The Promise and Perils of Artificial Intelligence in Health Professions Education Practice and Scholarship.

ABSTRACT: Artificial intelligence (AI) methods, especially machine learning and natural language processing, are increasingly affecting health professions education (HPE), including the medical school application and selection processes, assessment, and scholarship production. The rise of large language models over the past 18 months, such as ChatGPT, has raised questions about how best to incorporate these methods into HPE. The lack of training in AI among most HPE faculty and scholars poses an important challenge in facilitating such discussions. In this commentary, the authors provide a primer on the AI methods most often used in the practice and scholarship of HPE, discuss the most pressing challenges and opportunities these tools afford, and underscore that these methods should be understood as part of the larger set of statistical tools available.Despite their ability to process huge amounts of data and their high performance completing some tasks, AI methods are only as good as the data on which they are trained. Of particular importance is that these models can perpetuate the biases that are present in those training datasets, and they can be applied in a biased manner by human users. A minimum set of expectations for the application of AI methods in HPE practice and scholarship is discussed in this commentary, including the interpretability of the models developed and the transparency needed into the use and characteristics of such methods.The rise of AI methods is affecting multiple aspects of HPE including raising questions about how best to incorporate these models into HPE practice and scholarship. In this commentary, we provide a primer on the AI methods most often used in HPE and discuss the most pressing challenges and opportunities these tools afford.

Statement of Partnership and Humility: A Structural Intervention to Improve Equity and Justice in Medical Education.

PROBLEM: Due to generational exposure to the Black Lives Matter movement, other anti-bias social movements, and diverse peer advocacy groups, health professions students are often more knowledgeable than their teachers about ways in which systemic racism and bias have led to scientific inaccuracies that contribute to health inequities. However, traditional hierarchies and concerns about retaliation may limit educational communities from benefiting maximally from students' contributions. APPROACH: In spring 2021, faculty and students at the Vagelos College of Physicians and Surgeons, Columbia University, designed a structural innovation to engage faculty and students in partnership toward decreasing bias in medical education. This article discusses development and implementation of a Statement of Partnership and Humility (SPH) disclosure slide on which faculty acknowledge consideration of potential teaching biases and invite student feedback. OUTCOMES: The initial primary goal of the SPH slide was to increase faculty awareness and engagement in anti-bias topics; however, the unexpected dividends of decreasing faculty anxiety about receiving student feedback and promoting student engagement have proven equally powerful in promoting a healthy, inclusive learning environment. NEXT STEPS: Next steps include gathering qualitative and quantitative data to elicit both faculty and student perspectives on the use of the SPH slide, particularly with regard to psychological safety and openness to feedback.

Implementing Public Health Modules as an Approach to Improve Knowledge and Attitudes of Medical Students: A Student-Led, Multiyear Study.

PURPOSE: Public health is a necessary focus of modern medical education. However, while numerous studies demonstrate benefits of public health education during medical school among self-selected students (i.e., those interested in public health), there are few educational models shown to be effective across the general medical student population. This study examined the effect of a multiyear, case-based, longitudinal online public health curriculum required for all medical students at an urban, research-focused U.S. medical school. METHOD: The authors created 11 short public health modules to supplement a year-long, organ-based preclerkship course at Columbia University Vagelos College of Physicians and Surgeons. Beginning in 2020, all students were required to complete these modules, with repeated surveys to assess changes in attitudes and knowledge of public health over time. The authors compared responses for these domains before and after each module, across multiple time points throughout the year, and cross-sectionally to a 2019 cohort of students who were not provided the modules. RESULTS: Across 3 cohorts, 405 of 420 (96.4%) students provided responses and were included in subsequent analyses. After completing the modules, students reported perceiving a greater importance of public health to nearly every medical specialty ( P < .001), more positive attitudes toward public health broadly ( P < .001), and increased knowledge of public health content ( P < .001). These findings were consistent across longitudinal analysis of students throughout the year-long course and when compared to the cohort who did not complete the modules. CONCLUSIONS: Case-based, interactive, and longitudinal public health content can be effectively integrated into the required undergraduate medical education curriculum to improve all medical students' knowledge and perceptions of public health. Incorporating evidence-based public health education into medical training may help future physicians to better address the needs of the communities and populations in which they practice.

Racial Implicit Bias and Communication Among Physicians in a Simulated Environment.

Importance: Racial implicit bias can contribute to health disparities through its negative influence on physician communication with Black patients. Interventions for physicians to address racial implicit bias in their clinical encounters are limited by a lack of high-fidelity (realistic) simulations to provide opportunities for skill development and practice. Objective: To describe the development and initial evaluation of a high-fidelity simulation of conditions under which physicians might be influenced by implicit racial bias. Design, Setting, and Participants: This cross-sectional study, performed on an online platform from March 1 to September 30, 2022, recruited a convenience sample of physician volunteers to pilot an educational simulation. Exposures: In the simulation exercise, physicians saw a 52-year-old male standardized patient (SP) (presenting as Black or White) seeking urgent care for epigastric pain, nausea, and vomiting. The case included cognitive stressors common to clinical environments, including clinical ambiguity, stress, time constraints, and interruptions. Physicians explained their diagnosis and treatment plan to the SP, wrote an assessment and management plan, completed surveys, and took the Race Implicit Association Test (IAT) and Race Medical Cooperativeness IAT. The SPs, blinded to the purpose of the study, assessed each physician's communication using skills checklists and global rating scales. Main Outcomes and Measures: Association between physicians' IAT scores and SP race with SP ratings of communication skills. Results: In 60 physicians (23 [38.3%] Asian, 4 [6.7%] Black, 23 [38.3%] White, and 10 [16.7%] other, including Latina/o/x, Middle Eastern, and multiracial; 31 [51.7%] female, 27 [45.0%] male, and 2 [3.3%] other), the interaction of physicians' Race IAT score and SP race was significant for overall communication (mean [SD] ß = -1.29 [0.41]), all subdomains of communication (mean [SD] ß = -1.17 [0.52] to -1.43 [0.59]), and overall global ratings (mean [SD] ß = -1.09 [0.39]). Black SPs rated physicians lower on communication skills for a given pro-White Race IAT score than White SPs; White SP ratings increased as physicians' pro-White bias increased. Conclusions and Relevance: In this cross-sectional study, a high-fidelity simulation calibrated with cognitive stressors common to clinical environments elicited the expected influence of racial implicit bias on physicians' communication skills. The outlined process and preliminary results can inform the development and evaluation of interventions that seek to address racial implicit bias in clinical encounters and improve physician communication with Black patients.