Drafting Referral Letters to Primary Care and Allied Health
Use AI to draft structured, clear referral and correspondence letters to GPs, dietitians, and other specialists — faster and more consistently.
Drafting Referral Letters to Primary Care and Allied Health
The Problem
A substantial portion of your administrative time goes into correspondence. Letters back to GPs after clinic. Referrals to dietitians for patients with obesity or newly diagnosed diabetes. Requests to ophthalmology for diabetic eye screening follow-up. Updates to other specialists when a pituitary or adrenal finding changes the picture.
These letters need to be professional, clear, and complete. They also need to be written in slightly different registers: a letter to a GP needs to translate your specialist thinking into practical action points; a letter to a dietitian needs to explain the metabolic context clearly enough for a colleague who is not an endocrinologist.
Writing each one from scratch is slow. Using the same generic template for every letter produces letters that fail to communicate the specific clinical nuance.
How AI Helps
AI is well suited to drafting structured letters. You give it the key clinical points you want to communicate — in plain notes or bullet form — and tell it who the letter is going to and what action you need. It produces a professional first draft. You review, adjust, and send.
You are not dictating clinical decisions to the AI. You are giving it the substance and asking it to arrange that substance into a coherent letter. The clinical content is entirely yours.
A Real Example
Dr Obi has just reviewed "Mr K", a 47-year-old with poorly controlled type 1 diabetes (insulin-dependent diabetes) who would benefit from structured dietitian input focused on carbohydrate counting. She wants to write a letter to the community dietitian service.
She types the following into an AI tool, using no patient identifiers:
"Write a referral letter from an endocrinologist to a community dietitian service. The patient is a 47-year-old man with long-standing type 1 diabetes on an insulin pump. The aim is structured input on carbohydrate counting to support better glucose management. He has no known eating disorder. No specific blood results or drug details — just make clear this is for specialist dietary support in the context of insulin therapy. Professional tone, about 200 words."
The AI produces a clean letter draft. Dr Obi adds the correct clinical detail about the patient's current management approach (without identifiers — she will add those in her clinical system), adjusts one phrase, and has a usable draft in under two minutes.
Try It Yourself
Write a referral letter from an endocrinologist to {{referral_recipient}} for a {{patient_age}}-year-old patient with {{main_condition}}. The purpose of the referral is {{referral_purpose}}.
Key points to communicate:
- {{clinical_point_1}}
- {{clinical_point_2}}
- {{clinical_point_3}}
Do not include specific drug names, doses, or blood results. Professional tone. Around 200 words. End with a clear request for action.
Do not include real patient names, dates of birth, NHS numbers, GP names, or any other identifying information in your prompt.
Things to Watch For
The AI does not know the recipient. It will produce a professional letter, but it does not know whether your local dietitian service has a specific referral format or whether your GP colleagues prefer a particular structure. Edit for your local context.
Clinical accuracy is your responsibility. The AI is writing from your bullet points. If you leave something out, it will not know to include it. If you include something inaccurate in your notes, the letter will reflect that.
Do not paste letter templates containing patient details. If you have an old letter to use as a structural guide, remove all patient identifiers before pasting it.
Letters go into patient records. What you send is a clinical document. Review the full draft before sending, as you would any letter you had dictated.
Remember: AI is a helpful assistant, not a clinician. You make the call.
Was this lesson helpful?
Related lessons
AI for Endocrinologists: A Practical Introduction
A warm welcome to AI tools for endocrinology clinicians, with concrete examples of where AI saves time in daily practice.
Drafting Patient Education for Newly Diagnosed Type 2 Diabetes
Use AI to draft clear, personalised education materials for patients newly diagnosed with type 2 diabetes, saving time while improving consistency.
Explaining Thyroid Conditions in Plain Language
Use AI to translate complex thyroid physiology into clear explanations patients can actually understand and remember.