Healthcare Referral Tracking Spreadsheets: Use case and AI Confidentiality

An allied health clinic tracks incoming referrals in a spreadsheet so that no patient waits unseen: who referred them, how urgent it is, whether they have been seen and how many days they have waited. Someone then uploads the sheet to ChatGPT or Claude to summarise waiting times or to build a chart for a meeting. Initials look anonymous, but next to an age, a referral type and a referrer they point to one person in a small clinic. This page marks what a routine referral tracking sheet exposes under the Privacy Act, and shows how the same sheet is worked on Ermos inside the practice.

The referral tracking spreadsheet

Kept by the allied health coordinator to see who has been referred and seen, to prioritise urgent referrals and to report waiting times at the monthly team meeting. It is often shared by email, which is how it reaches an AI tool.

fx
ABCDEFGH
1ReceivedPatientReferralReferrerUrgencyStatusSeen onDays waiting
27
311
410
514
67
79
88
96
103
11Total referrals=COUNTA(B2:B10)
12Not yet seen=COUNTIF(F2:F10,"<>Seen")
13Average days waiting=ROUND(AVERAGE(H2:H10),1)
Referrals

An invented spreadsheet as at 11 November 2026. All patients are fictional and shown by initials and age. Days waiting is the number of days from the received date to the date seen, or to 11 November 2026 if not yet seen. Hover or tap any highlighted item for the analysis; the full list is below.

Referral waiting list tracker

An Agent+ that keeps the referral tracking spreadsheet current from the referrals and booking lists the practice supplies. It adds new rows, updates status, and prepares a short waiting time summary for the monthly team meeting. Urgent referrals and entries without a booking are flagged.

What it does

  1. Reads new referrals and the latest booking lists
  2. Updates each row's status in the practice's sheet
  3. Flags urgent referrals and entries without a booking

What stays with your people. The allied health coordinator checks the sheet and sets priorities. Clinicians decide urgency; the agent only organises what is on file.

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
    Initials and ages

    Initials with an age look anonymous, but in a small clinic they point to one person, and personal information includes information about an individual who is reasonably identifiable (Privacy Act s 6(1)). Next to a referral type the row is health information (s 6FA). Privacy Act guide.

  2. 2
    The referral types

    What each person was referred for, from a lumbar problem to a mental health plan and an NDIS functional capacity assessment. The last two reveal a mental health condition and that a person is an NDIS participant. Health information is sensitive information (s 6(1)). Patient data guide.

  3. 3
    The referrers

    Which doctor or service referred each person, and a support coordinator for an NDIS participant. It shows the patient’s care network and who holds the referral.

  4. 4
    Urgency ratings

    A triage judgement against each patient. The clinical weight of urgent or routine is the clinic’s own judgement, and it shows who the clinic put first.

  5. 5
    Dates, status and bookings

    When each person was referred, seen or booked. Information about a health service provided or to be provided is health information (s 6FA(a)(iii)), and the dates show a treatment timeline. Patient data guide.

  6. 6
    The totals and the average

    A count of referrals, a count not yet seen and an average wait. Aggregate figures with no name in them, although they come from rows that have names.

  7. 7
    The days-waiting formula

    A date calculation fixed at 11 November 2026 so the figures do not change when the sheet is opened later. It holds no patient information. The same formula can be shared safely, and the sheet it sits in cannot.

  8. 8
    Other tabs, hidden columns and file properties

    A spreadsheet file generally carries more than the visible tab: other sheets, hidden columns, cell comments and properties such as the author. Uploading the file sends all of it, including what is off screen.

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 healthcare practices. See how this works in practice: Referral letters on Ermos.

Common questions

Do ChatGPT and Claude use the information in a referral tracking spreadsheet?
An uploaded sheet 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 carries patient initials with ages, referrals for a mental health plan and an NDIS functional capacity assessment, and an urgency rating against each person.
Is it safe to upload a referral tracking spreadsheet to ChatGPT or Claude if it only has initials?
No, because initials do not make the rows anonymous. 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. In a small clinic, T.M. aged 41 with an NDIS assessment, or J.W. aged 29 with a mental health plan from Dr Voss, points to one person. Each row is health information, which is sensitive information under 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.
Can I upload a referral tracking spreadsheet to ChatGPT or Claude instead of pasting the figures?
An uploaded file leaves the clinic as a whole file, which can carry other sheets, hidden columns, comments, tracked changes and file properties as well as the part you meant to share. The OAIC's guidance also tells organisations to review carefully the terms and settings that will apply to their use of any AI product. This workbook has a second tab, Archive 2025, behind the Referrals tab, and the file may hold comments and the author's name that never show on screen.
How can an allied health clinic use AI for a referral tracking spreadsheet 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 waiting time summary is produced from the clinic's own referral records and its own report format. The practice manager checks the counts and the average before they go to a meeting.
Are patient initials and an age still personal information?
They can be. Section 6(1) of the Privacy Act defines personal information as information or an opinion about an identified individual, or an individual who is reasonably identifiable. In a small clinic, initials, an age, a referral type and a referrer may be enough for a colleague or neighbour to know who a row is about. Once a health service is attached, it is health information.
Does a waiting-list spreadsheet count as health information?
Yes. Section 6FA of the Privacy Act includes information or an opinion about a health service provided, or to be provided, to an individual and other personal information collected in providing a health service. A row that says a person is waiting for a mental health plan appointment, or an NDIS assessment, is health information, and health information is sensitive information.
Is an average waiting time safe to share outside the practice?
A figure with no names in it carries little privacy risk by itself. The risk is in the sheet it came from. Uploading the sheet to an AI tool sends every row, hidden column and comment as well as the average. The safer step is to calculate the figure inside the practice and share the result.

See Dominion answering from your own documents.

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

Book a demo