GP care plans and AI patient privacy: what a chronic condition care plan gives away in ChatGPT or Claude.

A chronic condition care plan is a complete summary of a person’s health in two pages: every condition, every medication, the latest results, what the patient wants to achieve and who is helping them. It is long, repetitive to write and reviewed every few months, which is why GPs and practice nurses paste last year’s plan into ChatGPT or Claude to update it. This page marks every item in a routine care plan that the Privacy Act, Good medical practice and the NSW health records law bear on, and shows how the same plan is drafted on Ermos inside the practice.

The chronic condition care plan

Prepared by the GP, often with the practice nurse, for a patient with one or more chronic conditions, then reviewed at set intervals. Copies go to the patient and to the allied health practitioners named in the plan.

An invented care plan. The Medicare number is a published test value and the mobile and provider numbers are masked. Hover or tap any highlighted item for the analysis; the full list is below.

What leaves the practice

1. 1
   Name, date of birth, address and Medicare number

   The identifiers that tie the whole plan to one man. The Medicare number is an identifier assigned by a government agency, and APP 9.2 limits when an organisation may use or disclose one. Privacy Act guide.

2. 2
   Every chronic condition and a recent fall

   A list of diagnoses is health information, which the Privacy Act treats as sensitive information with stricter rules on use and disclosure. Patient data guide.

3. 3
   The medication list

   Doses that show each condition being treated and a dose reduced for kidney function. Part of the record Good medical practice 10.5 asks a doctor to keep accurate and secure. Patient data guide.

4. 4
   Pathology results

   Laboratory values against a named patient, with the laboratory and date. Patient data guide.

5. 5
   Goals and a fear of residential care

   What the patient said about his life and what frightens him. Private detail Good medical practice 4.4 asks doctors and their staff to hold in confidence. Patient data guide.

6. 6
   The actions, including a driving question

   The GP’s plan for this patient, and a note that his fitness to drive is under discussion. A third party reading it learns something the patient has not yet decided to share.

7. 7
   The allied health team

   Named practitioners outside the practice. Their names are public, but the plan links them to this patient’s care.

8. 8
   A named carer with her mobile number

   Personal information about someone who is not the patient, collected because she helps with his medications. Privacy Act guide.

9. 9
   The consent line

   Records that the patient agreed to the plan going to the named care team. It covers those recipients; an AI provider is a different recipient the patient was never asked about. Patient data guide.

10. 10
    The GP and practice nurse

    Names on the public register and a masked billing identifier, linked here to this patient’s record.

What happens in Frontier Models (ChatGPT, Claude)

• The whole document is transmitted to the provider’s servers. On personal plans the user cannot choose where it is stored or processed, and for the major providers that has generally meant outside Australia, under that country’s legal process. Some business plans offer storage in Australia, but the provider still receives the content.
• On consumer tiers, prompts can be retained and used to train future models.
• The provider’s staff and systems can reach the content under the provider’s own policies.
• Once sent, the practice no longer controls it. Retention and deletion happen on the provider’s terms.

The same document on Ermos

Part of AI data leakage in general practitioner (GP) practices. See how this works in practice: Care plans on Ermos.

Common questions