Denials remain a top (and pressing) concern for revenue cycle leaders. But as our team talks directly with revenue cycle leaders, attends events and digs into client data, we hear more and more concerns about revenue loss that doesn’t technically count as or get categorized as “denials.” Post-payment takebacks (or, the euphemism, “recoupments”) are pulling earned revenue out of periods health system finance teams have already closed. DRG downgrades are landing on accounts that hit zero balance a year ago. Underpayments are posting as routine adjustments. Revenue cycle leaders can feel the missing revenue, but their denial reports don’t show it.
Denials haven’t gone anywhere. They’re still the biggest line most revenue cycle teams track. But the share of reimbursement loss that never touches a denial code is growing, and we’re getting asked most about payer takebacks and post-payment audits.
Prior auth is getting easier. The money has to come from somewhere.
The front end is under real pressure. Payers that took the 2025 pledge say they’ve pared back prior authorizations, reporting an 11% drop in volume, and Jan. 1, 2027 is the go-date for the API requirements established by the Interoperability and Prior Authorization final rule, which the industry is racing toward. Provider teams fought for this, and they should hold payers to it. But, it doesn’t change the economics on the payer side. Payers have medical loss ratios to manage and margins to protect. When one lever gets regulated and publicly scored, they just shift to a different lever. Right now, those are the ones that act after payment, take months to surface, and don’t carry a denial code. Across the claims Sift has analyzed, these five patterns sit outside standard denial reporting.
- Post-payment takebacks arrive long after the period closes. When a claim is subject to a takeback, the payer eventually recoups about 40% of what it originally paid. Roughly 2% of all paid dollars come back out this way. The median takeback lands about 170 days after first adjudication, and 90% of takeback dollars take around 550 days. By then, it shows up as a change to a period you already reported, and last quarter’s net collection rate revises downward after you’ve shared it.
- DRG downgrades take close to two years to finish. The payer strikes the secondary diagnosis holding the severity tier and reprices the claim one tier lower. Half of a cohort’s downgrade activity lands past the one-year mark. The disputed diagnosis comes in a letter, not on the 835, so it usually ends up in a spreadsheet somewhere (??!!).
- Some underpayments post as paid. Aetna Medicare is paying some inpatient claims at observation-level rates, signaled by RARC N610 paired with CARC 45. That pairing doesn’t create a denial record in most workflows. Across the low-severity Aetna cases we’ve analyzed, the gap between expected and actual reimbursement has run about $6,000 per case.
- Bundling codes cover real shortfalls. CARC 234, 96, and 55 get used for legitimate bundling and for reimbursement that simply never arrives. At one health system, 9% of outpatient claims in 2025 carried CARC 234 or 96. Most teams route these out of denial logic, so the shortfalls stop being counted.
- Itemized bill requests are a payment delay tactic. They make up under 0.05% of claims and up to 10% of denied dollars in a month. Most get overturned and paid close to in full, but the payer holds the cash the whole time, and the carrying cost never appears in a report.
Denial reporting misses post-payment takebacks
Denial reporting counts events that carry a denial code on the date they post. These losses surface at three different points: 1) pre-bill as concurrent review letters and portal notices, 2) at adjudication as remit codes that post as paid and 3) post-bill as post-payment takebacks and downgrades, sometimes two years later. Each lands in a different system owned by a different team, and the denial report only sees the slice that looks like a denial.
That’s why revenue cycle leaders feel the gap before they can measure it. Expected reimbursement and what the organization keeps drift apart while the denial line on the dashboard holds steady. You can measure it in your own data. Take a discharge cohort at least 18 months old and compare total expected reimbursement against what you kept after recoupments. Then compare that gap to the denied dollars your denial report shows for the same cohort. The difference is the part of your loss living outside denial reporting.
Tracking the full scope of adverse payment outcomes
Sift’s RevProtect Payments Intelligence tracks every payer action that leaves reimbursement below the contracted amount, and we call that full set adverse payment outcomes. Denials are one of them. Post-payment takebacks, DRG downgrades, and underpayments coded as adjustments are others. Tracking them takes payment intelligence that follows each claim through its whole payment life, by payer, long after the denial report has closed it out. That longitudinal record also lets RevProtect predict which cases are likely to trigger a specific payer’s adverse action while there’s still time to fix the documentation or the claim (pre-bill denials prevention).
To see what Sift tracks, our new Denial Deep Dives guide covers different adverse payment outcomes and includes a query for each pattern. And if you want to learn more about how RevProtect can work for your health system, let us know.
