Prior authorization support, explained.

What prior authorization support means for one patient and one prescription

One specialty drug request in five steps, and the step where the proof gets built

1. The prescriptionThe doctor chooses a therapy. The office learns the plan wants prior authorization before it will pay.
2. The plan's listThe coordinator finds the plan's current policy for this drug and this disease, with its form or electronic questions.
3. The proof, line by lineEach line is matched to a chart page, gaps are chased, the form is filled and the letter is drafted. This is the step this guide is about.
4. The signature and the sendThe prescriber reviews the request, signs it and it goes to the plan, by electronic request, portal, fax or phone.
5. The decisionThe plan approves, asks a question or denies. A denial starts an appeal, and the patient keeps waiting.

What new PA coordinators find surprising about the work

What PA, step therapy, an LMN and ePA mean

Fourteen terms are defined in the PA support summary and glossary.

What the prescriber's office prepares, what the plan decides, and where a drug maker's program may help

In shortThree parties touch a specialty request. The office owns the request and the prescriber signs it. The plan or its pharmacy benefit manager (PBM) reviews and decides. A manufacturer's support program may help offices understand the process, under limits explained in section 12.

People search for "prior authorization automation" and find two different markets. One serves plans, whose reviewers decide requests. The other serves offices, whose staff prepare them. A tool for one is close to useless for the other, because the two jobs start from opposite documents. A reviewer starts with a request and asks whether it meets the policy. A coordinator starts with a chart and asks what in it proves each line, and what is missing.

A specialty policy is short to read and slow to satisfy. The fictional Crestline policy in this guide has six lines, and two of them are step therapy. Those two lines are where the evidence is hardest to collect, because they reach back across months of care, several drugs and sometimes several pharmacies.

Coverage can also sit in two places. A self-injected drug usually falls under the pharmacy benefit, reviewed by a PBM. A drug given in a clinic usually falls under the medical benefit, reviewed by the plan's medical team. A coordinator has to know which applies before choosing the form. The Centers for Medicare & Medicaid Services (CMS), the federal agency that runs Medicare, treats the two differently in its drug proposal, as section 05 explains.4

CMS-0057-F, the prior authorization rule: what began January 1, 2026, what comes January 1, 2027, and why drugs sit outside it

In shortCMS's Interoperability and Prior Authorization final rule was released January 17, 2024. Decision deadlines and specific denial reasons began January 1, 2026, the first public metrics were due March 31, 2026, and the data links known as APIs generally follow by January 1, 2027. Every part of it excludes drugs.

Which plans CMS-0057-F reaches, and the exchange plan exception

The rule names its "impacted payers": Medicare Advantage organizations, state Medicaid and Children's Health Insurance Program (CHIP) fee-for-service programs, Medicaid managed care plans, CHIP managed care entities, and qualified health plan issuers on the federally facilitated exchanges, the federally run marketplaces for health coverage.1

Those exchange plans are excluded from the decision deadlines, a detail many summaries drop.1 Employer plans and most other commercial coverage are outside the rule altogether: they answer to ERISA, the federal law on employer benefit plans, the Affordable Care Act's claims rule and state law, as section 04 sets out.

How the deadline works when it does apply

• Standard requests: seven calendar days. Expedited requests, the fast-track kind: 72 hours. CMS measures both in "calendar time, not business hours".8
• The clock starts when the payer receives the request, and it "does not stop or restart" if the payer later asks for documentation it had not disclosed to the provider.8 Some programs may grant an extension of up to 14 calendar days.8
• A denial must carry a specific reason, whichever way the request was sent.1
• Metrics go on each payer's website every year, starting with 2025 data by March 31, 2026.8

Why a specialty office should still follow this rule

Three things reach the office anyway.

• The systems. The Prior Authorization API and the CMS measure that asks clinicians and hospitals to request PA electronically put these requests onto the health record systems offices use. They start with non-drug care, and CMS has proposed extending them to drugs.4,7
• The metrics. They give the office a first public view of how each large insurer behaves, and KFF's first analysis of them appears in section 10.
• The denial reason. The habit of a specific reason sets an expectation an office can use when it asks a commercial plan or a PBM to explain itself.

How long a plan has to answer a drug prior authorization: 24 hours in Medicaid, 72 in Part D, up to 15 days in an employer plan

In shortDrugs left out of CMS-0057-F still have deadlines, set by the rule for each type of coverage. Medicaid and expedited Part D requests run on 24 hours. Standard Part D runs on 72. An ERISA employer plan may take 15 days on a pre-service request. Knowing which clock applies tells the office when to chase.

Part D's rule also sets what the prescriber must say to win an exception to a step. The statement has to say the drug is medically necessary, and give a ground: the required alternatives have not worked, are likely not to work for this patient, or are likely to cause harm.18 The plan may ask for that in writing and may ask for documents.18 A coordinator who knows this drafts for the prescriber the facts that support one of those grounds, then lets the prescriber decide whether it is true.

Electronic prior authorization for drugs: CMS-0062-P is still a proposal, and the NCPDP dates that are already set

In shortCMS proposed in April 2026 to bring drugs into electronic prior authorization, with most changes from October 1, 2027. As of November 2026 it has not been finalized. Part D already has its own electronic standard, with a firm date of January 1, 2028. ePA moves the questions faster. It does not find the proof.

What the drug proposal would do, if finalized

• Status: CMS-0062-P was published at 91 FR 19890 on April 14, 2026, and the comment period "closed on June 15, 2026".4,19 CMS's page still called it proposed on November 3, 2026, and we found no final rule.19 One part has moved: the FY 2027 inpatient payment final rule adopted updated electronic PA data standards first listed in the proposal, effective October 1, 2026, according to WEDI.20
• Who: Medicare Advantage, Medicaid and CHIP programs and plans, exchange plans, and small-group exchange issuers.4,17
• How: drugs under a medical benefit through the FHIR Prior Authorization API; drugs under a pharmacy benefit through NCPDP standards. FHIR and NCPDP are data formats that let systems exchange requests.4
• When: October 1, 2027 for most changes, with drug metrics starting in 2028.4,17
• Part D: covered Part D drugs stay under their existing rules. Plans that offer both Medicare Advantage and Part D would report drug metrics only for Part B drugs.4,17

What ePA changes in the office, and what it leaves alone

An electronic request sends the plan's questions into the prescriber's system and the answers back. It removes faxes, retyping and hold music. CAQH's yearly index covers medical and dental transactions, and reports electronic adoption of medical PA rising from 31% in its 2023 edition to 40% in its 2025 edition. We read a trade report of the index, which gives no figure for drug requests.23

The same index puts potential medical savings from automating remaining manual transactions at $18.7 billion a year, across all medical transaction types.23 We found no public figure for the share of specialty drug requests sent electronically.

Speed depends on what the electronic question finds. One network vendor reports approvals "returned in just 18-seconds" on 83 medications, with 45,275 prescribers enabled across 20 health systems in 2025. That is a company's own claim, and the release gives no method.24 A result like that needs the answers to be ready. An electronic question that asks "documented trial of methotrexate, at least 3 months?" still needs someone to know whether the answer is yes and which page shows it.

The insurers' 2025 pledge: fewer requests so far, and no figures for drugs

In June 2025, health insurers made a voluntary pledge to reduce and simplify prior authorization. By April 2026 they reported about 11% fewer requests, roughly 6.5 million, and over 15% fewer in Medicare Advantage.25

AHIP, the health insurers' trade group, describes a 2027 commitment that at least 80% of approvals on electronic requests with all necessary documents will come in real time.26 The reports we read do not show these figures reaching prescription drugs, so an office should not plan on them for a biologic. Physicians are cautious too: in the American Medical Association's December 2025 survey, 33% believed the pledge would make a meaningful difference.26

Payer medical-necessity criteria for biologics: step therapy, specialist rules and limits that have grown stricter than the label

In shortA plan's policy is a list of conditions, and its small words decide requests: AND, OR, "at least", "documented in", "within". Commercial specialty policies have tightened: by 2021, 51.7% were stricter than the drug's FDA-approved label, and step therapy and prescriber rules had both grown. Step therapy laws help in some states and do not reach most employer plans.

The six kinds of line in a specialty policy, read from the fictional Crestline one

How specialty coverage has tightened since 2017

Those figures come from a Tufts analysis of 4,418 policies in 2017 and 4,998 in 2021, across 357 drug and indication pairs.27 They are shares of policies and stop at 2021, the latest published year.

An earlier study found step therapy on 38.9% of specialty coverage policies, ranging from 20.6% at one plan to 57.5% at another.28 Medicare has moved the same way: Part D plans restricted an average of 31.9% of compounds in 2011 and 44.4% in 2020, counting prior authorization, step therapy and exclusions from the plan's covered drug list, on non-protected-class compounds.29

What step therapy laws do and do not reach

A 2026 article in the American Bar Association's Health eSource counts 38 states plus D.C. with some form of step therapy regulation.30 What those laws do for a patient depends on who stands behind the plan.

The federal Safe Step Act (S. 2903), introduced September 18, 2025, would close that gap, and we found no record of its passage as of November 2026.31 So the same step may be contestable for one patient and untouchable for the next, in the same waiting room. An office needs to record which kind of plan each patient has, because the appeal letter changes with it.

Plans also rewrite these lists. When a new version arrives, a request that was fine last week may fail today. A separate guide on this site, payer coverage policy tracking, covers that work from a drug maker's seat. An office's version of the problem is smaller: know which version applies to this request, and use only that.

Where the proof for a PA criterion hides, and how to count weeks on therapy, lab age and either-or lines

In shortChart exports leave things out: telehealth notes, outside labs and fills at another pharmacy. Most errors in this work are small sums done by hand: weeks from fill dates, days since a lab, which option of an either-or line applies. Working these through with one fictional patient shows why a person should see the page behind each number.

Four sums every coordinator does, on one fictional chart

When visit notes and pharmacy fills disagree, the prescriber sees both

Notes can be vague, and fills can overstate. A fill proves only that a pharmacy handed over a box. The patient may have missed doses, and a fill after a drug was stopped says nothing about treatment. A note can round "about three months" up. Offices that compare the two see disagreements early. The right response is to show the prescriber both pages and let the prescriber say what happened, even when one version would meet the policy and the other would not.

Wording matters as much as dates. "Failed methotrexate" in a template is weaker evidence than "inadequate response at 18 weeks, BSA 15%". If the note says less than the policy asks, the fix is a new or amended note written by the clinician who knows. A stronger word in the letter fixes nothing.

HIPAA, the federal health privacy law, also matters here. Its minimum necessary standard applies to what you send a plan: reasonable efforts to limit information to what the purpose needs.32 It is good practice as well, because a reviewer reading 17 pages with passages marked finds the proof faster than one reading 51.

What a good letter of medical necessity contains, why every sentence must trace to the chart, and who has to sign it

In shortA letter of medical necessity explains, in the prescriber's voice, why this patient needs this treatment. It works when each statement points to a chart page and the prescriber has read it. Templates fail in two ways: they say nothing specific, or they claim more than the chart holds. A clinician signs, always.

The five parts of a one-page letter

Why the prescriber's name on a request has real weight

A prior authorization request is a statement to a payer on behalf of a named clinician. In November 2021 the US Department of Justice announced a $12.7 million False Claims Act settlement with a drug maker, kaléo. The government alleged that pharmacies submitted false prior authorization requests for its product, misrepresenting that prescribers had submitted them, between March 14, 2017 and April 30, 2020.33 The release states that these are allegations only, with no determination of liability.33

For any office, the lesson is that the person whose name is on the request reviews it, the login belongs to that person, and nobody else submits as them.

Coordinators may draft. Software may draft. A clinician reads what was drafted, changes what is wrong and decides whether to sign. Anything clinical in a request, including a claim about why a treatment failed, is that clinician's statement.

Missing notes, stale labs, old forms: eight failures a specialty PA request carries to the plan

In shortMost failures are preventable in the office: a missing note, a lab out of date, a form from last year, a letter that says too little. No public study counts these eight causes across practices, so the ranking below is our judgment. Your own denial letters will rank them for you.

Two small studies show how much of this comes back. At one complex dermatology clinic, 51% of 51 prior authorizations were first denied, with systemic drugs denied more than topical ones.34 At one rheumatology center, 21% of infused-drug requests were first denied, and 82% of those were eventually approved.35 Neither is a national rate, and the second suggests something a coordinator knows well: many first denials end in approval, after a delay the patient should not have had.

What PA denial letters leave out, what the first public metrics show, and how often an appeal wins

In shortDenial letters often do not say who decided or what would have worked. Published data show low appeal rates and high success rates, but the Medicare Advantage figures do not separate drugs and the newer metrics leave drugs out. For a drug request, your own denial and appeal log is the best evidence you will get.

That review, published in the Journal of the American Academy of Dermatology, also found that 41% of letters gave no information on step therapy or covered alternatives, and that among appeal denials only 31% engaged with the prescriber's arguments.36,2 We read a report of the paper and not its full text, and that report does not give the number of letters. A denial that names no missing fact is hard to answer.

The Medicare Advantage funnel from request to overturned denial

KFF's data are Part C organization determinations, the plans' coverage decisions, and CMS does not report them by service type, so they do not tell you about drugs under Part D or a specialty biologic under an employer plan.3 A high overturn rate also does not prove the first denial was wrong. It may mean only appeals with good evidence were filed, or that the missing page finally arrived. Denial rates ranged from 4.2% to 12.8% among large insurers.3

What the first CMS-0057-F metrics reveal, and where they stop

KFF's analysis, published August 13, 2026, read the metrics of 14 insurers. Median response was about one day on standard requests, and on expedited requests about 0.4 days in Medicare Advantage, 0.8 in Medicaid managed care and 1 day in the Marketplace.37 On standard requests, appeals overturned denials 67% of the time in Medicare Advantage, 47% in Medicaid managed care and 43% in the Marketplace.37

KFF lists the gaps: percentages with no counts, no service categories, optional templates, no share of decisions beyond the deadline, and nothing on self-funded employer plans. It also confirms that drugs are not in these metrics.37

Five moves that answer a drug denial: reason, deadline, pages, call, log

1. Read the stated reason against the chart. If the plan says "a trial of 8 weeks", find out which page it read.
2. Check the deadline. ERISA plans must give at least 180 days to appeal.13 Part D denials must describe the standard and expedited redetermination processes; redetermination is Part D's first level of appeal.9
3. Answer that reason and nothing else first. Quote the note, cite the fills, and attach only those pages.
4. Offer the prescriber a peer-to-peer call, a conversation with the plan's clinician, if the plan allows one. The prescriber decides whether to take it.
5. Log the outcome by plan, drug and reason. Over a year that log tells you where requests fail.

How long patients wait for a biologic, and the hours prior authorization takes from a practice

In shortPublished waits for a biologic run from about a week to well over a month, depending on the plan and the setting. Physicians report 40 prior authorizations a week and 13 hours of work. A US count of specialty drug requests does not exist in public data. Measure your own.

Published waits for specialty drugs, by plan and by denial

Those patients, with psoriasis or psoriatic arthritis, came from one academic health system, and the need for prior authorization significantly predicted a longer wait.39 Same-day approval went to 63% of Medicare patients and 11.5% of those in PPOs.39 The records run from 2000 to 2019, so today's numbers may differ.40 It remains one of few published views of biologic waits for this disease.

That study also found higher steroid exposure in the three months after a denial: a median of 605 mg against 160 mg of prednisone equivalent when no authorization was needed.35 A delay here keeps patients on older treatment while they wait. It is a single-center study, and it cannot prove the denial caused the difference.

The American Academy of Dermatology's survey found that dermatologists and staff spent a mean of 3.3 hours a day on PAs, 60% of dermatologists interrupted patient visits for them, and 12% of PAs led to delaying or abandoning treatment.41 A separate tracking study gave a median of 30 minutes a request, with a middle range of 17 to 105 minutes, and a median 12-day delay to treatment.34 That range is wide because easy renewals and hard biologic starts are different jobs.

The size of the work, and what no one publishes

Medicare Advantage insurers made 52.8 million determinations in 2024, about 1.7 per enrollee.3 We found no public figure for US drug prior authorizations, for the cost of one, or for coordinator hours by step. Any number you see for those in a sales deck is an estimate, and you should ask where it came from. You can measure all of them in your own office within a week, as section 18 shows.

What a manufacturer's patient services team may do on a prior authorization, and what stays with the practice

In shortA drug maker's patient support service, called a hub, and its field reimbursement team often offer prior authorization support. The conservative line is clear: the practice owns the request and the prescriber signs it. The program sees what the patient has agreed to share, and any help must carry no reward for prescribing. Get counsel to review the design.

Drug makers run patient support programs, with hub case managers and field reimbursement managers who help offices understand a plan's process. Patient enrollment and benefit checks are separate pieces of work, covered in their own guides. This section is about one question: how far that help may reach into the request itself.

Anti-kickback, HIPAA and false statements: the rules that limit a program's help

• Anti-kickback. The federal health department's Office of Inspector General (OIG) issues guidance on the law against paying for referrals. Its 2003 guidance for drug makers says of reimbursement consultation and similar product support: "Standing alone, services that have no substantial independent value to the purchaser may not implicate the anti-kickback statute." It warns about offering them in tandem with another service that benefits a referring provider, such as a reimbursement guarantee.42 OIG's site lists the 2003 text as archived and shows no publication date for a replacement.43 Take this to counsel, because the analysis depends on the facts.
• HIPAA. Sharing a patient's information with a manufacturer's program generally needs a valid authorization, which names the information, who shares it, who receives it, the purpose and an expiry, and carries the patient's signature.44
• False statements. The kaléo settlement described in section 08 is a warning about anyone other than the prescriber submitting in a prescriber's name.33

A program that keeps to status and general education avoids most of these questions. It can tell an office the status of a case the patient has agreed to share, explain publicly what kinds of evidence plans commonly ask for, and leave the rest to the office. If a program wants to do more, such as pay for software an office uses, that is a question for counsel before anyone signs anything, because a free tool can have independent value.42

Where AI fits between a chart and a PA policy, which rules reach it, and what a clinician keeps

In shortSoftware can read charts, match lines to passages and draft letters, and early studies are promising under tidy conditions. The hard part is the mess: incomplete records and administrative details. Most AI guidance for drug makers does not reach this office work, while HIPAA does. A clinician signs anything clinical.

What the early evidence on AI-drafted PA letters says

The most direct study we found is a preprint dated March 31, 2026, which has had no peer review. Two independent researchers ran three commercial language models across 45 physician-validated, synthetic scenarios in five specialties, 135 letters in all.

Their checks of all 135 letters, with five read line by line, detected no invented clinical facts, and the models averaged above 97% on the authors' rubric. They also found gaps that clinical scoring missed: absent billing codes, missing requests for an approval period and weak follow-up plans.45 The authors warn that all the facts were handed to the models in the prompt, and that the risk is likely higher with incomplete real records.45

Read that as an argument about how such a tool should be built. The drafting step works when the facts are already extracted and tied to pages. Most of the value and risk then lie in the steps before and after it: reading messy records accurately, saying when something is missing, and having a person check.

HIPAA applies to PA preparation, and most AI guidance for drug makers does not

Five ways AI drafting can go wrong on a real PA chart

• Records it never saw. Software can only cite what it was given. "No evidence found" must list where it looked.
• Poor scans and handwriting. Faxes from outside labs misread easily. A reader should show the page beside the value.
• Upgraded wording. A drafter may turn "some improvement" into "inadequate response". Quoting the note protects against this.
• Policy changes. Criteria change. A tool must say which version it used, and a person must approve a new one.
• Administrative detail. As the preprint found, codes, approval periods and form fields need their own checks.45

Who runs prior authorization in a specialty practice: coordinators, prescribers, managers and the pharmacy behind them

In shortA coordinator carries most specialty requests from a chart export, a saved PDF and a phone. The prescriber signs between patients. Help may come from a specialty pharmacy, a central team or a manufacturer's program. Time goes to chasing evidence and following up. No public study breaks this down by step, so measure it in your office.

Six roles a biologic request passes through in a practice

Where a coordinator's minutes go on one hard biologic start

Those minutes are our own. The one tracking study we found gives a 30-minute median and a middle range of 17 to 105 minutes across all kinds of request at one clinic.34 Our 130 sits above that range because a new biologic start is a hard case.

Coordinators, a central pharmacy team or software on top: how offices staff PA requests

One unmet line on day one, 18 weeks of methotrexate recovered, and Jordan Reyes's request signed the next morning

In shortPriya Ranvelle, the practice's PA coordinator, opens Jordan's request against Crestline Health's six-line policy. Five lines come back met and one unmet. A search of the practice's own record finds the missing proof. Six agents do the reading and drafting, and Dr. Maya Chen reviews and signs. In the sample, the plan approves 0.7 days after the prescription.

1

Crestline's six criteria sit beside an empty column

Jordan's chart arrived at 7:52 am: 46 pages in 8 documents. Priya opens his request against Crestline's policy CHP-RX-118, version 2026.10, Commercial. Each criterion shows in the plan's own words, with its verdict and evidence still empty. Across the top are the morning's other requests, each with its count and state, so she can see what else needs her.

On the right the ready-to-submit pack waits, with five empty parts: proven criteria, payer form, letter, cited pages and the prescriber's signature. It reads "nothing goes to the payer before it".

2

Three agents read the chart in view

Priya presses "Match the chart". The Chart Reader reads 46 pages and 8 documents in 38 seconds in the sample, every fact with its page. The Therapy History Builder finds 3 prior therapies and checks their dates against pharmacy fills in 6 seconds. The Criteria Matcher then takes the six lines one at a time in 11 seconds.

Each agent's progress is on the screen, so the coordinator sees what has been read before she trusts what comes back.

3

Five lines met, and the methotrexate step unmet

Body surface area 14%, including scalp and both palms, on the September 22 note, page 24. That meets two of the plan's three severity options. A board-certified dermatologist at the labeled dose. Fifteen weeks of Tenzira, the plan's preferred biologic, with no meaningful response, confirmed by four fills. A negative TB test from October 3.

Line 4 stays red: "No documented trial of methotrexate." The screen lists what was checked: topicals for 8 weeks and light therapy for 12 weeks are documented, and no methotrexate, cyclosporine or acitretin appears in the 46 pages or the specialty pharmacy fills. It offers two next steps: search the practice record, or ask Dr. Chen.

4

Any citation opens the page it came from

Priya clicks the first citation. The original consultation note opens with the passage marked, the page number and the reader's own label. She can check any verdict this way in a few seconds, before she relies on it.

The same click works for the prescriber at signing and for the practice manager at audit.

5

"Search the practice record" finds what the export left out

Priya presses the button on line 4. The agents look through telehealth encounters and outside pharmacy claims for Jordan's record since 2024. They find two telehealth visits with Dr. Chen on January 6 and May 12, 2026, and five methotrexate fills at a retail pharmacy.

They propose the proof and wait: "Methotrexate 15 mg weekly, 18 weeks, inadequate response." Priya can read the May 12 note first. Attaching is her choice.

6

Attached and re-checked: six of six

The chart is now 51 pages in 11 documents. Line 4 turns green at 99% in the sample, with three citations: methotrexate from January 6 to May 12, stopped for inadequate response and a rise in ALT, a liver enzyme, to 2.4 times the upper limit, on pages 47 and 49, and five retail fills on page 51.

The Criteria Matcher only uses proven passages, so the green line rests on the notes it can quote.

7

One click builds the form, the letter and the pages

The Form Filler completes the payer form at 32 of 32 fields and the electronic question set at 14 of 14, each answer cited. The Letter Drafter writes a one-page letter of medical necessity in Dr. Chen's voice, with 13 citations. The pack carries 17 of the 51 chart pages from 9 documents, with the proving passage marked on each.

This is the point where the work stops being Priya's. The next button reads "Send to Dr. Chen to sign and submit".

8

A letter where each clinical statement carries a number

The letter gives the request, the severity with its date, a treatment history table and the reasoning. Each number in the text leads to its chart page. The history table shows four earlier therapies, with dates, length and outcome.

No promotional wording appears, and the drafter claims nothing beyond the chart. If Dr. Chen disagrees with a sentence, she changes it, and the change is on the record.

9

Dr. Chen attests and signs between patients

The signing view shows what she is signing: 6 of 6 criteria, the form, the letter and 9 cited documents. Nothing goes to the plan until she signs. She ticks an attestation that she has reviewed the request, the letter and the attached records, that they are accurate, and that Veltrimab is medically necessary for this patient. She can also send it to her signature folder for later.

This setting is locked on, for every request, in the product.

10

Sent electronically, and the decision lands on the same panel

After the signature, the Status Watcher sends the form, the letter and the 17 cited pages by ePA and watches the clock. In the sample, Crestline records its decision minutes later: approved, CH-26-884213, valid 6 months to April 6, 2027, 0.7 days from prescription.

An approval this fast is the sample's result on a request with nothing missing. Your plans will behave as they behave, and the deadlines in section 04 are what you can hold them to.

Harder requests: an expedited Part D case with an unclear line, a plan's question, and an appeal that answers a misread trial

In shortMost mornings are not a clean six of six. The same screens handle a lab out of date, a statement a person must confirm, a plan asking for one more document, and a denial built on the wrong page. In each, an agent proposes and a named person decides.

Theo Lindqvist: a stale liver panel, and steroid dependence stated without dates

Theo, 67, has Crohn's disease and a Silverpine Medicare Advantage plan, with an expedited Part D request. The rule gives the plan 24 hours once it receives the request, which section 04 cites.10 Six of the 8 lines are met. His liver tests are from April 28, 162 days ago, and the plan allows 90. One click drafts a panel order for his gastroenterologist, Dr. Lena Ortiz, to sign, with "Ask Dr. Ortiz" beside it. A lab order is her decision.

Line 4 is unclear. The note says steroid-dependent and records no taper dates, so the Criteria Matcher's confidence falls to 84%. Below the practice's 90% setting, a person must look, and Dr. Ortiz still reads and signs the whole request at the end. The pack stays locked until both lines are settled.

Samuel Okoro: Larkfield Mutual asks for a date

The plan asks for documentation of a topical calcineurin inhibitor trial of at least 4 weeks. The answer is already in the chart, on page 9. The banner shows the question, the citation and a reply button. The Status Watcher has drafted the reply, and it waits for a person. The practice's clock on this one reads "Reply by Oct 9 to keep it open".

Rafael Duarte: a denial that read the week-8 check-in

Northgate Rx denied Dravenzo on September 30, because "the records show a TNF blocker trial of 8 weeks". TNF blockers are a class of drug designed to neutralize a protein behind inflammation. The Status Watcher compares that reason with the chart: the 8 is from a check-in visit, and the note of September 15 records 14 weeks with inadequate response. Four 28-day fills from June 9 to September 1 agree. The policy asks for 12.

The Letter Drafter writes the first-level appeal in Dr. Omar Haddad's voice. It answers that one reason, with both pages cited, and asks the plan to overturn its denial. It waits for his signature. The list shows the appeal window, 180 days from the denial, ending March 29.

Beyond the criteria screen: the queue, the evidence map, the payer form, plan versions, the dashboard and who signs what

In shortAround the criteria screen sit the request list, a workspace for each request, a list of approved plan policies, a practice dashboard and the settings. Six agents read and draft. People confirm, approve versions and sign.

What each PA agent reads or drafts, and the line it may not cross

Who signs a request, a reply and an appeal at the practice

On a request, the coordinator and the prescriber are separate people with separate roles, so each signature is one person's checked work on another's draft. The controls here support your own HIPAA duties: minimum necessary pages, role-based access and a full audit trail. Botminds is SOC 2 Type II audited, with information security managed to ISO 27001.

From chart notes and plan policies to a signed request: what the solution takes in and hands back

Where PA data travels: the health record, pharmacy fill history and the ePA network

Three design targets: 1.2 days to a decision, 84% first-pass approval, 24 minutes to verdicts

The solution data states three design targets: 1.2 days median from prescription to the plan's decision, 84% of requests approved the first time they are sent, and 24 minutes from chart in to every criterion met, unmet or unclear. They have not been measured at a customer. Part C helps you work out your own starting point.

What a specialty PA denial costs your practice: rework hours, a queue that never clears and patients who wait or walk away

In shortYou will not find a published cost per drug prior authorization. You can measure yours in a week, from your denial letters, your staff's time and your own dates from prescription to first dose. Start with the numbers below for context, and replace them with yours.

From coordinator minutes to patients who never start: finding your PA numbers in a week

For a drug maker's patient services team, the same six rows apply, with one change: you see few of these numbers directly. Ask practices for counts, with their agreement and no patient details.

What a denial log never shows: patients who drift, prescribers who give up, a coordinator who leaves

What PA support software could return on your request volume: hours, fewer denials and days back for patients

In shortValue comes from four places: hours on each request, denials that never happen, fewer days a patient waits, and an audit record that already exists. The calculator counts the first three and is yours to fill in. Every default marked illustrative is an assumption, and the formulas are printed.

Reading time, avoided denials, earlier starts and the audit file: the formula for each

Enter your request count, denial rate and rework hours

Try three sets of inputs. A small office with 300 requests a year and the other defaults has less to gain than the tool costs, so its decision turns on patients and staff time rather than dollars. A practice at the default 1,500 requests shows a modest positive total.

At 5,000 requests the total passes $150,000 with the tool's cost held at the default, so ask what the tool would cost at that volume. Volume is the main driver, and hours saved per request is the second. If your first-pass denial rate is already low, the hours line carries the case.

Can your practice automate PA yet? Score your denial log, time baseline, records and sign-offs

In shortScore each question from 0 to 2. The result says whether to build your baseline first, run a pilot, or plan the rollout. Your answers stay on this page.

Prior authorization vendor questions: proof on your charts, policy versions, the prescriber's signature and privacy

In shortAsk these of every vendor, including us. A good answer shows you the thing on your own files. Most of these cannot be answered well by a slide.

Test PA software for 12 weeks on one drug class and two plans, alongside your coordinators

In shortPick a drug class with enough volume, and two plans you know well. Score the tool first on past requests your best coordinator has checked (a gold set), then run live in parallel, and finally let the tool's work reach the prescriber. Decide on your measured results.

Six numbers to agree before the prior authorization pilot starts

Answers on CMS-0057-F, drug deadlines, step therapy, appeals and the letter of medical necessity

In shortShort answers, each one complete by itself, for coordinators, practice managers and patient services teams.