Training in the Age of Generative AI

Residency Reflections #10

Author

Jared Dashevsky, MD
November 27, 2024 • Estimated Reading Time: 9 minutes

RESIDENCY REFLECTIONS

Training in the Age of Generative AI

I’m training during two inflection points in medicine, and it’s making residency an incredibly exciting journey. The first is the rise of generative AI—its accessibility and impact are reshaping workflows and redefining what’s possible in patient care. The second is the introduction of GLP-1 medications, which are rewriting the script for tackling metabolic syndrome. I’ll dedicate two articles to reflecting on these pivotal changes.

In this article, I’m diving into the experience of training alongside the rapid evolution of generative AI.

My residency began just months after OpenAI launched ChatGPT. Back then, I wrote an article on the healthcare use cases, predicting three areas that would benefit most from generative AI:

  • Patient care and delivery

  • Research, diagnostics, and treatment

  • Clinical and non-clinical workflow

I won’t rehash the details—you can check out the full article here.

So far, I’ve leaned heavily on AI for “research, diagnostics, and treatment” and “clinical and non-clinical workflows.” The first category, “patient care and delivery,” is still waiting on meaningful EHR integration to make a real impact.

DoximityGPT and OpenEvidence have truly been the most impactful platforms for me. In fact, here was the first sentence of my article I wrote on the Top AI Medical Search Platforms Transforming Healthcare in 2024:

Artificial intelligence has 10x the speed and accuracy of my research, both in clinical practice and for Healthcare Huddle. Research questions that once took me 30 minutes to an hour to find answers for can now be accomplished in mere seconds. This rapid access to precise, evidence-based answers from medical-focused large language models (LLMs) has significantly contributed to my growth and efficiency as a physician.

And I truly meant every word.

When I have a research question, I toss it into OpenEvidence. Within seconds, I’ve got an answer, backed by primary sources I can verify and dive into for deeper understanding. Compare that to the old way—searching PubMed, flipping through Harrison’s textbook—which, while rewarding, was a massive time sink. As a resident working 70+ hours a week and running Healthcare Huddle, I need a streamlined research workflow.

I love OpenEvidence so much, that I’ve gotten several of my resident colleagues to also love it. Now, we nerd out over it. If there’s a debate—like “what’s the efficacy of albumin resuscitation in shock?”—someone inevitably says, “Let’s see what OpenEvidence says.” We’ll dig into the sources it cites, from RCTs to clinical guidelines (and yes, the albumin example actually happened).

For administrative tasks, DoximityGPT is my go-to. Writing a medical necessity letter for durable medical equipment? Easy. I give it some de-identified patient history and a few bullet points about why the equipment is needed, and in seconds, I’ve got an elegant letter ready to go. What used to take me 30 minutes of mental effort now takes less than five. My motto: work smarter, not harder, especially when it comes to admin work.

Eventually, tools like Pieces will take this efficiency to the next level, summarizing hospital courses, writing discharge summaries, optimizing billing, and more.

There’s a lot of hoopla about AI “taking over” physicians. I’ll save my full thoughts for another article, but here’s the short version: the idea that AI will replace physicians is the wrong mindset. Instead, we need to focus on how AI augments our ability to care for patients. That’s the real conversation we should be having.

Training during the rise of generative AI feels like being on the front lines of a technological revolution in medicine. As a resident, this means learning to balance tradition with innovation, embracing new technologies while staying grounded in the fundamentals of human connection and clinical judgment. It’s an exciting, challenging time to train, but more than anything, it’s a reminder that the future of medicine will always require both cutting-edge tools and compassionate care.

Enjoy this Reflection? Share it with your colleagues:

INSIDE THE HUDDLE

Healthcare Huddle
Sunday Newsletter

Huddle+
Inefficiency Insights

Huddle+
Huddle #Trends

Huddle+
Huddle University

Healthcare Providers
Residency Reflections

COMMUNITY SPOTLIGHT

Upcoming Event

During our coffee break, let's catch up on this week's healthcare happenings. We'll take a quick 15 minutes to recap what we've seen, from fundraising to policy.

Reply

or to participate.