Referral Letter Generator

Create referral letters from encounter notes

Writing a referral letter well takes more effort than it should. You need to extract the relevant clinical details from the encounter, frame them for a colleague who has never seen this patient, include enough history for context without burying the actual question, and do it all while the next patient is already in the room.

Many clinicians default to brief, templated letters that convey the minimum: a diagnosis, a line about the reason for referral, and a request to evaluate. These get the job done procedurally, but they often leave the receiving specialist guessing about what prompted the referral now, what has already been tried, and what the referring clinician is actually concerned about.

Dictum takes a different approach. Because it has access to the full encounter — either through ambient recording or your post-visit dictation — it can draft a referral letter that reflects the actual clinical conversation. The assessment you shared with the patient, the medications you discussed, the examination findings that raised your concern: these details are already captured. Dictum organizes them into a structured letter that a specialist can read quickly and act on. You review and edit the draft before it goes anywhere.

Common referral letter components

A good referral letter answers the receiving clinician's first three questions: why is this patient being sent to me, what do I need to know about them, and what does the referring physician want me to do? The structure below covers those questions in a format most specialists expect.

• Reason for referral — a clear statement of why you are requesting this consultation. Not just the diagnosis, but the clinical question: is this atypical presentation? Is current management failing? Does this patient need a procedure you do not perform?
• Relevant history — the medical background the specialist needs to understand this referral. Prior treatments, relevant comorbidities, family history if pertinent, and how long the current issue has been present.
• Clinical findings — examination results, lab values, imaging findings, or other objective data from the encounter that support the referral.
• Assessment or concern — your clinical impression. This is where you share what you think is going on and why you want another set of eyes.
• Requested consultation or next step — what you are asking the specialist to do. Evaluate and advise? Co-manage? Take over care? Perform a specific procedure? Being explicit here saves back- and-forth.
• Attachments or supporting documentation — a note about what you are including with the letter: recent labs, imaging reports, prior specialist notes, or medication lists.

Dictum structures its referral letter drafts around these components. If your practice uses a different format or your receiving specialists prefer a particular layout, you can rearrange or edit the draft before sending.

How Dictum preserves clinical context from the encounter

The gap between what happens in an encounter and what ends up in a referral letter is often significant. Details get lost — not because they are unimportant, but because writing the letter is a separate task that happens later, sometimes hours after the visit, when the specifics have faded.

Dictum closes that gap by generating the referral letter from the same source material as the clinical note. The nuances of the conversation — why you are concerned about this particular presentation, what the patient said about symptom progression, which treatments you already discussed and ruled out — are available to the AI because they were part of the recorded encounter.

This means the draft letter often includes clinical reasoning that would typically be left out of a manually written referral. A family medicine physician referring a patient to cardiology, for example, might record during the encounter that the patient's exertional dyspnea has worsened despite optimizing their beta-blocker. That context flows into the referral letter draft automatically, rather than being reduced to "please evaluate for cardiac workup."

The AI SOAP note generator produces the clinical documentation from the same encounter, so both documents stay internally consistent without extra work on your part.

Use cases by specialty

Family medicine and primary care

Primary care clinicians write more outbound referral letters than any other group. A typical AI scribe for family medicine workflow might produce three or four referrals in a single clinic session. Each one requires enough context for the specialist to understand the patient without duplicating the entire chart history.

Dictum helps here by pulling the relevant details from each encounter and organizing them into a focused letter. For a patient being referred to endocrinology, the draft would include the A1c trend, current diabetes medications, any complications discussed, and the specific question — whether that is insulin initiation, thyroid nodule workup, or something else. The clinician reviews the draft, adds anything specific to the specialist's preferences, and sends it through their usual channel.

Specialists coordinating care

Referral letters are not just a primary care responsibility. Specialists frequently refer to other specialists — rheumatology to ophthalmology for uveitis screening, orthopedics to neurology for nerve conduction studies, gastroenterology to surgery for patients who need a procedure beyond the scope of endoscopy.

In these cases, the referral letter needs to be precise about what has already been done. A surgeon receiving a referral from a gastroenterologist needs to know the endoscopy findings, pathology results, and the specific reason surgery is being considered — not just a diagnosis and a request to evaluate. Dictum captures these details from the encounter and organizes them in the draft, reducing the chance that critical information gets left out.

Return-to-referrer communication

Specialists sending patients back to the referring physician can also use Dictum to generate a consultation letter. This letter summarizes what was found, what was done or recommended, and what the primary care physician should monitor going forward. The same structured format applies — Dictum drafts it from the encounter, and the specialist reviews before sending.

Review, edit, and send workflow

Dictum generates a draft. It does not send anything on your behalf. The workflow is straightforward:

1. Record or dictate — capture the encounter using Dictum's ambient recording or enter a post-visit dictation.
2. Generate the referral letter — Dictum produces a structured draft alongside your clinical note and, if requested, an after-visit summary for the patient.
3. Review and edit — read through the letter. Add clinical details you want to emphasize, remove anything that is not relevant to the referral, and adjust the framing for the receiving specialist.
4. Send through your existing systems — copy the letter into your EHR referral module, paste it into a secure message, print it for fax, or export it however your practice handles referrals.

Clinicians should review AI-generated documentation before adding it to the medical record and should use Dictum in accordance with their organization's policies and applicable laws.

Example referral letter structure

Below is the general structure Dictum uses when drafting referral letters. The content is generated from your encounter, but the format follows this outline:

This format works for most outbound referrals. If a specialist or institution requires a different layout, you can restructure the draft before sending.

Referral letters and the rest of your documentation

A referral letter is one of several outputs Dictum can generate from a single encounter. You can also produce a structured clinical note using the AI SOAP note generator and a patient-facing summary with the after-visit summary generator. Each document is generated and reviewed independently — editing the referral letter does not change the clinical note, and vice versa.

To compare what is included at each plan level, see the Dictum pricing plans page.