Sign Up
Login
Try for Free
Login
A prior authorization request is the clinical justification packet a provider submits to a health plan to get a mental health service approved for coverage before it starts or continues. Practice staff or the clinician assemble it from the record: diagnosis, symptom severity, functional impairment, treatment history, risk, and a plan with measurable goals. No universal form exists; each payer or carve-out vendor sets its own.
Practice staff (billers, care coordinators) or the treating clinician
The payer or its behavioral health carve-out vendor: utilization review staff and peer reviewers
300 to 600 words plus attachments · 20 to 45 minutes by hand (clinical team estimate)
Payer utilization management form plus attachments (compare: prior authorization letter, insurance appeal letter)
Before a plan-gated service starts, and at concurrent review before approved sessions run out
Payer-specific by contract, no universal form; law sets the decision clocks; denial risk is high
A prior authorization request is the clinical justification packet a provider submits to a payer to get coverage approved before delivering or continuing a mental health service. It is a payer-facing utilization management document assembled from the clinical record, and it is not the claim: the authorization establishes that a service will be covered, while the claim asks for payment after the service is delivered. Nobody designed it. It grew out of managed care utilization review in the 1980s and 1990s and remains a payer-industry artifact, which is why it answers to so many names: prior authorization or PA, pre-authorization, precertification, treatment authorization request (TAR, the Medicaid term), service authorization request, outpatient treatment report (OTR, the older behavioral health carve-out term), and, for continuing care, concurrent review or continued-stay review.
Two boundaries carry this page. First, the request is not the prior authorization letter: the request is the payer's structured form or portal submission, while the letter is the clinician's narrative support that travels with it when the form cannot hold the complexity. Second, the obligation to obtain authorization is almost never law. It is payer contract and policy, plan by plan, and traditional Medicare imposes none of it on outpatient behavioral health: GAO reported in 2025 that CMS "does not currently require prior authorization for any behavioral health services" in traditional Medicare (GAO-25-107342). What the law does regulate is how payers run the process: decision clocks, licensed review of denials, and mental health parity. And for ongoing outpatient therapy, the operative version is usually a concurrent review triggered after a session threshold rather than an upfront gate, because outpatient care is already the least restrictive level.
Solo and group outpatient practices, intensive outpatient and partial hospitalization programs, psychological testing practices, and psychiatry clinics requesting services such as transcranial magnetic stimulation all meet the request, and the person completing it is as often a biller or care coordinator as the treating clinician. The gate sits with Medicare Advantage plans, Medicaid managed care organizations, and commercial plans, usually operated through a behavioral health carve-out vendor (Optum, Carelon, Magellan, Evernorth) whose portal and criteria control the process regardless of the logo on the member's card. The gate is also narrower than most clinicians assume: in GAO's 2025 review, 8 of 9 selected Medicare Advantage organizations required prior authorization for behavioral health, almost always for higher levels of care, and none required it for in-network outpatient counseling and psychotherapy. What outpatient clinicians actually meet is concurrent review once length or frequency passes the plan's threshold, and it uses the same packet. Use this template for the form-driven submission itself. When the form cannot carry the clinical argument, attach a medical necessity letter; when the payer's decision comes back, it arrives as a utilization review summary, which tells you what any appeal will have to rebut.
No universal form exists: no federal rule prescribes one for outpatient psychotherapy, so every payer and carve-out vendor builds its own. The sections below are the superset those forms draw from. Treat them as the checklist and map them onto whatever labels your payer's portal uses.
Member, provider, and service header. Member name, ID, and plan; the behavioral health carve-out vendor if one manages the benefit; provider name, credentials, license, and NPI; and the request stated in billing terms: CPT codes, units, frequency, and the date range. Pitfall: authorization obtained from the medical plan when a carve-out vendor actually manages behavioral health; a number from the wrong entity does not bind the one paying the claim.
Diagnosis. The specific ICD-10-CM code the record supports, written as code plus name. Pitfall: a Z-code or rule-out diagnosis standing alone; reviewers treat both as failing to anchor medical necessity, and it is one of the most common behavioral health documentation failures.
Symptom severity and functional impairment. Current symptoms with severity, current measure scores (PHQ-9, GAD-7), and what the client cannot do this week: work, school, relationships, self-care. Pitfall: severity adjectives with no function. "Severe anxiety" approves nothing; "missed three of five workdays" is something a reviewer can defend approving.
Level-of-care justification. Why this setting and intensity, argued against the criteria set the reviewer will score: LOCUS for adults, CALOCUS-CASII for children and adolescents, ASAM for substance use, or the plan's InterQual or MCG criteria. Pitfall: asserting the level ("weekly outpatient is appropriate") instead of scoring against the named criteria; the reviewer is completing a criteria worksheet, so write to it.
Treatment history and prior response. What has been tried and what happened: prior episodes, medication trials, earlier courses of therapy, partial responses, and the failed or insufficient alternatives that make the requested service the right next step. Pitfall: omitting response to treatment so far; in a continued-care request, measured progress plus remaining impairment is the whole argument, and its absence reads as care without direction.
Risk factors. Current risk status stated plainly, either way, with protective factors and the monitoring plan. Pitfall: silence on risk; a packet that never mentions it invites the reviewer to assume it was never assessed.
Treatment plan, goals, and requested course. Measurable goals tied to the impairments above, the interventions that address them, and the requested frequency and duration restated as the plan's arithmetic. Pitfall: vague, untracked goals ("improve coping skills") with no baseline or target; reviewers list this among the classic behavioral health documentation failures.
Urgency. Standard or expedited, chosen deliberately. Expedited review exists in law: 72 hours under Medicare Advantage rules and CMS-0057-F, and 72 hours for urgent care claims under ERISA plans. Pitfall: filing routine when a delay could seriously harm the client; the standard clock can run 7 to 15 days depending on plan type, and the expedited flag is the only thing that shortens it.
PRIOR AUTHORIZATION REQUEST (payer-neutral worksheet)
Member: ____________________ ID: ____________ DOB: __________
Plan: ______________________ BH carve-out vendor: _____________
Provider: __________________ Credentials/License: _____________
NPI: ____________ Phone/fax/portal: ___________________________
REQUESTED SERVICE
CPT code(s): ____________ Units: ______ Frequency: ___________
Start date: __________ End date: __________
[ ] Initial request [ ] Concurrent / continued-stay review
Urgency: [ ] Standard [ ] Expedited (delay risks serious harm)
DIAGNOSIS (specific ICD-10-CM code plus name; no rule-out alone)
________________________________________________________________
SEVERITY AND FUNCTIONAL IMPAIRMENT (scores plus this week's function)
Measures: PHQ-9 ____ GAD-7 ____ Other: ____________
Impairment: ____________________________________________________
LEVEL-OF-CARE JUSTIFICATION (LOCUS / CALOCUS-CASII / ASAM / plan criteria)
________________________________________________________________
TREATMENT HISTORY AND PRIOR RESPONSE (what was tried, what happened)
________________________________________________________________
RISK FACTORS (current status, protective factors, monitoring)
________________________________________________________________
TREATMENT PLAN AND GOALS (measurable, with baseline and target)
________________________________________________________________
ATTACHMENTS: [ ] treatment plan [ ] recent progress note
[ ] score sheet [ ] medical necessity letter
Clinician signature/credentials: ______________ Date: _________Free to use and share, no signup. The PDF includes a one-page cheat sheet with section-by-section pitfalls and a pre-sign checklist; the DOCX is the blank template, ready to adapt.
Scenario: concurrent (continued-stay) review request submitted to a commercial plan's behavioral health carve-out vendor, seeking 12 more weekly psychotherapy sessions for an adult with recurrent major depressive disorder. All details are fictional.
Prior Authorization Request (concurrent review). Member: R.D., 41 · Plan: [commercial HMO], behavioral health managed by [carve-out vendor] · Submitted: 08/04/2026 · Provider: T. Alvarez, LCSW, NPI [on file]
Requested service: Individual psychotherapy, 90834, 12 sessions at 1 per week, 08/11/2026 to 11/03/2026. Concurrent review of authorization A26-01187 (16 sessions approved 04/2026; 14 used as of this request). Urgency: standard.
Diagnosis: F33.1, major depressive disorder, recurrent episode, moderate. Confirmed at intake 04/07/2026; unchanged at treatment plan review 07/29/2026.
Severity and functional impairment: PHQ-9 on 07/29/2026: 12, down from 19 at intake. GAD-7: 9. Client has reduced missed workdays from about 2 per week in April to 1 in the past month, but continues to avoid team meetings, has resumed only one of three previously dropped activities, and reports early waking twice weekly with 5 to 6 hours of sleep.
Level of care: Outpatient individual therapy remains the indicated level under LOCUS dimensions: moderate symptom severity, no acute risk indicators, engaged and adherent, recovery environment supportive, functional recovery incomplete. No indicators for a higher level of care. Step-down to biweekly is planned once PHQ-9 remains below 10 for four consecutive weeks.
Treatment history and prior response: Second lifetime episode; 2019 episode remitted after approximately 6 months of CBT. Current episode began 01/2026. Sertraline 100 mg daily prescribed and monitored by PCP since 03/2026, adherent. Fourteen CBT sessions to date with partial response: PHQ-9 19 to 12, behavioral activation targets met, cognitive restructuring of self-critical beliefs in progress.
Risk factors: Denies current suicidal ideation; passive ideation reported at intake resolved by session 4 and has not recurred. No history of attempts. Protective factors: engaged in treatment, stable housing, supportive partner. Screened each session; monitoring continues.
Treatment plan and goals: Goal 1: PHQ-9 below 10 by 11/2026 (baseline 19, current 12). Goal 2: full workweek attendance for 4 consecutive weeks. Goal 3: resume two of three dropped activities. Interventions: weekly CBT with behavioral activation and cognitive restructuring; medication coordination with PCP; re-administer PHQ-9 and GAD-7 every 4 weeks.
Attachments: Updated treatment plan (07/29/2026), PHQ-9/GAD-7 score sheet, most recent progress note. Signed: T. Alvarez, LCSW, 08/04/2026.
This sample is fictional and for educational purposes. It does not describe a real patient.
Writing these after every session? BastionGPT drafts complete notes from bullets, dictation, or a transcript.
Generate a note from bulletsThe packet is assembled from the designated record set and then leaves your hands, so know what a payer can and cannot demand. Payers can require the medical-necessity elements of the treatment record: diagnosis, severity, functional impairment, the treatment plan, and progress. Psychotherapy process notes kept separate from the record carry heightened HIPAA protection and generally cannot be required as a condition of payment, so build the packet from progress notes, the plan, and outcome measures, never from process notes. File the request, the authorization number, and the payer's response alongside your billing note trail, because an approval is not payment: it confirms medical necessity was accepted, and the claim can still fail on coding, eligibility, timely filing, or place of service. Retention of the packet has no uniform national rule in any of the three countries this library covers; state or provincial licensing, privacy law, and the payer contract set it, and the longest applicable period wins.
The clocks around the request are law even though the request itself is not. Medicare Advantage plans must decide within 14 days standard and 72 hours expedited under 42 CFR 422.568 and 422.572, and an approved authorization is valid for the enrollee's full course of treatment. Employer group plans under ERISA owe answers in 15 days pre-service and 72 hours urgent under 29 CFR 2560.503-1. From 2026, the CMS Interoperability and Prior Authorization Final Rule (CMS-0057-F) holds Medicare Advantage, Medicaid, CHIP, and federal exchange plans to 7 calendar days standard and 72 hours expedited, with a specific denial reason and public reporting of authorization metrics. Mental health parity adds one more layer: a plan that imposes prior authorization on mental health benefits must be able to show the limit is no more stringent than on comparable medical benefits, though the 2024 tightening of those rules sits under a non-enforcement posture while litigation runs. So work the process, not just the form: identify the carve-out vendor before the first session, run a session counter that flags at 75 percent of approved visits, file the concurrent request 3 to 5 business days before the authorization expires with a current treatment plan review and re-administered measures, and if the answer is no, the next document is an insurance appeal letter, not a resubmission.
The denial data here is unusually good, and it points at process failures more than clinical ones. The HHS Office of Inspector General found that 13 percent of sampled Medicare Advantage prior authorization denials were for care that met traditional Medicare coverage rules (OEI-09-18-00260, 2022), and that Medicaid managed care organizations denied 1 in 8 requests in 2019, more than double the Medicare Advantage rate. Per KFF's analysis of CMS data, nearly 53 million prior authorization requests went to Medicare Advantage insurers in 2024, only 11.7 percent of 2023 denials were appealed, and yet more than eight in ten appeals overturn the denial. The lesson sits in that gap: most denials are never challenged, and most challenges win. The BastionGPT Clinical Advisory Board sees the same errors most often in prior authorization request reviews:
| Aspect | United States | Canada | Australia |
|---|---|---|---|
| Status | Payer policy, not statute; traditional Medicare requires none for outpatient behavioral health, while Medicare Advantage, Medicaid managed care, and commercial plans gate services by contract inside legally set decision clocks | Structurally absent from public plans: psychotherapy outside hospitals and physician practices is mostly privately insured, and private plans reimburse claims rather than pre-approve care | No clinical prior authorization; Better Access is a rebate-eligibility gate (diagnosis, GP Mental Health Treatment Plan, referral), and private hospital cover runs an administrative eligibility check, not clinical review |
| Terminology | Prior authorization, precertification, TAR (Medicaid), OTR, concurrent or continued-stay review | Preapproval or predetermination (uncommon for therapy), extended health care claim, treatment plan request | Mental Health Treatment Plan, referral, eligibility check (ECLIPSE), Informed Financial Consent |
| What changes | Whether a gate exists at all varies by plan and carve-out vendor; higher levels of care are gated most, and outpatient therapy usually meets concurrent review instead | Annual dollar maximums and eligible regulated-provider rules replace clinical review; some insurers request a treatment plan or physician referral for intensive or higher-cost claims | Rebate caps (10 individual and 10 group services per calendar year) and referral validity replace payer review; clinical necessity stays with the treating clinician |
| Retention | No uniform national rule for the packet; state licensing, privacy law, and the payer contract govern | Provincial privacy law and college standards govern the underlying record (for example Ontario's PHIPA) | State rules and the Psychology Board's Code of Conduct (effective December 1, 2025) govern the record; no authorization-specific rule |
Only the US column describes true clinical prior authorization, and even there the gate is contractual and concentrated in managed care. Read any claim about "prior authorization" against the country and plan type it names. In Canada and Australia this template still organizes a clean medical-necessity summary, but no payer requires it in this form.
BastionGPT is specifically trained, tuned, and clinically tested on prior authorization requests.
See how clinicians use it day to day on the AI therapy notes page.
Many BastionGPT users report saving more than 90 minutes per day on documentation.
HIPAA-compliant with a signed BAA on every plan. Your data is never used to train models. BastionGPT drafts, you review and sign.
Usually not upfront. In GAO's 2025 review, none of the nine selected Medicare Advantage organizations required prior authorization for in-network outpatient counseling and psychotherapy, and traditional Medicare requires none for any outpatient behavioral health service. The gates concentrate at higher levels of care: inpatient, partial hospitalization, transcranial magnetic stimulation. What outpatient clinicians actually meet is concurrent review once sessions pass the plan's threshold, which uses the same packet. Verify plan by plan; the answer lives in the contract, not in a general rule.
No. The obligation to obtain authorization is payer contract and policy; no statute prescribes the packet, and no universal form exists. What the law regulates is the payer's side of the process: decision timeframes, licensed-clinician review of denials, and mental health parity, which requires plans to show that authorization rules on mental health benefits are no more stringent than on comparable medical benefits. The concept itself is US managed care: Canada's public plans have no equivalent gate for psychotherapy, and Australia's Better Access is a rebate-eligibility gate rather than payer review.
The clocks are law and depend on plan type. Medicare Advantage: 14 days standard, 72 hours expedited. Employer plans under ERISA: 15 days pre-service, 72 hours urgent, and 24 hours for a concurrent care extension requested at least 24 hours before approved care ends. From 2026, CMS-0057-F holds Medicare Advantage, Medicaid, CHIP, and federal exchange plans to 7 calendar days standard and 72 hours expedited, with a specific reason for any denial. A June 2025 industry pledge also commits most large insurers to real-time electronic approvals for most requests by 2027. If a delay could seriously harm your client, request expedited review and say why.
The request is the payer's structured form or portal submission: fields, codes, criteria. The prior authorization letter is the clinician's narrative support letter that travels with it when the form cannot capture the situation: an atypical presentation, extended sessions, an out-of-network exception. If coverage has already been denied, neither is the right tool; that calls for an insurance appeal letter, which runs on post-denial deadlines and preserves external review rights.
On most current plans, no upfront authorization, but it draws elevated scrutiny. UnitedHealthcare and Optum required authorization for 90837 until 2018 to 2019 and then dropped it on many plans; some plans still limit extended-session frequency. Because 90837 pays more than 90834 and requires a documented minimum of 53 minutes, missing start and stop times is a frequent clawback trigger in post-payment audits. Document the time and the clinical reason the longer session was necessary.
They risk denial under CO-197, the authorization-absent code, and because of the CO group code the balance cannot be billed to the patient. Some plans allow retroactive authorization within a short window or for urgent situations; many do not, and it is never guaranteed. Verify benefits and the authorization requirement before the first gated session. Even with an approval in hand, payment still depends on coding, eligibility, timely filing, and the claim matching the authorized codes, units, and dates.
Payers can require enough of the treatment record to establish medical necessity: diagnosis, severity, functional impairment, treatment plan, and progress. Psychotherapy process notes kept separate from the record have heightened protection under HIPAA and generally cannot be required as a condition of payment. Build the packet from progress notes, the treatment plan, and outcome measures, and keep process notes out of it.
Track the approved session count and flag at 75 percent. Submit the continued-stay packet 3 to 5 business days before the authorization expires, with an updated treatment plan review, re-administered outcome measures, and a current diagnosis list. If it is denied, request a peer-to-peer review with a same-specialty reviewer and appeal inside the plan's window, commonly 30 to 180 days. The appeal math favors you: only about one denial in nine is appealed, and more than eight in ten appeals overturn the denial.
Yes. Give it the treatment plan, recent notes, and scores, and it drafts the packet: specified diagnosis, severity paired with function, level-of-care justification against the criteria you name, prior response, risk, and measurable goals, ready to map onto the payer's form. It can also turn a denial into the outline of an appeal. BastionGPT is HIPAA-compliant with a signed BAA on every plan, and your data is never used to train models.
Educational content, not legal or billing advice. Sample notes are fictional. Follow your organization's policies and your board, payer, and jurisdiction requirements.