The Role of Large Language Models in Transforming Emergency Medicine: Scoping Review.

BACKGROUND: Artificial intelligence (AI), more specifically large language models (LLMs), holds significant potential in revolutionizing emergency care delivery by optimizing clinical workflows and enhancing the quality of decision-making. Although enthusiasm for integrating LLMs into emergency medicine (EM) is growing, the existing literature is characterized by a disparate collection of individual studies, conceptual analyses, and preliminary implementations. Given these complexities and gaps in understanding, a cohesive framework is needed to comprehend the existing body of knowledge on the application of LLMs in EM. OBJECTIVE: Given the absence of a comprehensive framework for exploring the roles of LLMs in EM, this scoping review aims to systematically map the existing literature on LLMs' potential applications within EM and identify directions for future research. Addressing this gap will allow for informed advancements in the field. METHODS: Using PRISMA-ScR (Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews) criteria, we searched Ovid MEDLINE, Embase, Web of Science, and Google Scholar for papers published between January 2018 and August 2023 that discussed LLMs' use in EM. We excluded other forms of AI. A total of 1994 unique titles and abstracts were screened, and each full-text paper was independently reviewed by 2 authors. Data were abstracted independently, and 5 authors performed a collaborative quantitative and qualitative synthesis of the data. RESULTS: A total of 43 papers were included. Studies were predominantly from 2022 to 2023 and conducted in the United States and China. We uncovered four major themes: (1) clinical decision-making and support was highlighted as a pivotal area, with LLMs playing a substantial role in enhancing patient care, notably through their application in real-time triage, allowing early recognition of patient urgency; (2) efficiency, workflow, and information management demonstrated the capacity of LLMs to significantly boost operational efficiency, particularly through the automation of patient record synthesis, which could reduce administrative burden and enhance patient-centric care; (3) risks, ethics, and transparency were identified as areas of concern, especially regarding the reliability of LLMs' outputs, and specific studies highlighted the challenges of ensuring unbiased decision-making amidst potentially flawed training data sets, stressing the importance of thorough validation and ethical oversight; and (4) education and communication possibilities included LLMs' capacity to enrich medical training, such as through using simulated patient interactions that enhance communication skills. CONCLUSIONS: LLMs have the potential to fundamentally transform EM, enhancing clinical decision-making, optimizing workflows, and improving patient outcomes. This review sets the stage for future advancements by identifying key research areas: prospective validation of LLM applications, establishing standards for responsible use, understanding provider and patient perceptions, and improving physicians' AI literacy. Effective integration of LLMs into EM will require collaborative efforts and thorough evaluation to ensure these technologies can be safely and effectively applied.

Initiating medical abortion in an emergency department in the United States.

OBJECTIVES: The primary objective of this study was to assess the feasibility of initiating medical abortions in a large, academic emergency department (ED) in the United States. METHODS: A retrospective case series analysis was conducted to evaluate a protocol for initiating medical abortion in the ED implemented from January 2020 to October 2023 at an academic, tertiary care hospital in California, USA. Participants included ED patients diagnosed with pregnancies in the first trimester that were undesired and who opted for medical abortion. The medical abortion protocol was collaboratively designed by a multidisciplinary team and follow-up was conducted by our institution's gynaecology department. Data were sourced from a data repository of electronic health records and subjected to descriptive statistical analysis. RESULTS: A total of 27 eligible patients initiated medical abortions in the ED during the study period. The cohort was diverse in terms of racial and ethnic backgrounds and almost evenly split between private and public insurance. No patients had significant complications identified in the medical record. Two patients required uterine aspiration by the gynaecology team; one patient in clinic and one during a return visit to the ED. CONCLUSIONS: Data from this case series suggest that initiating medical abortion in the ED is feasible. The ED may be considered as an additional access point for abortion care services, especially in areas where other care options are not readily available. Educational, legal and regulatory frameworks that allow emergency physicians to take a greater role in providing this care should be considered.

Brain versus bot: Distinguishing letters of recommendation authored by humans compared with artificial intelligence.

Objectives: Letters of recommendation (LORs) are essential within academic medicine, affecting a number of important decisions regarding advancement, yet these letters take significant amounts of time and labor to prepare. The use of generative artificial intelligence (AI) tools, such as ChatGPT, are gaining popularity for a variety of academic writing tasks and offer an innovative solution to relieve the burden of letter writing. It is yet to be determined if ChatGPT could aid in crafting LORs, particularly in high-stakes contexts like faculty promotion. To determine the feasibility of this process and whether there is a significant difference between AI and human-authored letters, we conducted a study aimed at determining whether academic physicians can distinguish between the two. Methods: A quasi-experimental study was conducted using a single-blind design. Academic physicians with experience in reviewing LORs were presented with LORs for promotion to associate professor, written by either humans or AI. Participants reviewed LORs and identified the authorship. Statistical analysis was performed to determine accuracy in distinguishing between human and AI-authored LORs. Additionally, the perceived quality and persuasiveness of the LORs were compared based on suspected and actual authorship. Results: A total of 32 participants completed letter review. The mean accuracy of distinguishing between human- versus AI-authored LORs was 59.4%. The reviewer's certainty and time spent deliberating did not significantly impact accuracy. LORs suspected to be human-authored were rated more favorably in terms of quality and persuasiveness. A difference in gender-biased language was observed in our letters: human-authored letters contained significantly more female-associated words, while the majority of AI-authored letters tended to use more male-associated words. Conclusions: Participants were unable to reliably differentiate between human- and AI-authored LORs for promotion. AI may be able to generate LORs and relieve the burden of letter writing for academicians. New strategies, policies, and guidelines are needed to balance the benefits of AI while preserving integrity and fairness in academic promotion decisions.

A Conference (Missingness in Action) to Address Missingness in Data and AI in Health Care: Qualitative Thematic Analysis.

BACKGROUND: Missingness in health care data poses significant challenges in the development and implementation of artificial intelligence (AI) and machine learning solutions. Identifying and addressing these challenges is critical to ensuring the continued growth and accuracy of these models as well as their equitable and effective use in health care settings. OBJECTIVE: This study aims to explore the challenges, opportunities, and potential solutions related to missingness in health care data for AI applications through the conduct of a digital conference and thematic analysis of conference proceedings. METHODS: A digital conference was held in September 2022, attracting 861 registered participants, with 164 (19%) attending the live event. The conference featured presentations and panel discussions by experts in AI, machine learning, and health care. Transcripts of the event were analyzed using the stepwise framework of Braun and Clark to identify key themes related to missingness in health care data. RESULTS: Three principal themes-data quality and bias, human input in model development, and trust and privacy-emerged from the analysis. Topics included the accuracy of predictive models, lack of inclusion of underrepresented communities, partnership with physicians and other populations, challenges with sensitive health care data, and fostering trust with patients and the health care community. CONCLUSIONS: Addressing the challenges of data quality, human input, and trust is vital when devising and using machine learning algorithms in health care. Recommendations include expanding data collection efforts to reduce gaps and biases, involving medical professionals in the development and implementation of AI models, and developing clear ethical guidelines to safeguard patient privacy. Further research and ongoing discussions are needed to ensure these conclusions remain relevant as health care and AI continue to evolve.

Opportunities, Challenges, and Future Directions of Generative Artificial Intelligence in Medical Education: Scoping Review.

BACKGROUND: Generative artificial intelligence (AI) technologies are increasingly being utilized across various fields, with considerable interest and concern regarding their potential application in medical education. These technologies, such as Chat GPT and Bard, can generate new content and have a wide range of possible applications. OBJECTIVE: This study aimed to synthesize the potential opportunities and limitations of generative AI in medical education. It sought to identify prevalent themes within recent literature regarding potential applications and challenges of generative AI in medical education and use these to guide future areas for exploration. METHODS: We conducted a scoping review, following the framework by Arksey and O'Malley, of English language articles published from 2022 onward that discussed generative AI in the context of medical education. A literature search was performed using PubMed, Web of Science, and Google Scholar databases. We screened articles for inclusion, extracted data from relevant studies, and completed a quantitative and qualitative synthesis of the data. RESULTS: Thematic analysis revealed diverse potential applications for generative AI in medical education, including self-directed learning, simulation scenarios, and writing assistance. However, the literature also highlighted significant challenges, such as issues with academic integrity, data accuracy, and potential detriments to learning. Based on these themes and the current state of the literature, we propose the following 3 key areas for investigation: developing learners' skills to evaluate AI critically, rethinking assessment methodology, and studying human-AI interactions. CONCLUSIONS: The integration of generative AI in medical education presents exciting opportunities, alongside considerable challenges. There is a need to develop new skills and competencies related to AI as well as thoughtful, nuanced approaches to examine the growing use of generative AI in medical education.

The evolving role of emergency medicine in family planning services.

PURPOSE OF REVIEW: The emergency department serves as an essential access point for a variety of healthcare services. This review will examine the recent expansion of family planning and reproductive health services in the emergency department. RECENT FINDINGS: An increasing number of patients present to emergency departments for early pregnancy loss (EPL), abortion care, and contraceptive management. Availability of comprehensive EPL management varies dramatically, possibly due to lack of provider knowledge or training. Particularly in remote settings, educational interventions - such as providing information about medication management and training in uterine aspiration - may standardize this management. Restrictive abortion laws raise concerns for changing and increased patient presentations to the emergency department for complications related to unsafe or self-managed abortion. Emergency medicine providers should anticipate that more patients will present without a prior ultrasound confirming intrauterine pregnancy prior to initiating no-touch or self-managed abortions. Particularly among pediatric patients, there may be a role for contraceptive counseling during an emergency department visit. Novel strategies, including web-based interventions and emergency department-based curricula for contraceptive counseling, may help reach those who otherwise may not seek reproductive healthcare in a clinic setting. SUMMARY: The intersection of emergency medicine and reproductive healthcare is a promising frontier for providing immediate, patient-centered, family planning care. Continued research and provider education are necessary to refine these approaches, address disparities, and respond to the changing reproductive healthcare landscape.

Characteristics of Emergency Medicine Residency Programs With Unfilled Positions in the 2023 Match.

STUDY OBJECTIVE: The unprecedented number of unfilled emergency medicine post-graduate year 1 (PGY-1) residency positions in the 2023 National Resident Matching Program shocked the emergency medicine community. This study investigates the association between emergency medicine program characteristics and the likelihood of unfilled positions in the 2023 Match. METHODS: This cross-sectional, observational study examined 2023 National Resident Matching Program data, focusing on program type, length, location, size, proximity to other programs, previous American Osteopathic Association (AOA) accreditation, first accreditation year, and emergency department ownership structure. We constructed a generalized linear mixed model with a logistic linking function to determine predictors of unfilled positions. RESULTS: A total of 554 of 3,010 (18.4%) PGY-1 positions at 131 of 276 (47%) emergency medicine programs went unfilled in the 2023 Match. In our model, predictors included having unfilled positions in the 2022 Match (odds ratio [OR] 48.14, 95% confidence interval [CI] 21.04 to 110.15), smaller program size (less than 8 residents, OR 18.39, 95% CI 3.90 to 86.66; 8 to 10 residents, OR 6.29, 95% CI 1.50 to 26.28; 11 to 13 residents, OR 5.88, 95% CI 1.55 to 22.32), located in the Mid Atlantic (OR 14.03, 95% CI 2.56 to 77.04) area, prior AOA accreditation (OR 10.13, 95% CI 2.82 to 36.36), located in the East North Central (OR 6.94, 95% CI 1.25 to 38.47) area, and corporate ownership structure (OR 3.21, 95% CI 1.06 to 9.72). CONCLUSION: Our study identified 6 characteristics associated with unfilled emergency medicine residency positions in the 2023 Match. These findings may be used to guide student advising and inform decisions by residency programs, hospitals, and national organizations to address the complexities of residency recruitment and implications for the emergency medicine workforce.

Teaching Principles of Medical Innovation and Entrepreneurship Through Hackathons: Case Study and Qualitative Analysis.

BACKGROUND: Innovation and entrepreneurship training are increasingly recognized as being important in medical education. However, the lack of faculty comfort with the instruction of these concepts as well as limited scholarly recognition for this work has limited the implementation of curricula focused on these skills. Furthermore, this lack of familiarity limits the inclusion of practicing physicians in health care innovation, where their experience is valuable. Hackathons are intense innovation competitions that use gamification principles to increase comfort with creative thinking, problem-solving, and interpersonal collaboration, but they require further exploration in medical innovation. OBJECTIVE: To address this, we aimed to design, implement, and evaluate a health care hackathon with 2 main goals: to improve emergency physician familiarity with the principles of health care innovation and entrepreneurship and to develop innovative solutions to 3 discrete problems facing emergency medicine physicians and patients. METHODS: We used previously described practices for conducting hackathons to develop and implement our hackathon (HackED!). We partnered with the American College of Emergency Physicians, the Stanford School of Biodesign, and the Institute of Design at Stanford (d.school) to lend institutional support and expertise in health care innovation to our event. We determined a location, time frame, and logistics for the competition and settled on 3 use cases for teams to work on. We planned to explore the learning experience of participants within a pragmatic paradigm and complete an abductive thematic analysis using data from a variety of sources. RESULTS: HackED! took place from October 1-3, 2022. In all, 3 teams developed novel solutions to each of the use cases. Our investigation into the educational experience of participants suggested that the event was valuable and uncovered themes suggesting that the learning experience could be understood within a framework from entrepreneurship education not previously described in relation to hackathons. CONCLUSIONS: Health care hackathons appear to be a viable method of increasing physician experience with innovation and entrepreneurship principles and addressing complex problems in health care. Hackathons should be considered as part of educational programs that focus on these concepts.

Creating a Safe Space for Simulation: Is it Time to Stop Calling Them Confederates?

SUMMARY STATEMENT: Use of the term "confederate" is often used in research literature to describe an individual allied with the research team. Confederate is used in simulation research to describe participants allied with the simulation facilitator. Confederate can also refer to the Confederate States of America and has connotations of racial injustice and slavery. Use of this term in simulation may adversely affect psychological safety of learners. Use of the term within the literature is a potential driver of use during simulation sessions. We completed a rapid review of the health care simulation literature to determine the frequency of the term confederate. From 2000 to 2021, 2635 uses of confederate were identified in 765 articles. There seems to be an increased trend in use of this word. We argue that alternative terms exist and should be used to maximize psychological safety of learners.

Identifying Social Media Competencies for Health Professionals: An International Modified Delphi Study to Determine Consensus for Curricular Design.

STUDY OBJECTIVE: The use of social media by health professionals is widespread. However, there is a lack of training to support the effective use of these novel platforms that account for the nuances of an effective health and research communication. We sought to identify the competencies needed by health care professionals to develop an effective social media presence as a medical professional, with the goal of building a social media curriculum. METHODS: We conducted a modified Delphi study, utilizing Kraiger's Knowledge, Skills, and Attitudes framework to identify appropriate items for inclusion in a social media curriculum targeted at health care professionals. Experts in this space were defined as health care professionals who had delivered workshops, published papers, or developed prominent social media tools/accounts. They were recruited through a multimodal campaign to complete a series of 3 survey rounds designed to build consensus. In keeping with prior studies, a threshold of 80% endorsement was used for inclusion in the final list of items. RESULTS: Ninety-eight participants met the expert criteria and were invited to participate in the study. Of the 98 participants, 92 (94%) experts completed the first round; of the 92 experts who completed the first round, 83 (90%) completed the second round; and of the 83 experts who completed the second round, 81 (98%) completed the third round of the Delphi study. Eighteen new items were suggested in the first survey and incorporated into the study. A total of 46 items met the 80% inclusion threshold. CONCLUSION: We identified 46 items that were believed to be important for health care professionals using social media. This list should inform the development of curricular activities and objectives.

Symptomatic Respiratory Virus Infection and Chronic Lung Allograft Dysfunction.

BACKGROUND: Chronic lung allograft dysfunction (CLAD) is a major cause of allograft loss post-lung transplantation. Prior studies have examined the association between respiratory virus infection (RVI) and CLAD were limited by older diagnostic techniques, study design, and case numbers. We examined the association between symptomatic RVI and CLAD using modern diagnostic techniques in a large contemporary cohort of lung transplant recipients (LTRs). METHODS: We retrospectively assessed clinical variables including acute rejection, cytomegalovirus pneumonia, upper and lower RVI, and the primary endpoint of CLAD (determined by 2 independent reviewers) in 250 LTRs in a single university transplantation program. Univariate and multivariate Cox models were used to analyze the relationship between RVI and CLAD in a time-dependent manner, incorporating different periods of risk following RVI diagnosis. RESULTS: Fifty patients (20%) were diagnosed with CLAD at a median of 95 weeks post-transplantation, and 79 (32%) had 114 episodes of RVI. In multivariate analysis, rejection and RVI were independently associated with CLAD (adjusted hazard ratio [95% confidence interval]) 2.2 (1.2-3.9), P = .01 and 1.9 (1.1-3.5), P = .03, respectively. The association of RVI with CLAD was stronger the more proximate the RVI episode: 4.8 (1.9-11.6), P < .01; 3.4 (1.5-7.5), P < .01; and 2.4 (1.2-5.0), P = .02 in multivariate analysis for 3, 6, and 12 months following RVI, respectively. CONCLUSIONS: Symptomatic RVI is independently associated with development of CLAD, with increased risk at shorter time periods following RVI. Prospective studies to characterize the virologic determinants of CLAD and define the underlying mechanisms are warranted.

A patient self-collection method for longitudinal monitoring of respiratory virus infection in solid organ transplant recipients.

BACKGROUND: Methods for the longitudinal study of respiratory virus infections are cumbersome and limit our understanding of the natural history of these infections in solid organ transplant (SOT) recipients. OBJECTIVES: To assess the feasibility and patient acceptability of self-collected foam nasal swabs for detection of respiratory viruses in SOT recipients and to define the virologic and clinical course. STUDY DESIGN: We prospectively monitored the course of symptomatic respiratory virus infection in 18 SOT patients (14 lung, 3 liver, and 1 kidney) using patient self-collected swabs. RESULTS: The initial study sample was positive in 15 patients with the following respiratory viruses: rhinovirus (6), metapneumovirus (1), coronavirus (2), respiratory syncytial virus (2), parainfluenza virus (2), and influenza A virus (2). One hundred four weekly self-collected nasal swabs were obtained, with a median of 4 samples per patient (range 1-17). Median duration of viral detection was 21 days (range 4-77 days). Additional new respiratory viruses detected during follow-up of these 15 patients included rhinovirus (3), metapneumovirus (2), coronavirus (1), respiratory syncytial virus (1), parainfluenza virus (1), and adenovirus (1). Specimen collection compliance was good; 16/18 (89%) patients collected all required specimens and 79/86 (92%) follow-up specimens were obtained within the 7 ± 3 day protocol-defined window. All participants agreed or strongly agreed that the procedure was comfortable, simple, and 13/14 (93%) were willing to participate in future studies using this procedure. CONCLUSION: Self-collected nasal swabs provide a convenient, feasible, and patient-acceptable methodology for longitudinal monitoring of upper respiratory virus infection in SOT recipients.