United Kingdom version. Set to the United Kingdom: prompts follow NICE or SIGN, BNF names, mmol/L, NHS services and GMC confidentiality, and the privacy check masks NHS numbers and UK postcodes. US version

AI Prompts for Doctors (UK)

Clinical prompt templates for UK doctors: GP and hospital letters, IFRs, results in mmol/L and patient leaflets, with NHS number masking before you paste.

The interactive tool loads here when JavaScript runs.

About this tool

What is AI Prompts for Doctors?

This version starts with the region set to the United Kingdom. Every prompt it builds tells the assistant to write British English, to use mmol/L for glucose and cholesterol and mmol/mol for HbA1c, to point patients to their GP, NHS 111, A&E or 999, and to leave a [CHECK NICE/SIGN GUIDANCE] marker instead of quoting a recommendation from memory. Those lines sit under Constraints, so you can see exactly what was added.

The paperwork is the NHS kind. A request for a drug outside the local formulary goes to the integrated care board as an Individual Funding Request, which asks whether the patient is clinically exceptional rather than whether a payer's criteria are met, so the funding-letter template frames it that way. Referral letters, discharge advice, results letters and clinic summaries keep the structure UK consultants and GPs expect.

Before you copy, the privacy check looks for the identifiers that appear in UK notes: names after a title or "Patient:", NHS numbers (with or without the label, checked against the modulus 11 check digit), dates of birth, UK postcodes and phone numbers. Masking them is a first step, not permission to paste patient records into a consumer chatbot.

  • Ten templates covering notes, letters, payer appeals, evidence searches and patient education
  • Structured prompts with role, context, task, constraints and output format
  • PHI scan and one-click masking of names after titles, MRNs, dates, phone numbers, emails and ages over 89
  • Reading-level and language options for patient-facing text
  • Favourites, search and a Markdown export of the whole kit for your team
  • Runs locally in the browser with no account and no upload
Read the guide: AI for Doctors and Nurses: Safe Prompts for Notes, Letters and Teaching

Updated

How to use AI Prompts for Doctors

  1. Type your specialty and setting, for example outpatient cardiology, and choose your region so the prompt uses local spelling and conventions.
  2. Search or browse the templates and pick one, such as SOAP note from dictation or Prior authorization or appeal letter. Star the ones you use daily.
  3. Fill in the fields with de-identified notes, or press Try an example to see a finished prompt built from fictional data.
  4. Run the privacy check, mask any identifiers it finds, then copy the prompt and paste it into your approved assistant.
  5. Review the draft against the check-the-output list before anything goes into the chart or to a patient.

Worked examples

Masking a UK clinic note

Input

Patient: Margaret Hale, NHS number 943 476 5919, DOB 14/02/1951, lives at LS6 2QT. Seen in clinic, HbA1c 64 mmol/mol.

Output

Patient: [NAME-1], NHS number [RECORD-ID-1], DOB [DATE-1], lives at [POSTCODE-1]. Seen in clinic, HbA1c 64 mmol/mol.

Four identifiers are found: the name, the NHS number, the date of birth and the postcode. The HbA1c value is left alone because clinical values are not identifiers. A full postcode can still point to a handful of homes, which is why it is masked too.

Individual Funding Request instead of a prior authorisation

Input

Prior authorisation or appeal letter template, region United Kingdom, a biologic for psoriasis after methotrexate was stopped for raised ALT

Output

Prompt with the template's rules plus four UK rules and a note to frame the letter for the ICB panel and its exceptionality test

An IFR panel asks whether this patient differs from others with the same condition and would gain more benefit. A letter that only lists failed treatments, as a US appeal would, misses that question. Keep to facts in the record and let the model mark gaps as [MISSING].

Background and insight

Large language models write fluent clinical prose and also produce confident errors: a plausible lab value that was never measured, a guideline that does not say what the model claims, a dose with the wrong unit. The templates are built to make those errors visible. Each prompt asks for a list of gaps or unverified claims next to the draft, separates the physician's assessment from anything the model suggests, and forbids citations that are not in your material.

Privacy rules shape what you can paste. Under HIPAA a covered entity may share protected health information with a vendor only under a business associate agreement (45 CFR 164.502(e) and 164.504(e)), and HHS guidance from 2016 treats a cloud service that stores ePHI as a business associate even if it cannot read the data. Most consumer chat apps do not sign one. The Safe Harbor method in 45 CFR 164.514(b)(2) lists 18 identifiers to remove; the alternative is an expert determination. In the UK and EU, health data is special category data under GDPR Article 9.

The masking helper is pattern based. It catches common formats such as "Mr. Alvarez", "MRN: 4471203", 03/14/1957, phone numbers, emails and ages over 89, which are Safe Harbor identifiers, but it cannot recognise a name written without a title or a rare diagnosis that identifies someone in a small town. Read the masked text before you copy it.

Common use cases

  • Clinic documentation

    Turn dictated bullet points into a SOAP note in your preferred style, with gaps such as missing allergies flagged for you to fill.

  • Payer paperwork

    Draft prior authorization requests, appeals and peer-to-peer talking points that quote each policy criterion and point to the evidence.

  • Patient communication

    Explain results, reply to portal messages and write condition handouts at a reading level patients can act on.

  • Evidence questions

    Frame a bedside question as PICO with MeSH terms and a Boolean search string, then read the studies yourself.

Tips and limitations

  • If your trust or practice has an approved AI tool covered by a data protection impact assessment, use it for anything that cannot be fully de-identified.
  • Check every drug name and dose in a draft against the BNF and the record, not against what the model wrote.
  • Use the [CHECK NICE/SIGN GUIDANCE] markers as a to-do list: open the guideline and confirm the recommendation and its date before the letter goes out.

Sources and further reading

  1. Guidance on the use of AI-enabled ambient scribing products in health and care settings, NHS England. Information governance, DPIA and approval expectations for AI tools that handle patient data.
  2. Confidentiality: good practice in handling patient information, General Medical Council
  3. NHS Number (NHS Data Model and Dictionary), NHS England. The 10-digit NHS number and its modulus 11 check digit.
  4. Data Protection Act 2018, legislation.gov.uk

Frequently asked questions

Can NHS doctors use ChatGPT for clinic letters?

Only within your organisation's rules. Patient information is special category data under the UK GDPR and is also covered by the common law duty of confidentiality. NHS England guidance on AI scribing products expects a data protection impact assessment and an approved product before patient data goes into an AI tool. Without one, remove every identifier and paste only the excerpt you need.

Does masking the NHS number make a note anonymous?

No. It removes one direct identifier. Rare conditions, dates, places and family details can still identify someone, and pseudonymised data is still personal data under the UK GDPR. Read the masked text before you copy it.

How does the tool recognise an NHS number without the label?

A 10-digit number written as 3, 3 and 4 digits is flagged only when its last digit passes the NHS modulus 11 check, so most phone numbers and reference numbers are not mistaken for one. A number with the label "NHS number" is always masked.

Will the prompts make the assistant quote NICE guidance?

No. The UK rules tell it to name NICE or SIGN guidance only when it is in your material and otherwise to write [CHECK NICE/SIGN GUIDANCE: topic]. Language models misremember recommendations and their dates, so you look them up.

What does the GMC expect when doctors use AI to draft?

Good medical practice and the GMC confidentiality guidance apply to AI-assisted drafting as to anything else: you remain responsible for what you sign and send, and you share only the information the purpose needs. Your employer's policy on approved tools applies on top.

Can I use these prompts for a private medical insurer pre-authorisation?

Yes. The funding-letter template accepts the insurer's criteria as pasted text, and the UK note tells the model to use whichever process is named in the material: an ICB Individual Funding Request or an insurer's pre-authorisation.

Is there an in-depth guide to AI Prompts for Doctors?

Yes. “AI for Doctors and Nurses: Safe Prompts for Notes, Letters and Teaching” covers the background, how it works, common pitfalls and worked examples. Read the full guide.