Explore the vast range of titles available.

The U.S. government recently took steps to ensure that clinical decision support algorithms are safe for clinical use. The next and larger step will be teaching physicians how to use the algorithms effectively.

Artificial intelligence could have broad implications for medical education. Educators could lead the way when it comes to integrating this technology into clinical practice.

Emerging evidence suggests generative artificial intelligence (AI) may offer potential for autoimmune and rheumatic disease care, moving beyond traditional narrow AI applications to produce contextualized clinical content to support a wide spectrum of medical tasks. This article explores generative AI applications across autoimmune and rheumatologic clinical care, research, and administrative domains. However, significant implementation challenges remain, including clinical validation, model interpretability, data integration complexities, and evolving regulatory frameworks.

Background Sustainable Development Goals (SDGs) are universally recognised targets designed to combat poverty, inequality, and climate change. However, there exists limited awareness and understanding of these goals among nursing and midwifery students. To address this knowledge gap, a co-designed audio podcast was introduced as an educational tool to enhance students’ comprehension of SDGs and their relevance to the healthcare profession. Methods A prospective study was conducted at Queen’s University Belfast, involving 566 first-year nursing and midwifery students. A 60-minute SDG podcast, co-designed with students and stakeholders, was made accessible within the university’s learning management system. Pre- and post-test questionnaires were administered to assess changes in students’ knowledge levels and attitudes toward SDGs. Additionally, 37 participants engaged in focus group interviews six months after listening to the podcast to explore their experiences and reflections. Quantitative data was analysed using paired t-tests and descriptive statistics, while qualitative data was analysed thematically. Results The podcast significantly increased students’ awareness of SDGs and their understanding of the goals’ relevance to their profession and personal lives. Post-test scores showed substantial improvements across all three sub-scales: knowledge, professional relevance, and personal relevance. Moreover, participants rated the podcast as a valuable learning resource with high acceptability, although some expressed uncertainty about replay intentions. Focus group interviews revealed three themes, including 1) More than you know’, which described how participants developed new knowledge and understanding about SDGs, 2) ‘Nurse-Midwife Nudges’, which illuminated how participants made minor changes to their behaviour and 3) ‘Fitting Format’, which highlighted how participants favoured the use of an audio podcast to learn about SDGs. Discussion This study demonstrates the potential of audio podcasts as an effective and engaging tool for increasing awareness and understanding of SDGs among nursing and midwifery students. The results suggest that such interventions can positively impact students’ knowledge, attitudes, and behavioural intentions regarding the SDGs. The findings also emphasise the importance of co-design in developing educational resources tailored to the specific needs and preferences of students.

Increased clinical demands and newer means of self-directed learning (SDL) necessitate an understanding of how medical residents are supporting their learning. To examine the patterns of SDL engagement among internal medicine residents, their attitudes and behaviors with various resources, and evaluate the relationship between the clinical learning environment (CLE) and the time residents allocate to SDL and types of resources. This cross-sectional study used a systematic questionnaire informed by previous qualitative research on SDL among internal medicine residents. Internal medicine (IM) residents from 10 residency programs across the United States participated, providing a diverse representation of geographical and institutional contexts. Residents were asked to estimate weekly hours spent on SDL during their last clinical rotation, on which resources, and then to rank the usefulness of each resource. The survey also measured several variables, including attitudes and behaviors after using the resource they perceived to be the most useful, and the influence of training level, residency program type, clinical rotation, and number of hours worked clinically per week on reported time spent on SDL and types of resources. The response rate was 69.5% (783/1,126). Residents dedicated a mean of 18.2 (SD 18.6) hours per week (median of 10.5 hours per week) to SDL. Community-based programs reported more hours of SDL. There was no difference in hours spent on SDL based on the last clinical rotation, number of hours worked clinically, or PGY level. Senior residents favored digital resources, like podcasts, and were less likely to use traditional resources, like textbooks than interns. Our findings underscore the substantial time residents devote to SDL. In light of these results, educators and healthcare systems will need to work together to better support residents in optimizing the complex clinical learning environment.

Early studies on these technologies stunned the clinical reasoning community, showing proficiency in tasks that had previously only been in the domain of humans, such as solving complex cases, triaging patients in the emergency room, forecasting diagnoses, making complex management decisions in uncertain cases, and even taking a history directly from patient actors. Some of the results for testing LLMs in this space might be a reflection of materials in the models’ pretraining and to date there have been no large-scale, prospective, clinical trials investigating patient outcomes. Doctors, dealing with the demands of everyday patient care, knowing well the imperfections of human clinical reasoning, may soon have to face an uncomfortable reality when an LLM consistently exhibits superhuman performance on human tests of reasoning.

Medical internship is the final year of training before independent practice for most doctors in Botswana. Internship training in Botswana faces challenges including variability in participants' level of knowledge and skill related to their completion of medical school in a variety of settings (both foreign and domestic), lack of planned curricular content, and limited time for structured educational activities. Data on trainees' opinions regarding the content and delivery of graduate medical education in settings like Botswana are limited, which makes it difficult to revise programs in a learner-centered way. To understand the perceptions and experiences of a group of medical interns in Botswana, in order to inform a large curriculum initiative. We conducted a targeted needs assessment using structured interviews at one district hospital. The interview script included demographic, quantitative, and free- response questions. Fourteen interns were asked their opinions about the content and format of structured educational activities, and provided feedback on the preferred characteristics of a new curriculum. Descriptive statistics were calculated. In the current curriculum, training workshops were the highest-scored teaching format, although most interns preferred lectures overall. Specialists were rated as the most useful teachers, and other interns and medical officers were rated as average. Interns felt they had adequate exposure to content such as HIV and tuberculosis, but inadequate exposure to areas including medical emergencies, non-communicable diseases, pain management, procedural skills, X-ray and EKG interpretation, disclosing medical information, and identifying career goals. For the new curriculum, interns preferred a structured case discussion format, and a focus on clinical reasoning and procedural skills. This needs assessment identified several foci for development, including a shift toward interactive sessions focused on skill development, the need to empower interns and medical officers to improve teaching skills, and the value of shifting curricular content to mirror the epidemiologic transition occurring in Botswana. Interns' input is being used to initiate a large curriculum intervention that will be piloted and scaled nationally over the next several years. Our results underscore the value of seeking the opinion of trainees, both to aid educators in building programs that serve them and in empowering them to direct their education toward their needs and goals.

Large Language Models and the Medical RecordInstead of facilitating communication and transparency, the insertion of LLM-generated text directly into the medical record risks diminishing the quality, efficiency, and humanity of health care.

This comparative effectiveness research study examines the association between racial differences in pain assessment and false beliefs about biologization of race by large language models compared with a human baseline.

Large language models (LLMs) have shown promise in their performance on both multiple-choice and open-ended medical reasoning examinations, but it remains unknown whether the use of such tools improves physician diagnostic reasoning. To assess the effect of an LLM on physicians' diagnostic reasoning compared with conventional resources. A single-blind randomized clinical trial was conducted from November 29 to December 29, 2023. Using remote video conferencing and in-person participation across multiple academic medical institutions, physicians with training in family medicine, internal medicine, or emergency medicine were recruited. Participants were randomized to either access the LLM in addition to conventional diagnostic resources or conventional resources only, stratified by career stage. Participants were allocated 60 minutes to review up to 6 clinical vignettes. The primary outcome was performance on a standardized rubric of diagnostic performance based on differential diagnosis accuracy, appropriateness of supporting and opposing factors, and next diagnostic evaluation steps, validated and graded via blinded expert consensus. Secondary outcomes included time spent per case (in seconds) and final diagnosis accuracy. All analyses followed the intention-to-treat principle. A secondary exploratory analysis evaluated the standalone performance of the LLM by comparing the primary outcomes between the LLM alone group and the conventional resource group. Fifty physicians (26 attendings, 24 residents; median years in practice, 3 [IQR, 2-8]) participated virtually as well as at 1 in-person site. The median diagnostic reasoning score per case was 76% (IQR, 66%-87%) for the LLM group and 74% (IQR, 63%-84%) for the conventional resources-only group, with an adjusted difference of 2 percentage points (95% CI, -4 to 8 percentage points; P = .60). The median time spent per case for the LLM group was 519 (IQR, 371-668) seconds, compared with 565 (IQR, 456-788) seconds for the conventional resources group, with a time difference of -82 (95% CI, -195 to 31; P = .20) seconds. The LLM alone scored 16 percentage points (95% CI, 2-30 percentage points; P = .03) higher than the conventional resources group. In this trial, the availability of an LLM to physicians as a diagnostic aid did not significantly improve clinical reasoning compared with conventional resources. The LLM alone demonstrated higher performance than both physician groups, indicating the need for technology and workforce development to realize the potential of physician-artificial intelligence collaboration in clinical practice. ClinicalTrials.gov Identifier: NCT06157944.