CLINICAL · DOCUMENTS

Clinical Letters in Mental Health: Prior Authorizations, Referrals, and Treatment Summaries

Published June 15, 2026

Clinical notes get most of the attention in documentation conversations, but they are not the only writing that lands on a clinician's desk. Mental health practitioners are routinely asked to produce letters — a prior authorization so a medication stays covered, a referral to a specialist, a treatment summary for a new provider, an appeal after a denial, or a letter supporting an accommodation. Each one pulls from information you have already documented, yet each one means re-reading the chart, reformatting it for a different audience, and finding the right tone for a payer, a colleague, or an attorney.

These letters are a hidden documentation burden: low in volume compared to progress notes, but high in friction and often time-sensitive. This guide walks through the letters mental health clinicians write most often, what each one needs to contain, and how to produce them faster by drawing from the notes you have already written.


The letters mental health clinicians write most

Most clinical correspondence falls into a handful of recurring categories. They differ in audience and purpose, but they share a common backbone: a clear statement of the clinical situation, the relevant history, and a specific ask or conclusion.

Prior authorizations justify medical necessity to an insurer so treatment is covered. Referral letters hand off a patient to a specialist with the context they need. Treatment summaries bring a new provider up to speed on the course of care. Appeals contest a denial with documentation and reasoning. Return-to-work, accommodation, and ESA letters speak to employers, schools, or housing providers about functional needs. Each draws on the same underlying record — the diagnoses, medications, and progress you have already captured in your notes.

Prior authorization letters

Prior authorization is where letters most directly affect care: a delayed or denied authorization can interrupt a medication that is working. Payers are looking for a specific story — the diagnosis, what has been tried, how the patient responded, and why the requested treatment is medically necessary right now. The strongest prior auth letters read like the Assessment and Plan of a good SOAP note translated for a non-clinician reviewer: diagnosis with specifiers, objective measures, treatment history, and the consequences of discontinuation.

Example excerpt: "Ms. R. carries a diagnosis of Generalized Anxiety Disorder (F41.1) and Major Depressive Disorder, recurrent, moderate (F33.1). She has trialed sertraline (inadequate response at 150mg) and bupropion (discontinued due to side effects) prior to escalation to her current regimen. On the current medication her GAD-7 has improved from 18 to 7 over twelve weeks, with restored occupational functioning. Discontinuation or non-formulary substitution would pose a substantial risk of symptom recurrence and decompensation. Continued authorization is medically necessary."

Because prior authorizations hinge on diagnosis and treatment history, they are tightly linked to coding. Our guide to CPT codes is a useful companion when you are documenting the medical-necessity language payers expect.

Referral and treatment-summary letters

A referral letter exists to make the receiving clinician's job easier. It should answer one question clearly — what do you need this specialist to do?— and supply only the history that supports it: presenting concern, relevant diagnoses, current medications, and the specific clinical question. A treatment summary serves a related purpose when a patient transfers care or requests records: it narrates the course of treatment, what was tried, how the patient responded, and where things stand today.

Referral, in brief: "I am referring Mr. T. for neuropsychological evaluation to clarify the contribution of attentional versus mood-related factors to his cognitive complaints. He is currently stable on his mood regimen; the specific question is whether his persistent concentration difficulties are better accounted for by ADHD. Relevant history and current medications are summarized below."

The discipline in both letters is restraint: include what the reader needs and omit what they do not. A referral does not require the entire chart, and minimizing unnecessary protected health information is good practice as well as good etiquette.

ESA, accommodation, and return-to-work letters

Letters that speak to employers, schools, and housing providers carry a different weight because a third party will act on them. An emotional support animal letter, a workplace accommodation request, or a return-to-work clearance should state only what falls within your scope and what you can clinically support — the existence of a condition that substantially affects a major life activity, and the functional need being addressed — without overreaching into specifics that are not yours to certify.

Accommodation, in brief: "Ms. K. is under my care for a diagnosed anxiety disorder that, at present, intermittently affects her ability to concentrate in high-stimulation environments. A reduced-distraction workspace or the option to use noise-cancelling headphones would reasonably accommodate this functional limitation. I am happy to provide additional information as appropriate."

These letters reward conservatism. When a request exceeds what you can clinically support, the defensible move is to document what you can attest to and decline the rest — clinician judgment, not a template, governs what goes in.

Generating letters from your notes

Every letter above is a reformatting of information you have already documented. That is exactly the kind of work AI is suited to — not inventing clinical content, but restructuring your existing record for a new audience. The Documents AI feature in MH Scribe drafts letters from your finished clinical notes (never the raw transcript), drawing on a single session or the patient's entire history.

It ships with ready-made templates for the documents clinicians need most — referrals, treatment summaries, prior authorizations, appeals, and return-to-work and ESA letters — but it is not limited to them: you can describe any letter you need and generate it from scratch. Drafts are produced on your own letterhead with your signature applied, and you review and sign every document before it leaves your practice. The notes those letters draw from are the same ones you build during a psychiatric intake or follow-up visit.

Review, sign, and stay compliant

However a letter is drafted, the clinician remains responsible for its content. A generated draft is a starting point, not a final document: read it, correct it, and confirm it reflects your clinical judgment before signing. AI accelerates the writing; it does not transfer accountability.

The same data-protection standards that apply to your notes apply to letters generated from them. MH Scribe processes protected health information under a HIPAA Business Associate Agreement and does not use your data to train AI models — the reasoning behind embedding the BAA directly in our terms applies here just as it does to clinical notes. For a broader look at how MH Scribe compares on documents and letters specifically, see our AI scribe comparison.

Turn your notes into letters in seconds

Documents AI drafts referrals, treatment summaries, prior authorizations, and more — from your clinical notes, on your letterhead, ready for you to review and sign.