Write the Authorization Request the Reviewer Can Approve

The person who decides your authorization request never meets your patient. They read what you send, often in a few minutes, against criteria you may never see in full. Your evaluation can be thorough and your plan can be right, and the request still comes back denied, because the documents did not answer the questions the reviewer is required to ask. Those questions are knowable: why does this care need a therapist, why at this frequency, and why for this duration. A request that answers all three, in order, with evidence from your own records, is a request the reviewer can approve without a second round.

This article is about the content of the request: what the clinical documentation has to say and how to structure it. The operational side, which requests to track, who submits them, and how to keep visits from outrunning approved units, is covered in our authorization workflow article. The broader craft of writing notes that show skilled care is covered in the medical necessity article. Here we focus on the single document packet that stands between your plan of care and an approval.

Utilization review is a document check. The reviewer, often a nurse working from the plan’s criteria with therapist or physician review behind them, compares your request against a coverage policy. They cannot credit what the documents do not contain. A strong evaluation that never states why the patient needs skilled care reads, to a reviewer, as a patient who may not need it. The gap between what you know and what you wrote is where denials live.

The scale of that gap is documented. In 2022 the HHS Office of Inspector General reviewed a sample of prior authorization denials from the largest Medicare Advantage organizations and found that 13 percent of the denied requests met Medicare coverage rules. The services would likely have been covered under original Medicare. A common cause was plans applying clinical criteria that go beyond Medicare coverage rules. The same report found that when denials were appealed, plans overturned 75 percent of their own decisions, and that only 1 percent of denials were appealed at all.

Read those three numbers together and the lesson is practical, not cynical. A denial is the start of a process, not a verdict on your clinical judgment. The same documentation discipline that wins the first request also wins the appeal, because both are decided from the record. So the work is the same either way: build a request that answers the reviewer’s three questions from evidence already in your chart.

Medicare already tells you what the skeleton of a defensible request looks like. The Medicare Benefit Policy Manual requires every outpatient therapy plan of care to contain, at minimum: the diagnoses, the long-term treatment goals, and the type, amount, duration, and frequency of therapy services. Commercial and Medicaid reviewers ask for the same elements in different packaging, because those elements are the claim you are making: this patient, this problem, this service, this often, this long. An authorization request is that plan of care, argued.

That framing gives you two working rules. First, the request must agree with the chart. If the plan of care says three visits a week and the request asks for four, or the evaluation lists goals the request never mentions, the reviewer reads the inconsistency as a reason to deny or to pend the request for more records. Second, every sentence in the request should serve one of the three questions. Background that serves none of them is not harmless padding. It buries the sentences the reviewer is searching for.

Coverage for therapy services turns on skill. The reviewer is not asking whether the patient would benefit from attention, exercise, or practice. They are asking whether the service requires the judgment, knowledge, and decision-making of a qualified therapist, or whether a caregiver or an unskilled routine could deliver it safely and effectively. Your request answers by showing decisions, not activities. "Patient practiced word retrieval" describes an activity anyone could supervise. "Therapist adjusted cueing hierarchy from phonemic to semantic cues after error analysis showed perseverative errors" describes a skilled decision with a stated reason.

Three kinds of evidence carry this answer well. Show assessment driving treatment: what you measured, what the result changed about the plan. Show risk that requires clinical management: dysphagia with aspiration signs, fall risk during transfers, a wound or precaution that constrains activity. Show the plan to transfer skill out of the clinic: caregiver training and home program progression are skilled services when they require your teaching and grading, and naming them tells the reviewer you are not planning therapy forever.

Frequency is the question therapists most often leave unargued. The request says "2x/week for 12 weeks" and moves on, as if the number were self-evident. To a reviewer it is not. Tie the frequency to something in the chart: the intensity the condition requires at this stage, the rate of skill decay between sessions you have observed, the caregiver training schedule, or published practice guidance for the condition where it exists. A sentence of the form "twice weekly because X, and we will step down to weekly when Y" does two jobs at once. It justifies the number, and it shows the reviewer a clinician who is already planning the taper.

Duration is a forecast, and a reviewer trusts a forecast with a basis. The basis is the patient’s demonstrated rate of change: progress data from the current episode if you are requesting continued care, or prognosis grounded in the evaluation if the episode is new. State the endpoint as a condition, not only a date: the request ends when the patient reaches the named goals or when progress data says the plan needs to change. For continued-care requests, show what changed since the last authorization. A request that reports the same baseline as the previous one is asking the reviewer to fund a plateau, and the reviewer will decline. If progress has genuinely stalled, make the clinical decision first: change the approach, shift to a maintenance or caregiver-led plan, or discharge. Then write the request, or the discharge summary, to match the decision.

The framework below is payer-neutral. Most portals and forms ask for these elements in some order, and where a payer’s form is thinner than this, attach the narrative anyway. Every bracket is a prompt to pull something specific from your own chart. If a bracket has nothing in the chart to pull, that is a documentation gap to fix before you submit, not a line to skip.

A denial letter names, or is required to name, what the reviewer found missing. For Medicare Advantage, Medicaid, and CHIP plans, CMS rules require a specific reason with every prior authorization denial beginning in 2026, precisely so providers can correct and resubmit. Treat the stated reason as the outline of your appeal: if the reason is "medical necessity not established," the appeal leads with the skilled-care evidence; if it is "requested frequency exceeds guidelines," the appeal argues frequency with the step-down plan and the clinical basis. An appeal that re-sends the original packet unchanged asks the plan to re-grade the same answer.

Use the peer-to-peer review where the plan offers one. It is the one point in the process where a clinician who knows the case speaks to a clinician with authority over it, and it is most effective when you bring the same three answers in spoken form, with the progress data in front of you. And appeal more often than feels natural. The OIG data above says plans reversed three quarters of the denials that were challenged, while 99 percent of denials were never challenged at all.

Two CMS final rules changed what plans owe you, and a request writer should know them. The 2024 Medicare Advantage rule (CMS-4201-F) limits coordinated care plans to using prior authorization only to confirm diagnoses and medical necessity, requires an approval for a course of treatment to stay valid for as long as medically reasonable and necessary to avoid disruptions in care, and gives a patient who switches plans mid-treatment a minimum 90-day transition period during which the new plan may not require prior authorization for the active course of treatment.

The interoperability and prior authorization rule (CMS-0057-F) adds the clock. From January 1, 2026, Medicare Advantage plans, state Medicaid and CHIP programs, and Medicaid and CHIP managed care plans must decide expedited requests within 72 hours and standard requests within 7 calendar days, publish prior authorization metrics, and state a specific reason for every denial. Marketplace plans on the federal exchanges are outside these decision clocks, and the timing rules do not apply to drugs. Original Medicare is a different world again: it generally does not require prior authorization for outpatient therapy at all.