GP Pathology Results Recall Lists: Use case and AI Confidentiality

Every general practice keeps a pathology results recall list: the patients whose abnormal results mean they have to be brought back, how many times the practice has tried to reach them and how long each result has been open. The practice nurse or practice manager exports it on a Friday and uploads it to ChatGPT or Claude to count the overdue recalls, sort them by urgency or draft the reminder letters. Each row is a named patient beside a result they may not have heard yet, and the notes column records who at home must not find out. This page marks each column of an invented recall list, recomputes the totals, sets out what the RACGP Standards ask of a follow-up system, and shows how the same sheet is worked on Ermos inside the practice.

The pathology results recall list

Kept by the practice nurse or practice manager from the clinical software’s recall function and reviewed weekly with the doctors. It is exported to a spreadsheet to chase the oldest recalls, which is how it reaches an AI tool.

fx
ABCDEFGHIJ
1Result datePatientDOBTestResultGPAttemptsLast attemptDays openNote
2
3
4
5
6
7
8
9
10
11Open recallsDays open over 14
Open recalls

An invented recall list as at Friday 13 November 2026. All patients and the doctors’ initials are fictional. Days open is 13 November less the result date. Totals: nine open recalls, 19 contact attempts (2 + 1 + 1 + 1 + 3 + 2 + 3 + 2 + 4) and three recalls open for more than 14 days (16, 18 and 23 days; the recall at exactly 14 days is not counted). Hover or tap any highlighted item for the analysis; the full list is below.

Recall list reviewer

An Agent+ that builds the weekly review sheet from the recall export. It works from the spreadsheet the practice nurse supplies, sorts the oldest outstanding recalls to the top, carries each patient's contact instructions alongside, and drafts a short list of rows with gaps for the doctors to look at.

What it does

  1. Reads the exported recall spreadsheet
  2. Sorts outstanding recalls by how long they have waited
  3. Flags rows with missing contact details or instructions

What stays with your people. The practice nurse prepares the sheet and the doctors decide on every recall. The agent organises rows; it does not decide who is contacted or how.

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

  1. 1
    Patient names

    A name on a recall list says that person has a result the practice is worried about. The names were collected in providing a health service, so the practice holds them as health information under section 6FA of the Privacy Act, and it is covered by the Act whatever its turnover (s 6D(4)(b)). Privacy Act guide.

  2. 2
    Dates of birth

    With the name, enough to identify each patient with certainty. Privacy Act guide.

  3. 3
    The tests and the results

    A raised PSA, a positive bowel screen, HPV 16, a suspicious breast ultrasound, a chlamydia result and a pregnancy hormone that has not fallen. Health information is sensitive information under section 6(1) of the Privacy Act. Several of these patients have not yet been told, so the sheet knows before they do. Patient data guide.

  4. 4
    The notes column

    Free text with the instructions that keep a recall safe: email only, no voicemail, a wife who does not know, a partner who must not be used to pass a message. The RACGP’s guidance on its follow-up criterion (6th edition, CG6) describes discreet recall methods for patients at risk, such as avoiding detailed voicemails and text messages that others may read. Good medical practice 4.4 asks doctors to treat information about patients as confidential. Patient data guide.

  5. 5
    Contact attempts and dates

    How many times the practice has tried each patient and when. The RACGP Standards ask a practice to document each attempt to contact and recall a patient with a clinically significant result in the health record. The count is also the practice’s evidence if a recall is later said to have been missed.

  6. 6
    Result dates and days open

    When each result arrived and how long it has waited. A date arithmetic column with no clinical content, though beside a name it shows how long that person’s result has gone unanswered.

  7. 7
    The doctors’ initials

    Which doctor ordered each test and owns the recall. Staff details, linked here to named patients and to the oldest recalls.

  8. 8
    The totals

    Nine open recalls, 19 attempts and three open for more than 14 days. Aggregates with no name in them.

  9. 9
    The count formula

    A COUNTIF over the days open column. It holds no patient information, so the formula can be shared; the sheet it sits in cannot.

  10. 10
    The other tab

    The workbook carries a second tab of every recall closed during 2026, with the same columns. Uploading the file sends every tab, hidden column and comment, which turns nine patients into a year of abnormal results.

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 EdgeErmos Dominion
Where the document is processedProvider servers, generally offshore on personal plansYour instance in an Australian data centre; answers generated in AustraliaThe unit in your office
Who can read itProvider staff and systems, under policyYour staff, within the boundaries you set. The named Australian provider processes only permitted excerpts, under contract, and keeps nothingYour staff only
Is it keptOn the provider’s termsKept on your instance; the provider is barred from retaining itStays on the unit
Can it train a modelOn consumer tiers, yesNo, by contractOnly your own unit learns from it, on your hardware

Part of Use cases in general practitioner (GP) practices. See how this works in practice: The GP inbox on Ermos.

Common questions

Do ChatGPT and Claude use the information in a pathology results recall list?
Yes, an uploaded or pasted recall list is sent to the AI company's servers and processed there to produce a reply. As at October 2026, whether a conversation on a personal plan (ChatGPT Free, Plus or Pro; Claude Free, Pro or Max) can be used to train the models depends on a setting each user controls. Business plans (ChatGPT Business and Enterprise, Claude Team and Enterprise) and the developer APIs sit under commercial terms, and OpenAI and Anthropic each say they do not train on that content by default (OpenAI, Anthropic). On either product, the thumbs up and thumbs down buttons can send a whole conversation back for training. This sheet names nine patients with their dates of birth beside results such as a raised PSA, HPV 16 and a suspicious breast ultrasound. Its notes column records that one man's wife does not know he was tested.
Is it safe to upload a pathology results recall list to ChatGPT or Claude if we delete the name column?
No. Deleting the Patient column leaves a date of birth, a test result and a free-text note on every row. Removing names rarely makes a document anonymous. Information stays personal information while the person is reasonably identifiable, and the OAIC warns that information can be at risk of re-identification even when it has been de-identified or anonymised. A woman born in April 1973 with a biopsy booked for 17 November who has not told her daughters could be picked out by her family. So could a man whose partner is a patient at the same practice. Health information is sensitive information under section 6(1) of the Privacy Act. The Office of the Australian Information Commissioner (OAIC) recommends in its guidance on commercially available AI products that organisations do not enter personal information, and particularly sensitive information, into publicly available generative AI tools, because of the significant and complex privacy risks.
What else goes to ChatGPT or Claude when a recall list spreadsheet is uploaded?
Uploading sends the whole workbook, which can include other tabs, hidden columns, comments, tracked changes and file properties as well as the rows on screen. This one has a Closed 2026 tab listing every recall closed during 2026, with the same columns. An upload meant to sort nine open recalls would send a year of abnormal results and the notes beside them. The OAIC's guidance also tells organisations to review carefully the terms and settings that will apply to their use of any AI product.
How can a general practice use AI for a pathology results recall list without sending it to ChatGPT or Claude?
Ermos is built so the document stays under your control. On the Dominion AI unit it is drafted inside your office and none of it is sent to an AI provider. On Ermos Edge your documents and index stay on your own instance in an Australian data centre, and only the question and the excerpts the user is permitted to see go to a named Australian provider that is contractually barred from keeping them or training any model on them. The overdue count and the letters are produced from the practice's own recall export and letter template. A nurse or the practice manager checks every day count and total, and the doctor who owns each recall approves the letter.
What do the RACGP Standards say about recalling patients with abnormal results?
Criterion CG6 of the Standards for general practices (6th edition) asks a practice to act on all clinical information it receives in a timely manner, to have follow-up systems for recalling patients who have clinically significant results, and to document each attempt to contact and recall them in the health record. The 5th edition's indicator GP2.2 B says the practice recalls patients who have clinically significant results.
Which edition of the RACGP Standards applies to accreditation now?
The Australian Commission on Safety and Quality in Health Care says the 6th edition supersedes the 5th, and that accreditation under the National General Practice Accreditation Scheme currently uses the 5th edition, with arrangements for the 6th to be announced. It encourages practices to begin implementing the 6th edition. A practice should check the Commission's page for the current position before an assessment.
What is the difference between a recall and a reminder?
The RACGP Standards describe a recall as asking a patient to attend a consultation to discuss a significant clinical matter, where failing to recall may lead to an adverse outcome. A reminder is a prompt for routine or preventive care such as cancer screening or a vaccination. A recall list therefore holds the results a practice is most worried about.

See Dominion answering from your own documents.

A demonstration takes 30 minutes, uses no client data, and comes with no obligation.

Book a demo