Generative AI has a bright future in medical education. That goes not only for medical schools but also for postgraduate settings in which residents and fellows do most of their learning while also caring for patients.
“A core tenet of graduate medical education, or GME, is ‘graded authority and responsibility,’ where trainees progressively gain autonomy until they achieve the skills to practice independently,” several advanced GME trainees point out in a paper published this month in Frontiers in Medicine. “Additionally, trainees are expected to become ‘physician scholars.’”
What does GenAI have to do with any of that? As it turns out, plenty. The paper succinctly summarizes the relevant peer-reviewed literature on the subject and comments on risks as well as opportunities involving GenAI in the GME setting.
The authors are four clinical informatics fellows at the Baylor Scott & White health system in Texas. Here are excerpts from their section on opportunities.
1. EHR workload reduction.
Given their long work hours and stressful work environment, GME trainees are “particularly susceptible to burnout,” lead author Ravi Janumpally, MD, MHA, and colleagues point out, “with rates higher than their age-matched peers in non-medical careers and higher than early-career attending physicians.” More:
‘Given its ability to summarize, translate and generate text, GenAI demonstrates clear potential as a technological aid to alleviate the burden of clinical documentation.’
2. Clinical simulation.
Stakeholder interest is keen in the use of conversational GenAI to simulate patient encounters, although this application is more often focused on undergraduate medical education, the authors note. More:
‘Among the most interesting potential applications of GenAI in GME is the concept of using synthetic data as training material for visual diagnosis. For example, generative adversarial networks (GANs) and diffusion models have shown promise in generating realistic medical imaging data sets.’
3. Individual education.
“One-on-one tutoring delivered by humans is costly, and skilled teachers are not available everywhere, but GenAI tools may have some of the same benefits at a fraction of the cost,” Janumpally and colleagues write.
‘Large language models show promise as a tool for explaining challenging concepts to graduate medical trainees in a manner tailored to the learner’s level, and LLMs could be configured to act as personalized tutors.’
4. Research and analytics support.
GME trainees are required to participate in quality improvement (QI) projects, and these typically require quantitative data analysis, the authors note. “Trainees are often underrepresented in organizational QI activities, with one potential reason being the substantial time and effort needed for data collection and analysis.”
‘Large language models have some ability to facilitate straightforward data analysis and can generate serviceable code for statistical and programming tasks. LLMs are also adept at natural language processing tasks like extracting structured data from unstructured medical text.’
5. Clinical decision support.
GenAI for CDS is “an area of great potential and ways to improve performance are under development,” the authors write. However, they add, “GME faculty and trainees cannot yet rely on LLMs to directly guide clinical care.”
‘Studies done to evaluate the potential of LLMs for clinical decision support in various clinical contexts have shown mixed results so far, with limitations in their ability to handle nuanced judgment and highly specialized decision-making.’
This doesn’t mean GenAI has no future in CDS—just that the pairing needs more time and attention.
“LLMs can provide context-sensitive and specific guidance incorporating clinical context and patient data, they can be accessed through readily available communication channels, and—in contrast to rule-based alerts—they are interactive,” Janumpally and co-authors point out.
The paper is available in full for free.
