GP Notifiable Conditions Logs: Use case and AI Confidentiality

A GP practice keeps a running notifiable conditions log of the infectious diseases it has reported to the local Public Health Unit: who was seen, what was suspected, when it was reported and whether the report went in on time. The practice manager or a nurse uploads the sheet to ChatGPT or Claude to count the month’s notifications, find the late ones or write a summary for the clinical meeting. Every row is a named patient with an infectious disease, and the notes column adds a workplace, a school or a partner. This page marks each column of an invented log, recomputes the totals, and shows how the same sheet is worked on Ermos inside the practice.

The notifiable conditions log

Kept by the practice nurse or practice manager beside the clinical software, updated each time a doctor notifies a condition to the Public Health Unit, and reviewed at the clinical meeting. It is often exported or emailed, which is how it reaches an AI tool.

fx
ABCDEFGHI
1SuspectedPatientDOBConditionPhone conditionNotifiedWithin 24 hByNote
2
3
4
5
6
7
8
9
10
11Total
Notifications

An invented log as at 11 November 2026. All patients, the centre and the doctors’ initials are fictional. “Within 24 h” is the practice’s own target: notified by the end of the day after the condition was suspected. Totals: nine notifications, two of them for conditions NSW Health asks doctors to notify by telephone (measles and hepatitis A), and six of nine within 24 hours. Hover or tap any highlighted item for the analysis; the full list is below.

Notifiable conditions log reconciler

An Agent+ that keeps the practice's notifiable conditions log in order. It compares each row with the notification entries supplied from the record, updates the spreadsheet with new notifications, and prepares a summary for the clinical meeting using case references that lists gaps and late entries.

What it does

  1. Reads the log and the notification entries supplied
  2. Updates rows and fills gaps from the source entries
  3. Prepares a clinical meeting summary using case references

What stays with your people. The practice nurse or practice manager confirms every entry, and the doctors decide what is notified. The agent reconciles; it does not notify.

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 list of notified infectious diseases is already a diagnosis. The names were collected in providing a health service, so they are health information under section 6FA of the Privacy Act, and a practice that provides a health service and holds health information is not a small business operator (s 6D(4)(b)). Privacy Act guide.

  2. 2
    Dates of birth

    With the name, enough to identify each patient with certainty. NSW Health’s notification form asks for a date of birth, so the practice has it; the form asks for only the first two letters of the names for HIV, which shows how little identifying detail is asked for there.

  3. 3
    The notifiable conditions

    Whooping cough, measles, hepatitis A, syphilis and tuberculosis against named people. Health information is sensitive information under section 6(1) of the Privacy Act. NSW Health says doctors notify these conditions on reasonable clinical suspicion and that the notifications are legally required and kept strictly confidential. A paste into an AI tool is a different disclosure from the one the law requires. Patient data guide.

  4. 4
    The notes column

    Free text with the most candid detail: a pregnancy, an unvaccinated child, a patient who does not want his wife told and partners abroad, a cook at a named children’s centre, and a visa worry. It names workplaces and people who are not the patient in that row. Good medical practice 4.4 asks doctors to treat information about patients as confidential. Patient data guide.

  5. 5
    Dates and the late notifications

    When each condition was suspected and notified, and the three that missed the practice’s own 24 hour target, counted from the date each condition was suspected. NSW Health says a notification should be initiated within 24 hours of diagnosis. The sheet is the practice’s own record of where it fell short, written by name.

  6. 6
    The phone condition column

    Which conditions NSW Health asks doctors to notify by telephone, taken from its notification form. A public rule from the form, though on each row it sits beside a named patient and so hints at the condition.

  7. 7
    The doctors’ initials

    Which doctor notified each case. Staff details, linked here to named patients and to the pattern of late notifications.

  8. 8
    The totals

    A count of notifications, a count of phone conditions and a count notified within 24 hours. Aggregates with no name in them, though they come from rows that have names.

  9. 9
    The count formula

    A COUNTIF over a Yes and No 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 listing contacts the Public Health Unit asked about: parents, partners, an employer and classmates, none of whom is a patient. Uploading the file sends every tab, hidden column and comment.

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.

Common questions

Do ChatGPT and Claude use the information in a notifiable conditions log?
Everything pasted or uploaded 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 log lists nine patients by name and date of birth with whooping cough, measles, hepatitis A, syphilis and tuberculosis. Its notes column names a children's centre, a partner and a visa worry.
Is it safe to paste a notifiable conditions log into ChatGPT or Claude if we remove the names?
No, because each row still points to one person. 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 cook at a named children's centre with hepatitis A, or an infant admitted overnight after a seizure, would be recognised by colleagues and family. The diagnoses are health information under section 6FA 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 travels with a notifiable conditions log when it is uploaded to ChatGPT or Claude?
The whole workbook leaves the practice, including every tab and not just the sheet on screen. This one has a second tab listing contacts the Public Health Unit asked about, including parents, partners, an employer and classmates, none of whom is a patient. An uploaded file can also carry hidden columns, comments, tracked changes and file properties. 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 GP practice use AI for a notifiable conditions log 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 monthly counts and the late notifications are worked out from the practice's own log. A person at the practice checks every date and total before a summary reaches the clinical meeting.
Is a notifiable conditions log health information under the Privacy Act?
Yes. Section 6FA covers information about a person's health, including an illness, and other personal information collected in providing a health service. Health information is sensitive information under section 6(1). A practice that provides a health service and holds health information is not a small business operator, so the $3 million turnover exemption does not apply to it.
Does NSW law require GPs to notify infectious diseases?
NSW Health says doctors must notify certain conditions on reasonable clinical suspicion under the Public Health Act 2010, that the notifications are kept strictly confidential in line with privacy legislation, and that notification should be initiated within 24 hours of diagnosis. Some conditions, such as measles and acute viral hepatitis, are notified by telephone. The duty is owed to the Public Health Unit, and an AI provider is a different recipient.
What does the NSW Health notification form ask for?
The July 2026 doctor and hospital form asks for the patient's name, address, date of birth, language, country of birth, occupation or school, date of onset, Aboriginal or Torres Strait Islander status and risk factors, plus the doctor's details. For HIV it asks only for the first two letters of each name and no address, occupation or school. A practice log holding more than that is a choice the practice made.

See Dominion answering from your own documents.

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

Book a demo