Welcome back to season 10 of Write Medicine, where we explore best practices in creating continuing education for health professionals. 

I must say it was spectacular taking time off during the last couple of weeks of 2024. I spent some of that time reflecting and reviewing my business and my work in 2024 and planning strategy and goals for 2025. Like many of you, this is a quarterly practice for me, and it’s a practice I teach in WriteCME Pro, but the end of the year is special because it’s that point when, hopefully, you finally see many different threads coming together. That was certainly my experience this year, and I’m grateful for that as WriteCME Pro moves into its third year. During my vacation time, I watched a ton of movies like the Order, the Substance, Heretic, Juror #2, Wicked and more. I highly recommend each and every one of these if you get a chance to see them; they’re absolutely the kind of movie I love with a stellar cast, a moral dilemma at their core, and a story arc that encourages us to stop and think about our relationships to ourselves, to others, and to our wider communities.

And then, in the first week of January, it was off to the Alliance for Continuing Education in the Health Professionals annual conference in Orlando, Florida. I had a great time catching up with old friends and clients, meeting WriteCME Pro members, presenting on panels with colleagues, and learning from the many, many wonderful sessions. Today, we're diving into highlights from the three keynote sessions, which were among the best I’ve heard in many years. 

Keynote 1: AI in Healthcare Education

The first keynote, delivered by Dr. Immani Shephard, tackled the complexities and challenges of AI in healthcare education. Dr. Shephard, Director of Medical Education and Social Scientific Research at the University of Illinois, cautioned that while AI is rapidly generating information, it's not necessarily producing better or more accurate information. Many of us playing around with AI will likely agree with that assessment. 

She emphasized that AI learns from the data it's fed, and if that data reflects existing biases, AI will perpetuate those biases, potentially leading to inequitable healthcare outcomes. One striking example she gave is the GFR algorithm, which was initially based on the flawed assumption that African Americans have inherently higher GFR levels due to greater muscle mass. This bias resulted in delayed interventions, prolonged dialysis, and reduced transplant opportunities for this population. Actually, medical content strategist Sara Fagerlie also reminded us about bias amplification in a panel we did together on AI ethics with Nuria Negrao and Morgani Leafe. A recent article in Nature focused on this bias amplification through human-AI feedback loops. This refers to the cyclical process where AI systems, trained on potentially biased human data, not only adopt those biases but often amplify them. When we then interact with these biased AI systems, our own biases are reinforced, creating a self-perpetuating cycle of error.

So imagine an AI system designed to assist with medical diagnoses. If the training data contains biases that associate certain demographics with particular health conditions, the AI will probably overemphasize these associations in its recommendations. Clinicians, relying on the AI's input, might then unknowingly reinforce these biases in their own decision-making, leading to disparities in care.

Authors of the Nature article emphasize that this bias amplification effect is significantly greater in human-AI interactions compared to human-human interactions. The difference arises because AI systems tend to be more sensitive to minor biases in data and might exploit them to improve prediction accuracy. Additionally, humans often perceive AI as more objective and authoritative than fellow humans, making us more susceptible to AI's influence.

The implications of these feedback loops are profound, particularly in healthcare where accurate and equitable decision-making is paramount. As Dr. Shephard argued in the Alliance keynote, we need to critically examine why we're using AI and ensure its applications don't exacerbate existing inequalities. She challenged us to question What it means to be human in an increasingly AI-driven world and to consider how we can educate learners on the equitable use of AI and address biases within the medical field. She suggested several strategies, including teaching AI to identify and mitigate biases from the outset, encouraging skepticism towards AI outputs, and promoting a nuanced understanding of data. She also stressed the importance of qualitative research methodologies in CME, advocating for training that equips clinicians to recognize subtle cues like body language and non-verbal communication.

To guide CME providers in navigating the AI landscape, Dr. Shephard offered a practical checklist for making decisions about where and how AI fits in our education programs, prompting reflection on factors such as equity in outcomes, data accuracy, inclusion of social determinants of health, and the potential for AI to either improve lives or exacerbate existing inequities.

We need to help our learners develop "AI literacy" and develop it ourselves. This means:

First, teaching healthcare professionals to question AI outputs rather than accepting them as truth. Every time we integrate AI tools into our educational content, we should build in reflection points about potential biases.

Second, we need to embrace the argument that AI should augment, not replace, human clinical judgment. Dr. Shephard noted that medical students are increasingly turning to AI for diagnoses. Our role is to help clinicians understand when to use AI and when to rely on their human expertise in patient care.

Keynote 2: Integrating Art History in Healthcare Education

Our second keynote speaker, Dr. Siobhan Conaty from La Salle University, explored a different approach to enhancing healthcare education by integrating art history into curricula. She highlighted its numerous benefits, including promoting physiological reset, boosting dopamine levels, and simply bringing joy to the learning process—a major conference theme.

The session underscored how art history methods can cultivate transferable skills applicable to healthcare practice, such as attention to detail, pattern recognition, and cultural sensitivity. Think about it - what if we could help clinicians slow down and observe more carefully? 

Key takeaways about these methods for CME professionals included:

• Close looking, which involves slow, methodical observation of art to describe what is seen, identify objects, and resist interpretation. This technique encourages focus and mindful visual thinking, prompting us as viewers to analyze a piece by asking questions such as, "What's going on in this picture?" and providing a rationale to explain what it is you think you see. 
• Formal analysis, which looks at the grammar and syntax of art, examining elements like line, color, shape, and composition. For instance, different types of lines can evoke various emotional effects, a finding supported by neurological studies. 
• Understanding proportion, including the use of distortion or exaggeration for symbolic purposes, can help healthcare professionals recognize and address potential cultural biases about the body. 
• Similarly, appreciating balance in art, achieved through elements like light, shade, and symmetry, can translate to enhanced skills in medical imaging.

Here's a concrete example Dr. Conaty shared. Start with an artwork - it could be a portrait or a complex scene. Ask three simple questions:

• What's going on in this picture?
• What do you see that makes you say that?
• What more can we find?

These same questions can transform how clinicians approach patient assessment.

Ultimately, integrating art history methodologies and ways of looking into continuing education offers a unique and engaging way to foster critical thinking, observation skills, and a deeper understanding of the human experience, all valuable assets for healthcare professionals. These are valuable assets for us, too, in the education planning, design, and implementation process.

And talking of joy, one of the conference themes this year, I ran a Friday session focused on practices to boost joy at the end of the day on Friday. I’ll save the full rundown for another episode, but suffice it  to say the session included plenty of opportunity to practice movement, mindfulness, and breath-awareness techniques that enhance cognitive capacity, increase energy, and increase psychological safety. I messed up a bit though. I thought the session was meant to run for 60 minutes and designed my approach accordingly. But I was not paying attention and it turns out the session was scheduled to run for 30 minutes. Here’s the thing. Only a couple of people left. This tells me just how much we need every opportunity we can get to practice relaxation and allow ourselves to rest and replenish.  

Keynote 3: Social Networks and Behavior Change

In our final keynote session, Dr. Damon Centola of the University of Pennsylvannia guided us through the fascinating world of social networks and their impact on driving behavior and wider social change. Dr. Centola debunked common myths about how change occurs. While we might cling the the idea that behavior change often originates from highly connected individuals at the center of social networks who spread ideas, which is the guiding principle of diffusion theory, a core tenet in our field for many, many years, he argued that change often starts at the periphery of social networks and addresses beliefs as well as ideas. And so what, you might ask. Well, this insight invites us to rethink traditional approaches to disseminating information and influencing clinical practice. As Andy Bowser asked in the Q+A, should we keep inviting expert faculty to operate as highly connected individuals at the center of a field or try a different strategy? If so, what might that strategy look like? 

Dr. Centola had an answer for us. He emphasizes the power of "complex contagion," where change spreads through multiple reinforcing interactions within a social network, that is, the networks of which we are members in everyday and professional life, not just digital networks. This means that peer influence and social norms are crucial in driving behavior change, much more so than individual persuasion, even if experts are doing the persuading. 

He described one study for instance, recently published in Proceedings of the National Academy of Sciences (PNAS) where groups of physicians were given medical scenarios and asked to make diagnoses and treatment recommendations. The researchers found that the groups who shared information with each and worked together made significantly fewer errors than those who worked alone. Complex contagion highlights that behavior change, like adopting a new medical practice, is more likely to occur through repeated interactions within a network, rather than individual persuasion.

Here's how the study explains complex contagion:

• Networked Learning: Physicians working in structured information-sharing networks were exposed to multiple viewpoints and insights from their peers. This collaborative learning environment improved decision-making.
• Reinforcing Interactions: The exchange of information and perspectives within the networks created multiple reinforcing interactions, strengthening the adoption of accurate diagnostic and treatment strategies.

In practical terms, instead of relying solely on high-profile speakers or thought leaders, we can design programs that include:

• Small group discussions where clinicians can process new information together
• Peer feedback mechanisms
• Cross-departmental learning communities

What were Dr. Centola’s takeaways for CME professionals? First, we can leverage the power of social networks by fostering peer learning, promoting diversity within learning cohorts, and building "wide bridge" connections that span across different groups and organizations. Second, he reminded us that learning is a social process that involves vulnerability and unlearning old habits. So creating supportive learning environments where individuals can learn together as a cluster can accelerate the adoption of new knowledge and practices.

All three keynotes presented a compelling case for embracing innovation while remaining critically aware of potential pitfalls. As CME professionals, we have a responsibility to ensure that new technologies and approaches serve to enhance equity and improve healthcare outcomes for all.

So, my takeaways from these three keynotes?

1. When incorporating AI tools, build in explicit discussion of potential biases and limitations
2. Consider adding art observation exercises to enhance clinical observation skills
3. Design programs that leverage peer networks rather than relying solely on top-down expertise

That's all for this episode of Write Medicine. If you're using any of these approaches in your CME programs or content, I'd love to hear about it. Share your experiences via voicemail on the podcast webpage, email or LinkedIn.  

Remember to subscribe to the podcast and our newsletter, Write Medicine Insider for more insights on creating impactful continuing medical education. Until next time, keep on learning!