How to Create Interactive Patient Simulations (No Coding Required)

Imagine a world where every healthcare educator can code.  Even you.  Even if you’ve never coded before in your life.

"You think I have time to learn to code?”  Oh sure.  Right between grading exams, prepping lectures, attending committee meetings that really should’ve been emails, and spending time in the clinic. Side note: is being a health care educator possibly the busiest job on the planet?  But I digress.

AI is adding powers, not just for helping students - ahem - “perform better” on their assignments.  This time the superpower is yours to wield.

Now, I’m no tech bro.  I’ve never coded in my life (I learned a bit of html to build a website back in the day, but nothing substantial).  But using the latest model of Claude, I was able to generate a fully fledged interactive clinical encounter that I can now share with you.

So, I’d like to show you what I built and how to build something like it yourself. You know, with all that free time you have.

What You'll Need

Two quick tools to make students think you spent months building this.

1. A Patient Case to Repurpose

• Your own patient case (from the Word docs you’ve used for years), OR
• A free real patient case from ReelDx (requires a free login.  In case you haven’t heard: we rotate four public cases every few months, which are HIPAA-compliant and reviewed by our medical team, unlike, say, a random Youtube video).

2. Claude 3.7 Sonnet

• Never tried Claude? It’s like ChatGPT but with two key differences: 1. You can train it to sound like you (which is fun and slightly surreal to converse with yourself), and 2. With Claude Sonnet 3.7, you can publish anything - code, a document, an AI-created diagram - to a url that you can share with students and colleagues.
• Create a free account at anthropic.com/claude. A pro subscription starts at $18 a month, which mostly removes the character limits that can inhibit the progress of the coding (see below).

That’s it!  Claude can code and publish whatever scenario you dream up.

Try the version we made here:

It didn't take a sophisticated prompt (see below).  I first asked Claude for ideas about what interactive clinical experiences it could build and then asked it to do so.

As CEO of ReelDx, I have a library of HIPAA-compliant patient videos to pair such things with, which is what I did here.  The video case is hosted on our public.reeldx.com domain, but the interactive simulation code was created by Claude (and I posted it to our website).

One of the biggest challenges as CEO is applying our patient video cases in ways that serve a diverse set of learners: from medical schools to PA to NP to EMS. By giving educators a new way to bring patient scenarios to their students, AI now makes it easier than ever to use ReelDx video to ground the case in reality, and seamlessly blend it with Claude-powered interactivity.  Think of it as replacing a stack of flashcards with an interactive case study — students retain more without even realizing it. And with AI handling the setup… all that’s left is to teach.

Steps for Building Your Own Interactive Patient Case

Again, it seems daunting, but Claude is quite forgiving.  The prompt I used was… less than ideal but it got the job done.

Step 1: Prepare the patient case and use the following prompt

From this prompt alone, it came up with a simulation that involved:

• Branching Decision Paths
• Dynamic Vital Signs
• Immediate Feedback
• Clinical Notes Section

We followed-up by inserting the patient case data from ReelDx (without any identifying information, of course):

Prompt:

"Make the same scenario about the following patient ...

Gender:

Age:

Height:

Weight:

Vital Signs:

Signs and Symptoms:

History:

Differential Diagnosis:

Patient Workup:

Patient Outcomes: "

(I left the section above blank for you to fill in with your own patient case if you so choose.  I used the ReelDx case sections for those pieces of information since they’re real and peer-reviewed — harder for the AI to hallucinate if it’s based on real information).

Claude does the rest.  Well, sort of ...

Step 2: Patiently Ask Claude to Continue

Okay, here’s where Claude isn’t perfect yet.  It seems that Anthropic (the maker of Claude) made an annoying choice here: to limit the rate of Claude’s character generations when users aren’t present, it’s introduced a feature that will stop Claude in its tracks once it hits a certain character limit.  It looks like so:

Typing the word, “continue” will allow Claude to start generating again.  This game took four instances of “continue,” but your mileage may vary.

If you’re on the free plan, the limit is even stricter, which makes creating code tricky.  Generative AI still has a problem with creating more characters than you request, and that’s also true for coding.  What this means in practice is that if you’re going to execute this task, you’ll probably want a premium account.

Step 3: Need a patient case?  ReelDx has you covered

Forgive the self-promotion, but ReelDx has the world’s only library of HIPAA-compliant, real patient video cases.  Four of them are free at this url: https://public.reeldx.com/; and another 700 are available if you schedule a demo with yours truly (which you can do here). 

Why Take the Trouble to Code with Claude?

Look, you don’t NEED to lean into the newest tools to be a great teacher.  And Lord knows you’re plenty busy with the basic requirements of the job.

But we’d like to encourage you to be the best healthcare educator you can be, and that means teaching with the latest tools.  We hope to make those accessible - see above - but let’s pause to try to persuade you of the benefits of this approach.

1.  It's Surprisingly Enjoyable (for You and Students)

Creating these simulations isn’t just about student engagement — it makes teaching more engaging too. There’s a reason educators keep coming back to this approach. When students are actively participating, asking better questions, and actually thinking through a case, it turns every session into an engaging exchange of ideas, rather than a delivery of information.

2.  It Transforms Passive Learning into Active Engagement

Active learning has long been recognized for its advantages, but let’s be honest — it takes more effort from both educators and students. That’s why there’s often resistance to implementing it. The good news? AI is changing that.

With the right tools, active learning no longer means adding countless hours to your duties. Instead, it automates engagement — guiding students through realistic clinical interactions while allowing educators to focus on those high-impact teaching moments.

3.  It Inspires Creative Teaching Innovation

Many educators find that once they create their first interactive case, they discover applications across their curriculum. The initial exploration often leads to a cascade of "What if we tried this?" moments that revitalize teaching modalities.

4.  Help Your Students Build their AI Literacy

How do I stay at the top of my AI game?  I have a class full of ReelDx interns who help build these assignments with me.

I asked all of my students to sign up for Claude and build a healthcare education game.  Many got stuck on the character limit, but one student in particular went above and beyond, designing a simulation of an immigrant trying to navigate life in the United States, which you can try here:

The coding isn’t perfect (the scenario will time out after a certain number of choices), but you can clearly see what the future of “vibecoding” is: the ability for anyone to build an interactive experience, regardless of their programming chops.

What are the downsides of using AI?

Claude 3.7 Sonnet represents a significant advancement in reducing common AI limitations. Its contextual understanding and reduced tendency toward "hallucination" (generating incorrect information) make it particularly well-suited for educational applications.  But it will still hallucinate, and you should both fully test whatever it produces AND educate your students on taking some of its findings with a grain of salt.

Old dogs can learn new tricks

Look, we're not going to pretend that this one tool will magically transform your entire teaching career. Your lecture slides aren't going to suddenly start presenting themselves, which of course, is good.  We still need dedicated teachers like you!  AI isn’t coming for your teaching job, though a teacher who can use AI well might be.

The era of one-dimensional patient scenarios is fading, not because they weren't valuable, but because we finally have the tools to make them work harder for us without working harder ourselves. Your clinical knowledge and teaching experience remain the essential ingredients – Claude just helps serve them up in a more engaging package.

And if anyone asks how long it took you to develop such an impressive interactive simulation, well... that can be our little secret.