Verified against ChatGPT · 2026-08-10
Turn a pile of health research links into a doctor-visit prep sheet, not a diagnosis
Summarizes what someone has already read about a health topic into an organized set of questions and observations to bring to an actual appointment, explicitly refusing to render a verdict on their situation itself.
The prompt
Ready to copy — highlighted parts are example details you can swap.
You are helping someone organize their own health research ahead of an upcoming appointment. They have been reading about a topic and want the material turned into something useful to bring to their clinician — not a conclusion about what's wrong with them. TOPIC THEY'VE BEEN RESEARCHING Persistent afternoon fatigue and whether it could be related to thyroid function. WHAT THEY'VE READ OR BEEN TOLD Read two articles saying fatigue plus cold sensitivity can point to hypothyroidism, one forum thread saying it's often just poor sleep, no lab work done yet. THEIR OWN SYMPTOMS OR SITUATION Fatigue every afternoon for about 3 months, also noticed hair thinning recently, no other new symptoms. UPCOMING APPOINTMENT TYPE First visit with a new primary care physician, mostly a general checkup. WHAT THEY WANT OUT OF THIS Wants to leave with either a lab order or a clear explanation for why one isn't needed yet. RULES Organize what they've read into categories (things multiple sources agreed on, things sources disagreed on, things that seem specific to their situation versus general information) rather than just restating it back in the same order. Do not resolve disagreements between sources yourself and do not tell them which explanation is more likely correct for their case — that is exactly the judgment their clinician is there to make with information you don't have (an exam, test results, full history). Convert the research into a list of specific questions to ask at the appointment, phrased so a clinician could answer them directly, rather than vague questions like "is this serious." Separate anything in their own description that sounds urgent (a symptom that's rapidly worsening, or a described red-flag symptom they mentioned reading about) into its own flagged line recommending they not wait for the scheduled appointment if it applies, rather than filing it in with the general prep notes. WHAT NOT TO DO Do not state as fact any specific statistic, prevalence number, or named study finding that wasn't explicitly given to you in their notes — if they mention a statistic, you can reflect it back as "you noted that X source said..." but never generate a new one, and never state a treatment's effectiveness as settled fact. MANDATORY DISCLAIMER State plainly, before the summary, that this is a personal research organization aid to prepare for a real appointment, not a diagnosis or medical opinion, and that everything here needs review and interpretation by a qualified clinician who has actually examined them. OUTPUT FORMAT 1. The disclaimer. 2. Any flagged urgent item, if present, at the very top. 3. A short table: what sources agreed on / where sources disagreed / what seems specific to their situation. 4. A numbered list of questions to bring to the appointment.
Customize
Optional — swap in your own details for the highlighted parts above.
Why this works
The core risk in this task is a model quietly sliding from organizing research into interpreting it — GPT-5.1 is fluent enough to produce a confident-sounding synthesis that reads like a diagnosis even when it's just pattern-matching across the sources it was given, so the prompt explicitly forbids resolving disagreements or ranking likelihood, forcing the output to stay in the category of "here's what you found and where it conflicts" rather than "here's what's probably going on." Structuring output around agreement/disagreement/personal-relevance rather than a flat restatement gives the person something a clinician can actually use in a short visit, because a clinician's limited appointment time is better spent resolving the disagreement than re-reading a wall of undifferentiated notes. Requiring urgent-sounding symptoms to be pulled into their own flagged line matters because burying a red flag inside general prep notes defeats the purpose — if something in their own description suggests waiting for a scheduled appointment isn't appropriate, that needs to be impossible to miss, not one bullet among many. The rule against inventing statistics or study findings addresses a specific failure mode of consumer health prompts: a model asked to "summarize research" will sometimes fabricate a plausible-sounding number to fill a gap, which is worse than no number at all since it launders a guess as fact. Framing this explicitly as appointment prep, not diagnosis, both in the disclaimer and in the structural rules, keeps the tool doing the thing it can actually do well — organizing what a person already knows — instead of the thing only an examining clinician can do.
What you get back
Note: This organizes your own research for your appointment — it isn't a diagnosis. Sources agreed: fatigue + cold sensitivity are commonly discussed together with thyroid function. Sources disagreed: whether lab work is needed before or after ruling out sleep issues. Specific to you: the hair thinning wasn't mentioned in the general articles you read. Questions to ask: "Given the fatigue, hair thinning, and duration, would you order thyroid labs at this visit or wait?"
Verified against
ChatGPT GPT-5.1 · 2026-08-10
Changelog
- 2026-08-10 — Initial publish, verified against ChatGPT GPT-5.1.
Need this built into your business?
If a prompt isn't enough — what Scult builds, built and maintained for you — that's Scult's day job.
EXPLORE WHAT SCULT BUILDS
