How to use ChatGPT Health and prepare for a visit
Connect supported sources in ChatGPT Health, review what it can use, and prepare clear questions for your clinician. Keep personal health information separate from reusable preparation tips.
Updated 5 min read
To use ChatGPT Health, open Health in the sidebar, connect a supported source, and review its permissions before asking for help preparing for a visit. OpenAI describes Health (opens in a new tab) as a way to understand connected information and organize questions for your doctor. Keep reusable preparation methods in fenbs, with personal medical details kept separate.
Who can use Health, and what is it for?
OpenAI’s public Health page (opens in a new tab) lists eligible US users aged 18 and older on web and iOS. Access is rolling out. This is consumer visit preparation, separate from the professional workflows discussed in ChatGPT for clinicians.
Start with a narrow goal: understand a term, organize questions, or identify information you want your clinician to explain. Treat the output as a draft to check. A clear list of questions is a more useful outcome than a confident answer you cannot verify.
Health is not for diagnosis or treatment, according to OpenAI’s help guide (opens in a new tab). Seek professional help for urgent concerns rather than waiting for an AI conversation.
Where do you connect your health information?
Use the documented Health paths (opens in a new tab). Apple Health requires an iPhone:
- Open Health in the sidebar, under More if needed.
- Choose Connect Health or Get started, then a supported source and its permission steps.
- Add a provider through Health > Accounts > Add account.
- On iPhone, manage Apple Health through Health > more options (•••) > Apple Health.
- Review permissions in Settings > Plugins > Health. Use @Health in a message to indicate that connected data should inform the response.
Connect only what serves the question you want to ask. Before agreeing, read which account and information you are authorizing. Avoid treating every available connection as something you need to enable. Make a deliberate choice, then review the result before adding another source.
Keep account selection deliberate, too. If you use more than one login, the guide to switching ChatGPT accounts covers that separate task. Stop before granting access if you are unsure which account you are using.
What should you ask before an appointment?
Work toward a short question sheet you can review yourself. Ask for ordinary language, uncertainty, and a distinction between information in the source and the assistant’s interpretation. Do not ask it to decide which treatment you should follow.
- Define the goal: “Help me prepare questions for a routine visit. Do not diagnose or recommend treatment.”
- Ask for traceability: “For each point, identify the connected record it comes from and anything that needs checking.”
- Request clarification: “Which terms should I ask my clinician to explain in plain English?”
- Keep the result short: “Give me five questions, ordered by what I should clarify first.”
Review every line against the original information before taking it to the appointment. Separate “the record says this” from “I want to ask about this.” If an answer seems inconsistent, turn the inconsistency into a question for your clinician instead of asking the assistant to settle it.
For example, a generic preparation method might be: gather the source, check its date, define unfamiliar terms, and write questions for a professional. That method contains no patient details. It is the kind of repeatable approach you can keep in fenbs for later use; the personal question sheet should stay separate.
Which privacy choices should you review?
OpenAI says connected medical and Apple Health information, and chats using it, are excluded from foundation-model training and ad targeting (opens in a new tab). That specific protection does not mean every conversation has the same controls. Review what you are sharing before proceeding.
The Health guide (opens in a new tab) says relevant connected data can be used across chats with permission; Health conversations can create memories when memory is on. Disconnecting deletes synced source data within 30 days, but existing chat content remains until deleted. Earlier chats stay in the separate Health Project, not automatically in the new tab.
Before you connect anything, decide whether the benefit is worth sharing that information. Before you disconnect, review both the source connection and what you have already put into conversations. Keep those as separate checks, and do not assume one action has completed the other.
Public eligibility wording and the help guide’s discussion of linked-teen restrictions are not identical. Do not infer teen eligibility from the existence of a parent-managed switch. The separate parental-controls guide explains account linking without treating it as permission to use Health.
What if Health is missing or a connection is unavailable?
Do not solve missing access by sharing more information in another tool. First compare your account and device with the Health access guidance linked above. Do not assume that a missing tab can be enabled by granting another connection more access.
Write down the nonmedical part of the problem: what you tried, which screen you reached, and whether you saw a connection choice. Keep screenshots and error descriptions free of personal medical information. You can still prepare a blank visit-question template while you investigate access.
Avoid repeatedly authorizing extra accounts just to test whether the feature works. Return to your original purpose and ask whether you need a connection at all to draft generic questions. If access remains unclear, use the vendor guidance rather than a guessed menu path.
What should you keep after the session?
Keep your personal appointment notes in the place you already trust for health information. Outside approved health systems, retain only generic methods or lessons from public sources. Never save patient or confidential details. Removing a name from a personal medical summary does not turn it into a reusable method.
A suitable reusable note might say: “Ask for source dates, separate facts from interpretations, and finish with questions for a clinician.” Do not add the results, medications, symptoms, appointment dates, or excerpts that led you to that lesson. Save the approach in fenbs so you can use it again without copying a health conversation.
Never lose the work you did in an AI session
After a useful AI session, save it in fenbs: the key observations, findings, research and context. Then carry on from exactly where you left off, on any device and in any other AI. The next AI you connect reads it at the start of the session, so you never have to explain the same problem again.
Create a free accountFree for up to 5 people and 50 open tasks · No credit card · Works with Claude, ChatGPT, Cursor and more
How to keep a preparation method with fenbs
Keep the useful method so you can carry on from where you left off, on any device and in another connected AI. fenbs is not the right place for a personal health record or clinical work. For other everyday uses, browse AI at work and the separate ChatGPT and Apple Intelligence comparison.
- Create a free account.
- Write down only the generic preparation steps or a lesson from a public source. Never save patient or confidential details.
- Connect ChatGPT to fenbs, if your account allows the connection.
- Ask it to save that generic method and the next nonmedical step. Review the words before saving.
- When you return, ask your connected AI to read the saved method and help you use that approach again, without adding personal medical information.