Payer-Specific Denial Patterns at UnitedHealthcare, Cigna, and Aetna

Denial rates at the three largest commercial insurers in the country are not random noise. UnitedHealthcare, Cigna, and Aetna each run a distinct, repeatable denial logic, and the practices billing them can learn what that logic looks like before a claim ever gets submitted. This is the whole argument of this piece: pattern recognition beats appeal-writing, because by the time you're writing an appeal, you've already lost weeks and, in UnitedHealthcare's case, maybe the whole claim.

The backdrop makes this more than an academic exercise. The industry-wide initial claim denial rate hit 11.8% in 2024, up from 10.2% just a few years earlier, and it's still climbing. Medicare Advantage denials rose 4.8% from 2023 to 2024, while commercial plan denials grew only 1.5% over the same stretch: two different trend lines, two different sets of causes, two different fixes. Net revenue lost to denials grew 25% year over year, from $38.6 billion in 2024 to $48.4 billion in 2025, which tells you resolution complexity is growing faster than the raw number of denials. On top of a rising denial count overall, practices are facing denials that take longer and cost more to untangle once they land.

Most billing offices still find out about a denial spike at month-end, well after the appeal clock has run down. That lag is the real gap. What follows is a look at how three payers each generate their denials, and what a practice needs to do differently for each one.

How UnitedHealthcare's nH Predict algorithm reshaped its post-acute and in-office drug denial behavior

UnitedHealthcare's ACA marketplace denial rate dropped from 33% in 2023 to 20% in 2024, the sharpest year-over-year drop among major insurers. That number looks like good news until you notice it only covers a slice of UHC's total book. Employer plan denials are tracked separately, so the marketplace figure alone doesn't tell the whole story.

The more consequential shift happened in post-acute care. UHC rolled out the NaviHealth nH Predict algorithm to drive coverage decisions for skilled nursing and rehab stays, and the Senate Permanent Subcommittee on Investigations, in an October 2024 report, found that UHC's post-acute denial rate more than doubled after the tool went live. Plaintiffs in litigation over the tool allege it carries a 90% error rate, and that more than 80% of the prior authorization denials it generated were reversed on appeal. That reversal figure matters more than the error-rate allegation itself: if four out of five denials get overturned once a human looks at them, the denial was never really a clinical judgment. It was a flag thrown by a model, and flags can be beaten if you fight them.

In-office drug administration is the other soft spot. UHC's denial rate there sat at 15.32% in 2024, well above the roughly 10.75% average among its peers, and it only eased to 13.11% in 2025, still the highest rate among the payers tracked. UHC did commit, under public pressure, to cutting about 10% of its prior authorization requirements in 2025, but that concession is concentrated in low-volume PA categories and won't move the needle on post-acute or drug administration denials.

For a practice billing post-acute care, skilled nursing, or in-office infusion and injectable drugs to UHC, the risk is concentrated in these specific settings, and it's driven by an algorithmic flag rather than a clinician actually reading the chart. That changes what "good documentation" means. Medical necessity notes need to anticipate the kind of pattern-matching a model does, not just satisfy the clinical standard a human reviewer would accept.

UHC's 65-day appeal window and why it changes the math on every denial

UnitedHealthcare gives practices 65 days to appeal a commercial denial. Aetna, Cigna, and Blue Cross Blue Shield plans all allow 180 days. That gap is nearly three times the runway competitors give you, and UHC is the shortest in the industry by a wide margin.

Miss the 65-day window and the appeal right is gone, full stop, no matter how strong the clinical case would have been. Across the industry, fewer than 1% of denied medical claims ever get appealed, even though appeal success rates when pursued can be substantial. Most of that recoverable revenue is never chased; it's simply written off. Layer UHC's short clock on top of that pattern and you get a payer where the cost of slow triage is higher than almost anywhere else in the business.

There's also a legal thread worth watching, though it doesn't change what a billing office needs to do this week. In February 2025, a federal court allowed breach of contract and good faith claims to proceed in Estate of Gene B. Lokken v. UnitedHealth Group, a signal that UHC's denial processes remain under real legal scrutiny. Court rulings don't put money back in a practice's account, though. Only a workflow that catches a UHC denial the same week it lands, queues it, and files before day 65 does that.

Cigna's PxDx system and the pattern of high-volume, low-review denials

Cigna's denial engine runs on volume and speed rather than algorithmic prediction about future care needs, and the numbers are startling on their face. ProPublica's investigation found that Cigna's PxDx system denied 300,000 claim payment requests over a two-month span in 2022, with an average review time of 1.2 seconds per denial. One medical director, under this system, denied 60,000 claims in a single month. There is no meaningful chart review happening at that pace; there's a rules engine matching codes.

Cigna's own position is that PxDx applies only to a defined set of lower-cost tests and procedures, that the logic mirrors what other insurers and CMS already use, and that it isn't AI-driven. Whatever the label, a federal district court allowed a class action against Cigna to proceed in March 2025, with plaintiffs arguing that PxDx violated ERISA fiduciary duties by substituting an automated rule for the individualized medical necessity review the plan documents promised.

The mechanism behind most of these denials is a diagnosis-code mismatch, and it's a fixable problem once you understand it. PxDx checks procedure codes against diagnosis codes at the population level; if the pairing doesn't match what Cigna's system expects to see, the claim gets flagged as non-covered automatically. A test that is entirely appropriate for the specific patient in front of the clinician can still get denied, because the code combination doesn't match the population-level pattern Cigna's algorithm was built around. The denial usually has a specific, identifiable cause sitting in the code pairing, and once you find it, you can either correct the submission or build the appeal around exactly that mismatch.

Practices doing high-volume diagnostic or outpatient procedure work with Cigna should be checking their most common procedure-diagnosis pairings against Cigna's covered combinations before they submit, not after the denial shows up. Cigna's system is built for an environment where almost nobody pushes back. A figure cited in the litigation, drawn from KFF data, put the industry-wide appeal rate at just 0.2% of denied claims in 2021. A denial engine calibrated against that kind of silence doesn't need to be right very often. It just needs to be fast.

Aetna's prior authorization intensity and what its unusually high appeal overturn rate signals

Aetna's Medicare Advantage prior authorization denial rate runs at 11.6%, the second-highest among major MA insurers according to muni.health. Its post-acute care denial rate is worse: 25.9%, the highest of the three payers examined in the Senate PSI's October 2024 report, and roughly one and a half times the 7.7% industry average. Post-acute providers and hospital-based billers carry a disproportionate share of Aetna's denial burden, and that's not a coincidence; it's where the payer applies the most pressure.

The AMA's 2024 prior authorization survey of physicians, a randomly sampled group of 1,000 physicians released in February 2025, rated Aetna as a "high" or "extremely high" burden payer on prior authorization, more intensive on average than Cigna or Blue Cross Blue Shield, though still lighter in raw PA volume than UHC.

Here's where Aetna's profile gets genuinely strange, and genuinely useful for a billing team: 92.6% of Aetna's Medicare Advantage denials get overturned on appeal. That's the single most important number in this entire dataset. When Aetna's own appeals reviewers agree with the provider more than nine times out of ten, the initial denial was never a clinically defensible decision. It was a first-pass gate, and it was wrong almost every time someone actually pushed on it.

The gap between that overturn rate and what practices actually do is enormous. The industry-wide appeal filing rate for denied claims remains vanishingly low, even though win rates once an appeal is filed can be substantial. Aetna is also financially significant enough to justify the effort: commercial HMO reimbursement from Aetna averages 216% of Medicare, above most competitors. The data suggests every Aetna denial should be treated as presumptively recoverable rather than written off.

The prior authorization burden that sits behind all three payers' denial patterns

Underneath all three payer profiles sits the same structural weight: prior authorization has become a second full-time job for a huge share of medical practices. The AMA's 2024 survey put the average physician workload at 39 prior authorization requests per week, consuming 13 staff hours per physician per week. Forty percent of practices now have staff whose entire job is handling PA requests. MGMA's analysis found that practice spending on PA staffing jumped 43% between 2019 and 2024, a substantial rise over that period.

Scale is the other half of the story. Medicare Advantage plans alone process tens of millions of prior authorization determinations each year. No billing office reviews that volume claim by claim without some kind of systematic tooling; the math simply doesn't work by hand.

The clinical stakes are real, not just financial. Ninety-three percent of physicians in the AMA survey said PA causes care delays, 82% said it leads patients to abandon treatment altogether, and one in four physicians reported a serious adverse event tied to a prior authorization decision. The administrative cost of a single denied claim now averages $25, while AI-driven claim scrubbing can meaningfully improve first-pass acceptance rates. That gap, between a $25 denial and a 95% clean claim rate, is where most practices are quietly bleeding money.

The deeper problem isn't that prior authorization logic is hard to predict in the abstract. Each payer runs a different logic, updates it without warning, and practices relying on static training manuals or annual refreshers can't keep pace with rules that compound and shift underneath them month to month.

What payer-specific denial intelligence requires operationally to be useful in real time

Three payers, three different intervention points. UHC calls for pre-submission medical necessity documentation built to survive algorithmic review, plus same-day denial triage, because a 65-day window closes faster than most billing cycles run. Cigna calls for procedure-diagnosis code pairing checks before the claim ever goes out, since PxDx flags are a code-matching problem, not a clinical one. Aetna calls for an appeals workflow that treats every initial denial as recoverable by default, given a 92.6% overturn rate on MA claims.

None of that works if the knowledge lives in one person's head. Payer rules change without notice, and staff turnover means the institutional memory of "how Aetna denies things" or "what triggers a Cigna PxDx flag" walks out the door with whoever built it. That knowledge needs to sit in a system rather than a person, or the practice starts over every time someone leaves.

The real-time piece is not optional, either. A denial rate discovered at month-end is already a missed appeal window if the claim was denied by UHC in the first week of that month. Speed of detection is the whole game for at least one of these three payers.

External audit pressure is rising alongside all of this. Recent industry analyses have found notable year-over-year increases in at-risk audit amounts, with a growing share of that exposure tied to commercial payers. Denial management and audit exposure are converging into the same problem viewed from two angles.

What actually works, in practice, looks like this: claim scrubbing tuned to each payer's specific code-pairing and PA logic before submission, denial alerts that fire in real time and trigger an appeal workflow immediately rather than at the next billing cycle, and appeals that get filed rather than shelved, since a strong internal success rate only helps the practices that actually submit the paperwork. Add human review for the clinically complex exceptions, where a biller who understands a specific payer's behavior can catch what an automated flag misses, and integration into the EMR that doesn't force a front-desk retraining project just to get started.

The practices holding their ground against UHC, Cigna, and Aetna aren't necessarily the ones with the biggest billing departments. They're the ones that respond fastest, in a way that's specific to each payer's own behavior, rather than running one generic denial process against three very different opponents.