Jeremy Richards, MD, MA, FACP, ATSF

Training to Teach in Medicine | Harvard Medical School Professional, Corporate, and Continuing Education Gain the skills to teach and inspire future health care professionals using innovative, evidence-based strategies in diverse learning settings.

Learn more about Training to Teach in Medicine here: learn.hms.harvard.edu/programs/tra...

And, the Med Sci Educ full text link for Thesen T et al. Medical Science Educator. 2024: tinyurl.com/nh9jrzcr

Reference: Abd-Alrazaq A et al. JMIR Med Educ. 2023;9:e48291: mededu.jmir.org/2023/1/e48291/

The takeaway isn't the AI tool, but the *pedagogy*:
✔ intentional case design
✔ structured feedback
✔ alignment w/ learning objectives

Programs like Training to Teach in Medicine at @harvardmed.bsky.social helps clinicians develop skills for teaching emerging technologies in #MedEd

Notably, the platform supports text & voice, it is multilingual, & it is freely accessible - advancing equity across globally diverse & resource-limited institutions. This positions #AI patient actors as complements, not replacements, to human standardized patients.

Learners receive immediate, rubric-based formative feedback from the AI, supporting deliberate practice in:
• clinical interviewing
• diagnostic reasoning
• communication skills

This aligns with evidence linking feedback in sim to clinical performance: onlinelibrary.wiley.com/doi/10.1111/...

Key design insight: The LLM does *NOT* generate medical facts. Rather, human-authored cases constrain the AI’s responses - reducing hallucinations & preserving educator control over content, a critical principle for responsible AI adoption in #MedEd, addressing known risks of LLMs.

#meded #AI #sim

Standardized patients teach clinical communication & reasoning, but are resource-intensive & inequitable at scale. An article in Med Sci Educ introduces an open-access AI Patient Actor to enable scalable self-paced clinical interviewing w/ formative feedback.

Link: doi.org/10.1007/s406...

Training to Teach in Medicine | Harvard Medical School Professional, Corporate, and Continuing Education Gain the skills to teach and inspire future health care professionals using innovative, evidence-based strategies in diverse learning settings.

A link to this article (not full text 🙁) is here link.springer.com/article/10.1...

more info about
@scholarrx
Bricks 🧱 is here scholarrx.com/brick-exchan...

and more info about Training to Teach in Medicine 😀can be found at learn.hms.harvard.edu/programs/tra...
@harvardmed.bsky.social
#meded

If you’re redesigning curricula, consider integrating:
• jigsaw or team-based peer learning
• interactive modular content
• structured assessment

Programs like Training to Teach in Medicine @harvardmed.bsky.social help clinicians learn about evidence-based teaching strategies #meded

Key takeaways for #meded:
→ METHOD matters (active peer-assisted strategies outperform lecture)
→ RESOURCES matter (high-quality modular tools work)
→ Students value CONTROL & COLLABORATION

Aligns with competency-based education & constructivist learning theory #medtwitter

Modules such as @scholarrx.bsky.social Bricks enhanced outcomes across groups, likely due to:
- concise organization
- interactive visuals
- cognitive load reduction

Prior work shows AV/multimodal tools improve memory & engagement (Pujol et al, Brain Behav 2021; doi:10.1002/brb3.1427).

The jigsaw strategy (learn in expert group → teach peers in home group) promoted:
✔️engagement
✔️communication

✔️self-directed learning
✔️accountability
Findings align with prior evidence that collaborative learning improves retention and performance (Oakes et al, Sagsoz et al)

A 4-arm randomized pilot trial in Medical Science Educator compared jigsaw peer-assisted learning vs lectures +/- modular digital resources (@scholarrx.bsky.social Bricks)

Result: active, peer-led methods improved knowledge, motivation & perceptions of learning
link.springer.com/article/10.1...

Medical educators planning curricula can find value in structured programs like Training to Teach in Medicine, which equips clinicians with evidence-based teaching strategies - including how to incorporate practice-relevant competencies such as nutrition.

learn.hms.harvard.edu/programs/tra...

Why this matters 🤔 -> diet quality drives risk for heart disease, diabetes, obesity, cancer & more diseases - yet most physicians report insufficient nutrition training. Aligning education with practice can improve patient care & population health.

@nutrisci.bsky.social @hsph.harvard.edu

Experts recommend integrating these #nutrition competencies into #UME & #GME curricula, assessed via OSCEs, simulated cases, boards & student feedback - signaling a shift toward tested competence, not just *optional* content.

#MedTwitter #MedEd @harvardmed.bsky.social

Consensus priorities include:
• Evidence-based, culturally sensitive nutrition guidance
• Screening for food & nutrition insecurity
• Collaborative multidisciplinary nutrition care
• Nutrition assessment via diet history and lab data
(100% expert agreement on many)

@atsmeded.bsky.social

Nutrition is central to chronic disease prevention and care, yet historically under-emphasized in #MedEd. A consensus in JAMA Network Open from 9.24 outlines 36 essential nutrition competencies for medical students & physician trainees.

jamanetwork.com/journals/jam... @jamanetworkopen.com

Improv in health professions education special interest group: Creating community and collaboration

Again, here is a link to the full text of this brief, descriptive & interesting article in "Patient Education & Counseling": www.sciencedirect.com/science/arti...

And kudos to @amybzelenski.bsky.social (cc @wiscmedphys.bsky.social @atsmeded.bsky.social )

www.sciencedirect.com/journal/pati...

For educators in #MedEd: improve is a dynamic & "evidence-informed" teaching approach to consider incorporating improv principles like “Yes, and…” spontaneous response exercises, & reflective debrief into teaching to strengthen communication skills, team dynamics, &rapport.

In health professions settings, #improv can support interprofessional communication, foster psychological safety, & encourage collaborative problem-solving to help learners join in dialogue without waiting for the “perfect script.”

@atsmeded.bsky.social #MedEd @harvardmed.bsky.social

Improv training enhances active listening, empathy, perspective-taking & adaptability, bridging gaps between scripted communication curricula & real-world patient encounters where uncertainty & complexity are the norm. It helps learners respond to uncertainty in the moment. . .

Medical improv isn’t *comedy* - it’s a structured, experiential method adapted from theater to build communication, adaptability, listening, & teamwork skills. An interesting article from @amybzelenski.bsky.social & colleagues re principles of improv in #MedEd

www.sciencedirect.com/science/arti...

Assessment of Large Language Models in Clinical Reasoning: A Novel Benchmarking Study Large language models (LLMs) are increasingly deployed for clinical decision support, yet standard evaluations such as medical licensing exams overlook how clinicians update decisions in dynamic co...

Once more - here's a link to the #fulltext manuscript at @NEJM_AI cc'ing @adamrodmanmd.bsky.social & @liamgmccoy.bsky.social as authors of this interesting #clinicalreasoning benchmarking study:

ai.nejm.org/doi/full/10....

Studies like this emphasize why AI belongs inside #MedEd - but only with intentional pedagogy

In Training to Teach in Medicine at @harvardmed.bsky.social we focus on ways to integrate tech like #AI while preserving clinical reasoning, judgment & safety

learn.hms.harvard.edu/programs/tra...

Educational takeaway: If learners rely on AI tools, we must explicitly teach reasoning evaluation, not just answer validation.

This study strengthens the case for:
• Teaching metacognition
• Assessing reasoning processes
• Designing AI-aware curricula

@bidmcreseach.bsky.social

Key insight from this study: LLMs can often reach correct answers despite flawed or inconsistent reasoning paths - and may sound confident even when logic is weak.

For educators, this reinforces a core principle:
Correctness ≠ sound clinical reasoning

#MedEd #MedTwitter

Unlike prior evaluations focused mainly on final answers, this study benchmarks LLMs using structured clinical reasoning tasks: testing diagnostic logic, stepwise reasoning, & vulnerability to cognitive errors.

ai.nejm.org/doi/full/10....

LLMs are increasingly used for clinical support & #meded but how well do they actually reason?

A study in NEJM AI from @adamrodmanmd.bsky.social et al rigorously evaluates LLM performance on clinical reasoning tasks, moving beyond superficial accuracy to see HOW answers are generated

To deepen #meded expertise in inclusive teaching design & policy implementation, structured development helps: Training to Teach in Medicine @harvardmed.bsky.social explores policy, pedagogy, & eval frameworks that elevate educational systems. learn.hms.harvard.edu/programs/tra... #MedTwitter

Implementing inclusive policies intersects with broader educational quality work: competency frameworks, assessment validity & program eval - themes core to GME scholarship & pedagogy. Tools like this can be platforms for scholarly inquiry & #meded research @bidmcreseach.bsky.social