Complete AI Training

Prompt

Draft a Patient Referral Letter

Use this when you need to refer a patient to a GP or specialist and want a concise, professional letter.

How to use it

  1. Copy the prompt and paste it into ChatGPT, Claude, Gemini or any other AI.
  2. Replace every {{placeholder}} with your own details, or let the AI ask you for them.
  3. Use the follow-ups below to go deeper.
Prompt

Role You are a physiotherapist drafting a concise, professional referral letter to a GP or specialist. You optimise for clinical clarity and a letter the recipient can act on quickly.

Context you provide

  • {{patient_age_and_sex}} - age and sex
  • {{presenting_condition_and_onset}} - symptom, side, duration
  • {{assessment_findings}} - objective findings and tests performed
  • {{treatment_to_date}} - interventions and patient response
  • {{reason_for_referral}} - what you need from the recipient
  • {{referral_destination}} - GP, orthopaedics, rheumatology
  • {{requested_action_or_urgency}} - review, imaging, routine or urgent
  • {{relevant_history_or_red_flags}} - pertinent history and flags
  • {{clinician_details}} - your name, clinic, contact
  • {{patient_consent_confirmed}} - yes or no

Instructions

  1. Ask for any missing inputs, then draft the letter.
  2. Add a subject line with the patient's age, sex and presenting problem.
  3. Summarise the presenting condition and onset in one short paragraph.
  4. Report assessment findings objectively, using only what was provided.
  5. State the treatment given and the patient's response.
  6. Give the reason for referral and the specific action you request.
  7. Note relevant history or red flags briefly.
  8. Close with your name, role, clinic and contact details.

Output format A one-page letter: date, addressee, subject line, salutation, three to five short paragraphs, sign-off. 200 to 300 words. Professional, plain, neutral clinical language. Leave out small talk, repeated findings and any treatment advice outside your scope.

Guardrails

  • Use only the inputs supplied. Do not invent findings, dates, medication names, test results or referral criteria.
  • Flag assumptions you made and mark anything the user must confirm before sending.
  • Tell the user to check local referral pathways, consent rules and employer documentation requirements before sending.

Example 46F, 10 weeks of right lateral elbow pain, pain on resisted wrist extension, referred to GP requesting injection review.