ER Case Simulation

When Experience Can’t Wait for Reality

The High-Stakes Puzzle

Emergency medicine residents face a paradox: they need expert clinical reasoning to avoid misdiagnosis, yet their learning depends on unpredictable case exposure and limited direct observation. With misdiagnosis rates reaching 20%, the traditional "learn as you go" approach leaves dangerous gaps in training that patients ultimately pay for.

Residents need more reps, which means more cases to study. This is a collection of cases, all with the same chief symptom but different root causes.

From Skepticism to Possibility

What began as a skeptical conversation between two ER physicians wondering if AI could generate medically accurate training cases transformed when I suggested something more ambitious. What if large language models could not only create realistic cases but enable interactive learning experiences that reveal the hidden cognitive processes behind diagnosis?

Each case has realistic details, accompanied by the patient's personality and backstory.

Making the Invisible Visible

The ER Case Simulation evolved from hypothesis to breakthrough: an AI system that generates clinically authentic emergency scenarios while residents interact naturally—asking questions, ordering tests, and working through diagnosis in real-time. Unlike traditional simulations, it captures and visualizes their moment-by-moment reasoning process, creating a "cognitive map" that expert faculty can use for targeted feedback.

Users can interact "verbally" with patients, while also requesting clinical tests and physical exam maneuvers. They're able to log their clinical reasoning along the way.

From Prototype to Proven Approach

The prototype's potential has been recognized with dedicated research funding for 2025, when a formal study at UC Davis will measure its impact on clinical reasoning development. What started as a speculative question has evolved into a structured investigation of how AI can transform medical education by giving residents the volume and variety of cases they need — the opportunity to develop adaptive expertise (see: Fostering Adaptive Expertise Through Simulation) — while making their thinking processes transparent to mentors.

Beyond emergency medicine, this approach opens new possibilities for accelerating expertise in any field where complex decision-making carries high stakes. The breakthrough isn't just in generating content—it's in revealing the normally hidden journey from information to insight, allowing systematic improvement of how professionals think, not just what they know.

Reach Out

Are you interested in discovering new training or up-skilling opportunities enabled by generative AI for your team or your organization?

Request a consultation