Tutorial: Preparing a Teaching Case with AI

Preparing a teaching case takes time that most clinicians do not have. A real case study for trainees, a presentation for a multidisciplinary team meeting, or an anonymised example for a nursing education session — all of these involve the same challenge: turning detailed clinical information into a clear, structured educational document.

This tutorial shows you how AI can help with that task. We will work through an example from start to finish.

---

What You Will Produce

By the end of this tutorial, you will have a structured teaching case document covering:

• A brief clinical scenario
• Key learning points
• Questions for discussion

This is the kind of document that works for a 20-minute teaching session with registrars or a team education morning.

---

Step 1: Choose and Anonymise Your Case

Pick a case that taught you something — a diagnostic puzzle, an unexpectedly good outcome, a difficult communication challenge, or a situation where the standard approach was not right.

Before you do anything with AI, anonymise the case completely. Change or remove:

• The patient's name, age (use an age range), date of birth
• The clinic or hospital location
• Any names of colleagues or staff
• Any dates that could identify the case
• Any unusual circumstances that would identify the patient

Write out the anonymised case in plain terms. At this point, you are just describing a clinical scenario — not a real person.

---

Step 2: Decide on Your Teaching Goals

What do you want trainees or colleagues to take away? Write two or three learning points before you open the AI tool. For example:

• Understanding the decision to convert from fresh to frozen transfer in the context of OHSS (ovarian hyperstimulation syndrome — a condition where the ovaries overrespond to stimulation) risk
• How to communicate a cycle plan change to an anxious patient
• Recognising early signs of excessive stimulation response on monitoring

This step is important. If you skip it, the AI will guess at your goals — and it will guess generically.

---

Step 3: Open Your AI Tool

Go to claude.ai or chat.openai.com. Start a new conversation.

---

Step 4: Write Your Prompt

Type something like:

"Help me structure an anonymised teaching case for a group of IVF trainees. Here is the scenario: [paste your anonymised scenario]. The learning goals are: [paste your 2–3 goals]. Please structure the output as: (1) a brief scenario summary of around 150 words, (2) three to five key learning points in plain language, (3) four discussion questions that a facilitator could use with a small group. Tone: educational, clear, engaging. Aimed at doctors in the first two years of subspecialty training."

---

Step 5: Read the Output

The AI will produce a structured document. Here is an example of what it might look like for a scenario involving a decision to freeze all embryos due to OHSS risk:

---

Scenario Summary

A 31-year-old woman with polycystic ovary morphology attended for monitoring on day eight of stimulation. Her response was stronger than anticipated, with a large number of follicles developing across both ovaries. Her oestradiol level was significantly elevated. After discussion with the patient, the clinical team decided to trigger and freeze all resulting embryos rather than proceed with a fresh transfer, in order to reduce the risk of severe OHSS. The patient had previously been expecting a fresh transfer and was disappointed by this change.

Key Learning Points

1. Cycle modification — including a freeze-all strategy — may be the safest option when stimulation response is excessive, even when the patient is disappointed by the change.
2. Communicating a last-minute change in plan requires clear explanation of the clinical reasoning and acknowledgement of the patient's emotional response.
3. OHSS is preventable in the majority of cases with careful monitoring and willingness to modify the plan.
4. Freezing all embryos does not reduce the overall chance of pregnancy — it delays it.

Discussion Questions

1. At what point in stimulation would you begin to consider a freeze-all strategy? What factors would you weigh?
2. How would you explain this change of plan to a patient who had been expecting a fresh transfer?
3. What are the risks of proceeding with a fresh transfer in this situation?
4. How do you balance patient preference with clinical risk in this scenario?

---

That is a solid structure. Read it carefully.

---

Step 6: Revise for Accuracy and Your Context

Check every clinical statement. AI can get details wrong — for example, it might describe a threshold or a clinical parameter inaccurately. You know what the case actually involved.

Also check:

• Are the learning points what you wanted, or has AI guessed differently?
• Are the discussion questions at the right level for your audience?
• Is the tone right for your team?

Ask for revisions in the same conversation. For example:

"The third discussion question is too broad. Make it more specific — focus on the conversation with the patient, not the clinical decision."

---

Step 7: Add What AI Cannot Know

Copy the final draft into your own document. Now add:

• Any visual aids or scan images you want to include (anonymised)
• References to guidelines you want the group to read
• Specific points from your clinic's protocol that are relevant
• Your name as the session facilitator

---

Step 8: Prepare Your Facilitation Notes

AI can also help with this. In the same conversation:

"For the discussion question about communicating a change of plan to an anxious patient — give me a brief facilitator note with two or three key points I might want to draw out from the group, and a common wrong answer to watch for."

Use the output as your preparation notes, not as a script.

---

Step 9: Check the Final Document

Read the complete document from start to finish. Make sure it:

• Contains no identifying information
• Is clinically accurate
• Reads clearly to someone who was not in the original clinical situation
• Serves your stated teaching goals

---

Step 10: Use It and Iterate

Run the session. Note what questions prompted the best discussion and what fell flat. The next teaching case you prepare with AI will be better because you will know what to ask for.

---

Privacy reminder: Even for teaching cases, full anonymisation is essential. AI makes this more efficient, but the responsibility for removing identifying information rests with you before you ever open the AI tool.

Remember: AI is a helpful assistant, not a clinician. You make the call.