First-pass PA denial rates in specialty medicine — orthopedics, cardiology, rheumatology, neurology — consistently run higher than in primary care. Industry benchmarks suggest 25 to 30% of specialty PA submissions are denied on the first submission attempt, compared with closer to 12 to 15% in primary care. The difference is usually framed as payer stringency: specialty procedures and medications are expensive, so payers scrutinize them more carefully.
That framing is partially right but misses something important. Payer scrutiny is a constant. What varies between a 12% denial rate and a 28% denial rate isn't how hard payers look at the requests — it's how complete and correctly formatted the documentation is when it arrives. The majority of first-pass denials in specialty medicine are not clinical denials. They're administrative denials. Missing attachments, incomplete clinical narrative, CPT-to-ICD-10 mismatches, and missing step therapy documentation account for the bulk of first-round rejections.
The Most Common Denial Reason Codes in Specialty PA
Payer denial reason codes are standardized through X12 transaction sets (the 835/277CA remittance advice codes), but most payers layer their own clinical criteria denial language on top of the base codes. The patterns that appear most frequently in specialty PA denials break into a few consistent categories.
Incomplete clinical documentation. The submission references a clinical note as support for medical necessity, but the note itself doesn't contain the specific clinical findings the payer's criteria require. For a spinal surgery request, the payer's criteria may require documentation of six weeks of conservative management including physical therapy, specific pain scores using a validated instrument (like the VAS or ODI), and imaging results. A note that says "patient has failed conservative treatment" without quantifying any of those elements will be denied, often with denial code CO-252 (additional information needed) or CO-197 (precertification/authorization absent).
Step therapy failure not documented. Step therapy requirements — also called fail-first protocols — are among the most common sources of specialty denial, particularly in rheumatology and neurology. A new biologic request for a rheumatoid arthritis patient requires documentation that the patient tried and failed at least two conventional DMARDs (typically methotrexate and one other) at adequate doses for adequate durations. If the prior therapy documentation isn't extracted from the chart and explicitly stated in the submission, the payer will deny for insufficient step therapy evidence. The clinical reality may be perfectly clear in the physician's mind and even recorded somewhere in the EHR. What matters is whether it's in the submission package in a form the payer's reviewer can locate and verify.
ICD-10 and CPT code misalignment. Payers maintain internal coverage criteria that map specific ICD-10 diagnosis codes to approved CPT procedure codes. If the diagnosis code on the submission doesn't match the payer's expected code for the requested procedure, the submission may be automatically flagged or denied before clinical review. This isn't always obvious: M05.79 (rheumatoid arthritis with rheumatoid factor, other sites) and M06.09 (rheumatoid arthritis without rheumatoid factor, other sites) may both describe the same patient's condition, but one may map cleanly to the payer's biologic therapy criteria and the other may not.
Missing or outdated clinical attachments. Lab values, imaging reports, functional status assessments — many specialty PA requests require attachments that exist in the EHR but need to be explicitly pulled and attached to the submission. A neurology request for a high-cost multiple sclerosis therapy may require a recent MRI report showing active lesions, a current EDSS score, and documentation of at least one relapse in the prior 12 months. If the coordinator submits the request without the imaging report because they didn't realize it was required for this payer's criteria for this specific drug, the denial is administrative, not clinical. The patient medically qualifies. The paperwork doesn't.
Why Specialty Practices Are More Exposed
Primary care PA requests tend to cluster around high-frequency, lower-cost medications and straightforward referrals. The documentation requirements are relatively consistent across payers, and primary care teams that process the same requests repeatedly develop institutional knowledge of what each payer wants. The repetition creates informal standardization.
Specialty practices face a different problem. The range of requested procedures and medications is broader, the payer criteria are more complex and vary more significantly across payers, and the consequences of a missing data element are higher. A rheumatology group submitting PA requests across adalimumab, upadacitinib, tocilizumab, and belimumab — each for different diagnosis indications, across patients covered by four or five different payers — is maintaining a matrix of payer-by-drug-by-indication criteria that no person can hold in their head consistently.
The coordinator who handles PA at a busy rheumatology group typically manages 30 to 60 active prior auth cases at any time. They know the general pattern for each major payer. But when a specific payer updates its criteria for a specific drug — as Cigna, for instance, periodically updates its clinical policy bulletins for biologic therapies — that update doesn't automatically appear in the coordinator's workflow. It appears in a PDF on the payer's website. The coordinator finds out about the new requirement when the first submission under the new criteria gets denied.
The Documentation Assembly Problem
Consider a 12-provider rheumatology group in the Northeast processing approximately 50 PA requests per week across their patient population. Their case mix includes a significant volume of biologic therapy requests, some complex DME authorizations, and a steady flow of infusion therapy PAs. Their average first-pass denial rate, before any workflow changes, is around 27%.
When they categorize their denials by reason code, approximately 60% are administrative — missing documentation, step therapy gaps, ICD-10 alignment issues. Only 40% involve a clinical dispute where the payer's reviewer is actually questioning the appropriateness of care. The clinical disputes, they can't fully prevent. The administrative denials, in principle, are avoidable.
The root cause for most of the administrative denials in a scenario like this is the same: documentation assembly is a manual process that relies on coordinators knowing which data elements to include for each payer-drug-indication combination. When that knowledge is incomplete or when a criteria update hasn't propagated, gaps appear. A system that maintains an updatable map of payer criteria — what clinical evidence each payer requires for each CPT/ICD-10 combination — and pre-checks submissions against that map before they go out is what separates 27% first-pass denials from something closer to 12 to 15%.
Peer-to-Peer Review: When Denials Do Go Clinical
For the subset of denials that are genuinely clinical — where the payer's medical director believes the requested service isn't medically necessary based on the clinical presentation — peer-to-peer review is the standard recourse. A physician at the practice calls the payer's medical director, presents the case directly, and makes the argument for authorization.
Peer-to-peer review succeeds roughly 60 to 75% of the time when pursued. The problem is that it requires physician time. A 20-minute peer-to-peer call takes a physician away from clinical time, consumes bandwidth that could be spent seeing patients, and is generally scheduled at the payer's convenience (usually 24 to 48 hours after the physician requests it), adding further days to the authorization timeline.
We are not suggesting that automation eliminates the need for peer-to-peer review. Some clinical disputes are genuine, and having a physician advocate directly for their patient's care is appropriate and often necessary. What automation does is reduce the number of cases that end up in peer-to-peer because of avoidable administrative deficiencies — keeping peer-to-peer review for cases where it actually involves a clinical judgment call, rather than using it to rescue submissions that should never have been denied in the first place.
Gold-Carding: The Long Game Worth Playing
Several states have implemented gold-carding provisions as part of prior auth reform legislation. Gold-carding exempts physicians with a track record of approvals from the prior auth requirement for specific services or drugs — essentially, if your practice has a 95%+ approval rate for a particular biologic therapy over 12 months, you may be exempt from submitting PA requests for that drug from that payer going forward.
This is a meaningful long-term benefit, but it depends on building the approval track record in the first place. A practice with a 27% first-pass denial rate doesn't build a gold-card-qualifying track record. A practice that improves to 90%+ first-pass approvals on its core specialty medications — through documentation quality and criteria compliance — starts accumulating the approval history that gold-card provisions recognize.
The denominator for gold-card qualification is total submitted requests, not just approved requests. Every preventable denial that goes through appeal counts against the track record. Getting documentation right on first submission isn't just about turnaround time and staff hours — it's about building toward a future where the PA requirement itself becomes less of a burden for the practice's highest-volume services.
Where to Start
If you're trying to reduce your specialty practice's first-pass denial rate, the highest-leverage starting point is a denial audit by reason code. Pull 90 days of PA denials, categorize them by the payer's stated denial reason, and separate administrative denials from clinical denials. For most specialty practices, that audit reveals 50 to 65% of denials as addressable through documentation process changes rather than clinical appeals.
Once you know which specific documentation gaps are driving denials — missing step therapy notes, ICD-10 mismatches, absent imaging reports — you can build targeted process improvements or automation to address each gap. That kind of targeted fix, starting from actual denial data, is more effective than general documentation improvement initiatives that don't trace back to specific payer requirements.