deyrup A.T., Graves J.L., Ross L.F. Race in the reading: a study of problematic uses of race and ethnicity in a prominent pediatrics textbook. Acad Med. 2022;97(10):1521–1527. doi: 10.1097/ACM.0000000000004666. [DOI] [PubMed] [Google Scholar]
- 17.2024 FACTS: Applicants and Matriculants Data . 2024. Association of American Medical Colleges.https://www.aamc.org/data-reports/data/2024-facts [Google Scholar]
- 18.Perez M.A., Williams C., Henderson K., et al. Association of applicant demographic factors with medical school acceptance. BMC Med Educ. 2023;23(1):960. doi: 10.1186/s12909-023-04897-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Francone N.O., Simon M.A., Ortega P. Exploring race and ethnicity representational inequities in Illinois medical schools. Health Equity. 2021;5(1):526–533. doi: 10.1089/heq.2021.0026. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Lucey C.R., Saguil A. The consequences of structural racism on MCAT scores and medical school admissions: the past is prologue. Acad Med. 2020;95(3):351–356. doi: 10.1097/ACM.0000000000002939. [DOI] [PubMed] [Google Scholar]
- 21.Lau B.D., Haider A.H., Streiff M.B., et al. Eliminating health care disparities with mandatory clinical decision support: the venous thromboembolism (VTE) example. Med Care. 2015;53(1):18–24. doi: 10.1097/MLR.0000000000000251. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.