GP Medico-Legal Reports: Use case and AI Confidentiality
Medico-legal reports are the slowest paperwork in general practice. A solicitor sends five questions about a patient injured in a car accident, and the treating GP has to turn eight months of consultation notes into a history, a list of injuries, an account of work capacity and a prognosis. The obvious shortcut is to paste the notes and the letter of request into ChatGPT or Claude and ask for a first draft. The notes hold everything the patient has told the practice, including history she has asked the doctor to hold back. This page marks every item in an invented treating doctor’s report and the drafting note beside it, sets out what the Medical Board asks of doctors who sign reports, and shows how the same report is drafted on Ermos inside the practice.
The treating doctor’s medico-legal report
Written by the treating GP when a solicitor or insurer asks for a report on a patient with a compensation claim, usually weeks after the records have been sent. The GP drafts from the record, checks the draft with the patient where history is sensitive, signs it and the practice invoices the requester.
Wombeyan Street Family Practice
8 Wombeyan Street, Eastwood NSW 2122 · (02) 5550 0161
Treating doctor’s report · DRAFT · Friday 13 November 2026
To: Lucinda Nairn, Senior Associate, Maddox Ferrier Lawyers, Level 6, 50 Phillip Street, Sydney NSW 2000
Re: Petra Lindholm, DOB 11 December 1985; motor vehicle accident 14 March 2026; claim NM-0000000 with Norwell Mutual1
1. The request. Your letter of 9 November 2026 asks about her injuries, her treatment, any earlier condition or psychological history relevant to her recovery, her capacity for work and her prognosis. Her signed authority dated 26 October 2026 is on our file.2
2. History. Ms Lindholm told me on 16 March 2026 that she was stopped at traffic lights on Blaxland Road, Ryde, when her car was struck from behind. She was the driver, wore a seatbelt and did not lose consciousness.3
3. Injuries and treatment. Neck strain and a right shoulder injury. Cervical spine X-ray on 17 March showed no fracture. MRI on 22 April showed a partial thickness tear of the right supraspinatus tendon. She has had 24 physiotherapy sessions. Headaches three or four days a week; amitriptyline 10 mg at night since May.4
4. Work. She is a dental assistant. She was off work from 16 March to 17 July 2026 and returned three days a week on 20 July. She cannot hold her right arm raised for long procedures.5
5. Earlier history. She was admitted to a psychiatric unit for nine days in 2021 after a suicide attempt and has been stable on lithium since, with no change of treatment after the accident. She has reported low mood and poor sleep since April 2026.6
6. Prognosis. I expect the shoulder to improve over 6 to 12 months and have asked for an orthopaedic opinion. The headaches are likely to continue for some months.8
7. Limits of this report. I write as her treating general practitioner. I give no opinion on who was at fault or on permanent impairment.9
Dr Theodora Vasilakis, GP, Provider no. 4xxxxxxC. I have seen Ms Lindholm 11 times since the accident. Report fee to be invoiced by Odette Sorensen, Practice Manager.10
Drafting note, kept out of the report. Petra phoned on 30 October and asked that the 2021 admission is not sent unless it is relevant. The solicitor’s third question asks about psychological history, so I cannot answer it and leave the admission out. Call her before I sign. The 2023 sexual health entries are unrelated and stay out.7
An invented draft report that follows the solicitor’s records request for the same patient. The law practice, the insurer, the claim number and the patient are fictional and the provider number is masked. Hover or tap any highlighted item for the analysis; the full list is below.
Medico-legal report drafter
An Agent+ that drafts the treating doctor's report for a compensation claim. It reads the solicitor's questions and the patient's record, sets out the history and treatment under each question, and notes entries that are sensitive or unrelated to the claim so the GP can check them with the patient.
What it does
- Reads the solicitor's questions and the patient's record
- Drafts an answer to each question with record references
- Notes sensitive or unrelated entries for the GP to raise
What stays with your people. The treating GP checks the draft against the record, discusses sensitive history with the patient, forms any opinion and signs it.
An Agent+ is built once and anyone in the practice can run it. It works from your own documents and runs on Ermos Dominion, the AI unit in your office, or on Ermos Edge, your own instance in an Australian data centre.
What leaves the practice
- 1The patient, the accident and the claim number
Identifies the patient and an open compensation claim with a named insurer. A claim number is the key to a whole insurer file. The practice holds these details as health information because it collected them in providing a health service (Privacy Act s 6FA). Privacy Act guide.
Frontier Models (ChatGPT, Claude) Transmitted to the provider, generally offshore on personal plans Ermos Edge Stays on your instance in Australia; leaves only inside a permitted excerpt, per question, to the named provider Ermos Dominion Stays on the unit in your office Use case Medico-legal report drafter. Builds a draft report answering the solicitor's questions in order, each point traced to the entry in the record it came from. See the use case - 2The solicitor’s questions and the authority
What the solicitor asked and the patient’s signed authority. The authority is her consent to disclosure to that solicitor, which APP 6.1 allows. It names one recipient, and an AI provider used to draft the answers is a different one. Privacy Act guide.
Frontier Models (ChatGPT, Claude) A legal claim, the questions asked and the patient’s authority transmitted Ermos Edge Stays on your instance in Australia; leaves only inside a permitted excerpt, per question, to the named provider Ermos Dominion Stays on the unit in your office Use case Medico-legal report drafter. Builds a draft report answering the solicitor's questions in order, each point traced to the entry in the record it came from. See the use case - 3The patient’s account of the accident
What she told the doctor two days after the crash, with the date she said it. In a claim, the first account given to a doctor can become evidence, and the other side may compare it with later versions.
Frontier Models (ChatGPT, Claude) Her first account of the accident transmitted Ermos Edge Stays on your instance in Australia; leaves only inside a permitted excerpt, per question, to the named provider Ermos Dominion Stays on the unit in your office Use case Medico-legal report drafter. Builds a draft report answering the solicitor's questions in order, each point traced to the entry in the record it came from. See the use case - 4Injuries, imaging and treatment
Diagnoses, two imaging results, a medicine and a count of physiotherapy sessions. Health information is sensitive information under section 6(1) of the Privacy Act. Patient data guide.
Frontier Models (ChatGPT, Claude) Sensitive health information transmitted; on consumer tiers it can train future models Ermos Edge Stays on your instance in Australia; leaves only inside a permitted excerpt, per question, to the named provider Ermos Dominion Stays on the unit in your office Use case Medico-legal report drafter. Builds a draft report answering the solicitor's questions in order, each point traced to the entry in the record it came from. See the use case - 5Her job and her time off work
Her occupation, the dates she could not work and what she still cannot do. The loss of earnings part of the claim is built on these lines. Patient data guide.
Frontier Models (ChatGPT, Claude) Her occupation and time off work transmitted Ermos Edge Stays on your instance in Australia; leaves only inside a permitted excerpt, per question, to the named provider Ermos Dominion Stays on the unit in your office Use case Medico-legal report drafter. Builds a draft report answering the solicitor's questions in order, each point traced to the entry in the record it came from. See the use case - 6A psychiatric admission and a suicide attempt
Mental health history from 2021, years before the accident, with a medicine that identifies the condition. Good medical practice 10.9 asks doctors not to deliberately omit relevant information from a report, so the doctor may have to include it. That decision belongs to the doctor and the patient, and it is made before the report leaves the practice. Patient data guide.
Frontier Models (ChatGPT, Claude) A psychiatric admission transmitted before she and the doctor have decided whether it goes in Ermos Edge Stays on your instance; whether it goes in the report is still decided by the doctor and the patient Ermos Dominion Stays on the unit; whether it goes in the report is still decided by the doctor and the patient Use case Medico-legal report drafter. Builds a draft report answering the solicitor's questions in order, each point traced to the entry in the record it came from. See the use case - 7The drafting note
The patient’s instruction, the doctor’s reasoning about the third question and a mention of unrelated sexual health records. It was never meant for the solicitor. Good medical practice 4.4 asks doctors to treat information about patients as confidential. Patient data guide.
Frontier Models (ChatGPT, Claude) The patient’s instruction and the doctor’s reasoning transmitted Ermos Edge Stays on your instance in Australia; leaves only inside a permitted excerpt, per question, to the named provider Ermos Dominion Stays on the unit in your office Use case Medico-legal report drafter. Builds a draft report answering the solicitor's questions in order, each point traced to the entry in the record it came from. See the use case - 8The prognosis
The doctor’s opinion on recovery, which the value of the claim partly rests on. Good medical practice 10.9 asks doctors to take reasonable steps to verify the content of a report before signing it, and a drafted opinion still has to be the doctor’s own. Patient data guide.
Frontier Models (ChatGPT, Claude) A prognosis that bears on the claim transmitted Ermos Edge Stays on your instance; a drafted prognosis is still an opinion the doctor has to check and own Ermos Dominion Stays on the unit; a drafted prognosis is still an opinion the doctor has to check and own Use case Medico-legal report drafter. Builds a draft report answering the solicitor's questions in order, each point traced to the entry in the record it came from. See the use case - 9The limits paragraph
Standard wording that says the doctor writes as the treating GP and gives no opinion on fault or permanent impairment. It contains nothing about the patient.
Frontier Models (ChatGPT, Claude) Standard wording with nothing about the patient Ermos Edge Not about where data goes Ermos Dominion Not about where data goes Use case Medico-legal report drafter. Builds a draft report answering the solicitor's questions in order, each point traced to the entry in the record it came from. See the use case - 10The doctor, the visit count and the fee
A register name, a masked provider number, the number of consultations and who invoices the report. Practice details linked to this claim.
Frontier Models (ChatGPT, Claude) Register names and a masked identifier, tied to the claim Ermos Edge Stays on your instance in Australia; leaves only inside a permitted excerpt, per question, to the named provider Ermos Dominion Stays on the unit in your office Use case Medico-legal report drafter. Builds a draft report answering the solicitor's questions in order, each point traced to the entry in the record it came from. See the use case
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
Ermos Dominion
The document is drafted on a unit in your office, from your own templates and the patient record. Every marked item stays on the unit, which keeps working with the internet switched off. Nothing in it is transmitted to an AI provider, so there is no disclosure to analyse: the confidentiality work is the same as for your file server.
Ermos Edge
Your documents, the index and your access boundaries stay on your own instance in an Australian data centre. When a staff member asks for this draft, only their question and the excerpts they are permitted to see go to a named Australian provider, who is contractually barred from keeping the content or training any model on it. Nothing is processed outside Australia.
| Frontier Models (ChatGPT, Claude) | Ermos Edge | Ermos Dominion | |
|---|---|---|---|
| Where the document is processed | Provider servers, generally offshore on personal plans | Your instance in an Australian data centre; answers generated in Australia | The unit in your office |
| Who can read it | Provider staff and systems, under policy | Your staff, within the boundaries you set. The named Australian provider processes only permitted excerpts, under contract, and keeps nothing | Your staff only |
| Is it kept | On the provider’s terms | Kept on your instance; the provider is barred from retaining it | Stays on the unit |
| Can it train a model | On consumer tiers, yes | No, by contract | Only your own unit learns from it, on your hardware |
Part of Use cases in general practitioner (GP) practices. See how this works in practice: Certificates and reports on Ermos.
Common questions
Do ChatGPT and Claude use the information in a medico-legal report?
Is it safe to paste a medico-legal report into ChatGPT or Claude if we remove the patient's name?
Can ChatGPT or Claude be trusted to draft a medico-legal report?
How can a GP practice use AI for a medico-legal report without sending it to ChatGPT or Claude?
What does the Medical Board expect of a GP who signs a medico-legal report?
Can a GP leave a patient's mental health history out of a report for a solicitor?
Is a treating doctor's report different from an independent medical examination?
See Dominion answering from your own documents.
A demonstration takes 30 minutes, uses no client data, and comes with no obligation.
Book a demo