Orthopaedic Payer Underpayment Recovery in 2026: Finding the Revenue That Was Never Denied

An underpayment is a claim the payer adjudicated and paid, just at less than your contracted rate. It posts as a normal payment with a contractual obligation adjustment, usually a CO group code such as CO-45, and it closes the account. Nothing lands in a work queue. Nobody appeals it. It looks exactly like a clean paid claim on every report you run

That is the entire problem: denials are visible by default and underpayments are invisible unless you independently calculate what each claim should have paid and compare. Orthopaedics is the specialty where this hurts most, because your highest value claims are the ones subject to multiplicative payment adjustments, and every one of those adjustments is a place the payer's logic can silently diverge from your contract.

If your group has a denials committee and no expected reimbursement model, you are working the smaller of the two problems.

Denials and underpayments are not the same problem

A denial is a refusal to pay. It arrives as a zero pay with a claim adjustment reason code, generates a work item, and gets worked or written off. Someone owns it.

An underpayment is a payment. The payer accepted the claim, applied its own adjudication logic, and remitted an amount. If that amount is below the contracted allowable, the difference posts as a contractual adjustment, which your system treats as expected and correct by definition. The claim adjustment reason code set is maintained by X12, and CO-45 in particular indicates charges exceeding the fee schedule or maximum allowable. Whether that adjustment was right depends entirely on a number your system does not know unless you told it.

This asymmetry is why net collection rate is the detector of last resort. MGMA defines the metric as amount collected divided by expected collections based on contractual allowables, and frames the revenue opportunity as the gap between your actual insurance net collection rate and 98.5 percent. Every point below 98.5 percent is roughly one percent of contracted revenue lost to some mix of denials, underpayments and write offs. We could not verify an orthopaedic specific net collection rate benchmark from MGMA; the specialty breakouts sit behind DataDive, and the orthopaedic figures circulating in billing vendor content are not traceable to a methodology.

How big is this actually

There is no methodologically sound published estimate of what percentage of commercial claims are paid below the contracted rate for physician practices. Every source that claims one traces back to marketing material without disclosed methodology. Treat any such figure you are shown as unsourced until proven otherwise.

The closest historical proxy is the AMA's National Health Insurer Report Card, which measured whether payers' allowed amounts matched contracted rates. It reported a claims processing error rate of 19.3 percent in 2011, 9.5 percent in 2012 and 7.1 percent in 2013, and put the addressable prize at roughly $17 billion annually. The last edition we can locate is 2013 and no current edition exists, so treat it as discontinued and the figures as thirteen years stale.

For a directional dollar frame, a widely cited industry analysis estimates that 1 to 3 percent of provider net revenue is lost annually to commercial payer underpayment. That figure is vendor sourced and hospital oriented, so apply it as a hypothesis to test against your own data rather than as a benchmark. The more useful finding in the same analysis: of more than $200 million in underpayments one firm recovered in a year, 70 percent were identified outside the standard variance report. Variance reporting alone is documented as insufficient.

What is well established is how rarely practices look. In an MGMA poll, 26 percent of practice leaders said they audit payments against contracted rates monthly, 25 percent annually, 15 percent quarterly, 10 percent semiannually, and 24 percent never. A separate poll of 389 leaders found 58 percent review payer contracts annually and 17 percent do not review them regularly at all.

The five orthopaedic underpayment patterns worth modeling first

Orthopaedics concentrates underpayment risk because expected reimbursement is not a fee schedule lookup. It is the fee schedule after multiple procedure ranking, bilateral adjustment and assistant surgeon percentage. A model that loads the base schedule without replicating the payer's reduction logic will generate false variance on every multi procedure surgical claim and miss the real ones.

‍Multiple procedure payment reduction. MPPR is driven by the MULT PROC indicator on the CMS Physician Fee Schedule Relative Value File. Indicator 2, the standard surgical rule, pays ranked procedures at 100 percent, then 50 percent for each subsequent procedure, with additional procedures by report. Indicator 1 applies the older 100, 50, 25, 25, 25 pattern. Indicator 3 covers special endoscopy rules within a code family. Indicator 0 means no MPPR applies and indicator 9 means the concept does not apply.

Three failure modes to test for. MPPR applied to a code carrying indicator 0 or 9. Incorrect ranking, since procedures must be ranked by RVU with the highest first, and ranking by charge or by claim line order underpays the primary. And use of the wrong RVU set for the place of service. UnitedHealthcare's own Medicare Advantage policy specifies Facility Total RVUs for facility place of service codes including 19, 21, 22, 23, 24, 26, 31, 34, 41, 42, 51, 52, 53, 56 and 61, and Non Facility RVUs otherwise, with the highest RVU ranked first. The same policy confirms that assistant surgeon services are ranked separately from the primary surgeon's services, which is a separate ranking stack and a frequent source of error.

‍Bilateral procedures and modifier 50. Governed by the BILAT SURG indicator. Indicator 1 means the 150 percent adjustment applies, with payment at the lower of billed charges or 150 percent of the fee schedule allowed amount, billed on a single line with modifier 50 and a quantity of one. Indicator 0 or 9 means the 150 percent adjustment does not apply and modifier 50 should not be used. Indicator 2 means the RVUs already include the bilateral component, so both sides are paid at 100 percent without modifier 50. Indicator 3 pays 100 percent per side, typically radiology. ASCs cannot submit modifier 50 and use the units field instead.

The failures here are specific and common: the payer pays 100 percent instead of 150 percent on an indicator 1 code; the payer splits the line into two and then applies MPPR to the second side, producing a double reduction; or the billing format does not match what the payer's bilateral logic expects, so the logic never fires. Note that bilateral and MPPR interact, and UnitedHealthcare's policy states bilateral eligible services may also be subject to multiple procedure reductions.

‍Assistant at surgery. The Medicare Claims Processing Manual, Chapter 12, section 20.4.3, sets the fee schedule amount for a physician assistant at surgery at 16 percent of the amount otherwise applicable. A non physician practitioner billing modifier AS is paid 85 percent of that 16 percent, which is 13.6 percent of the fee schedule amount. Modifier 80 is full assistance by a physician, 81 minimal assistance by a physician, 82 physician assistance in a teaching facility where a qualified resident is unavailable, and AS a non physician practitioner. Eligibility is gated by the ASST SURG indicator, where 0 means the restriction applies unless medical necessity is documented, 1 means a statutory restriction with no payment, 2 means no restriction, and 9 means not applicable.

Watch for the payer applying the 13.6 percent rate to a physician billed with 80 or 82, denying outright on an indicator 2 code, or running the assistant's line through the primary surgeon's MPPR ranking stack.

‍Implant carve outs. These are structurally the hardest category, and it is worth being direct about why. A carve out ties expected reimbursement to an invoice value that lives outside the claim, so a standard fee schedule variance model cannot detect a carve out underpayment at all. Recovery requires a separate reconciliation against implant logs and invoices. We could not verify typical invoice plus percentages or documented failure rates from an authoritative source, so treat any specific number you are quoted as unverified and negotiate from your own cost data.

‍Unlisted codes. Codes such as 27599 for unlisted femur or knee procedures and 27299 for unlisted pelvis or hip procedures carry no RVUs and no fee schedule amount, so there is no contracted rate against which to compute variance. Payment is negotiated per claim. AAPC's guidance is to never substitute a close code, submit a cover letter with the operative note explaining the code choice, suggest a fee by comparing the procedure to a similar procedure with an established code and RVU value, and match the comparison code to the payer's own fee schedule vintage, which matters especially for workers' compensation carriers running older schedules. Practically, unlisted code claims are invisible to variance reporting and belong in a manual queue with a target reimbursement recorded at submission.

The clocks that decide whether you can still collect

Three separate deadlines get conflated constantly, and the distinction is the difference between a recoverable underpayment and a write off.

Timely filing is the deadline to submit the original claim, commonly 90 to 180 days from date of service and up to twelve months for some plans. Reconsideration is the deadline to ask a payer to re adjudicate a claim it already processed, and this is the clock that governs underpayments. It runs from the remittance date, not the date of service, and it is usually much shorter than timely filing. Formal appeal is the deadline to escalate after reconsideration is denied.

Published windows vary more than people assume. Aetna allows 180 calendar days from the initial claim decision to file a reconsideration, then 60 calendar days from the reconsideration decision to appeal. UnitedHealthcare publishes twelve months to complete reconsideration and appeal, while noting that timelines vary and directing providers to their participation agreement. The agreement overrides the public web page in both cases. Verify per contract rather than applying a single rule.

For self funded ERISA plans, the federal claims procedure regulation at 29 CFR 2560.503-1 gives 30 days for a post service claim decision, at least 180 days for the claimant to appeal an adverse determination, and 30 days per level for a post service appeal decision. The appeal right belongs to the claimant, so a practice invokes it through a valid assignment of benefits or authorized representative designation. Exhausting the plan's internal appeal process is the predicate to an ERISA section 502(a)(1)(B) suit, which means missing the 180 days can forfeit the federal cause of action entirely, not just the appeal.

Prompt pay interest, and the California change that took effect this year

State prompt pay statutes apply to state regulated insurance, meaning fully insured commercial plans and in most states Medicaid managed care. They set a clean claim payment clock, require a written contest or denial notice inside the same window, and impose automatic statutory interest once the clock expires. Interest generally accrues without the provider requesting it, which makes prompt pay interest one of the most routinely uncollected revenue lines in a practice. Self funded ERISA plans are exempt, so this leverage disappears on a meaningful share of a typical orthopaedic group's commercial book.

Three examples worth knowing.

‍California changed on January 1, 2026. AB 3275, signed September 27, 2024, converts the deadline from working days to 30 calendar days. Previously insurers had 30 working days and HMOs had 45 working days, which in practice ran to roughly 63 calendar days. The written contest or denial notice is also now due within 30 calendar days of receipt. Interest accrues at 15 percent per annum from the first day after the 30 day period, and failure to automatically pay accrued interest now costs the greater of $15 or 10 percent of the accrued interest, up from a flat $10. The statute is California Health and Safety Code section 1371. If you have California volume and have not reset your aging triggers, you are sitting on interest you are entitled to.

‍Texas has the strongest penalty regime. Thirty days for electronic claims and 45 for paper, with the clock starting the day after receipt. For payment 1 to 45 days late, the penalty is the lesser of billed charges minus the contracted rate, or 18 percent of the contracted amount annually. Critically, penalties are calculated per claim and cannot be waived by contract, which matters because payer contracts routinely attempt to waive them. See Texas Insurance Code sections 843.338, 1301.103, 843.342 and 1301.137.

‍New York sets 30 days for claims transmitted electronically and 45 days for paper or fax, under Insurance Law section 3224-a, with interest at the greater of the Tax Law section 1096 corporate rate or 12 percent per annum. No interest is owed where the amount due is under $2, so small dollar interest has to be pursued in aggregate. New York has no statutory definition of clean claim, which is a real practical weakness when disputing whether the clock ever started.

What gets written off, and why underpayments are worse

HFMA, citing Change Healthcare's analysis of $3 trillion in submitted claims, reports $262 billion denied, averaging nearly $5 million per provider, with up to 65 percent of denied claims never resubmitted. Ninety percent of denials are preventable and two thirds of preventable denials are appealable. Organizations with effective denial programs appeal 85 to 88 percent of their denials. MGMA, citing the same underlying index, reports 86 percent of denials as potentially avoidable while 24 percent of all denials are not recoverable.

No published source quantifies the write off rate for underpayments as opposed to denials. Given that underpayments never generate a work item in the first place, the realistic assumption is that the figure is worse than 65 percent, but that is inference rather than data and we are not going to present it as a statistic.

On the economics of pursuit, MGMA's $25.20 average cost to rework a claim remains the standard physician practice figure, sourced to a 2020 denials index. Do not substitute Premier's hospital figure of $57.23 per denied claim from CY2023 data; the settings and cost structures are not comparable.

Building the capability, in order

Load expected reimbursement, not just the fee schedule. HFMA's framework for measuring payer performance turns on the accuracy of expected payment calculations, and stresses that the modeling parameters used in contract negotiation must remain identical to those used in post adjudication audit, or account level variance is meaningless. For orthopaedics that means replicating MPPR ranking, bilateral logic and assistant surgeon percentages, not just loading a conversion factor.

‍Separate the queues. Route soft denials to billing staff, hard denials to clinical teams, and unresolved contract variance to someone who owns contract compliance. HFMA's guidance is hierarchical for a reason: underpayment work requires contract knowledge that a claims follow up team does not have and should not be expected to acquire.

‍Track prompt pay interest as its own line. It accrues automatically, it is rarely paid automatically, and in California and Texas the statutory rates are high enough to matter.

‍Carve out the categories your model cannot see. Implant carve outs and unlisted codes need manual queues with target reimbursement recorded at submission, because no variance engine will flag them.

‍Close the loop upstream. Every underpayment pattern above is detectable at the moment of coding, from the same information used to build the claim: which procedure ranks first by RVU, whether the bilateral indicator supports modifier 50 and in what format, whether the assistant modifier matches the practitioner type and the ASST SURG indicator, and whether the place of service selects the right RVU set. Claims built correctly with the payer's own logic in view do not generate variance to chase.

That last point is where an automated coding layer earns its place in this workflow. Maia's AutoCoder populates codes, modifiers and clinical justification inside the EHR before a coder opens the chart, applying the indicator level rules that determine what the payer should pay, and Maia's collections and contracting work is built on the same expected reimbursement logic. The goal is not a better variance report. It is fewer claims that need one.

Frequently asked questions

What is the difference between a denied claim and an underpaid claim?‍

A denial is a refusal to pay that arrives as a zero pay with a reason code and creates a work item. An underpayment is a claim the payer processed and paid, but below the contracted allowable, with the shortfall posting as a contractual adjustment such as CO-45. Denials are visible by default; underpayments look like clean paid claims and close the account unless you independently model expected reimbursement.

‍How do I know if we have an underpayment problem?‍

Start with net collection rate. MGMA frames the opportunity as the gap between your actual insurance net collection rate and 98.5 percent, since the denominator is contractual allowables. Then load expected reimbursement for your top two or three payers, including MPPR ranking and bilateral logic, and reconcile a quarter of surgical claims line by line. Note that MGMA polling found 24 percent of practices never audit payments against contracted rates.

‍How long do we have to dispute an underpayment?‍

It depends on the payer and, ultimately, on your participation agreement rather than the payer's public page. Aetna publishes 180 calendar days from the initial claim decision for reconsideration plus 60 days to appeal; UnitedHealthcare publishes twelve months to complete both steps. The reconsideration clock runs from the remittance date, not the date of service, and it is separate from and usually shorter than your timely filing window.

‍Are payers required to pay interest on late claims?‍

Under state prompt pay laws, yes, for state regulated plans. California moved to a 30 calendar day standard effective January 1, 2026, with 15 percent annual interest and a penalty of the greater of $15 or 10 percent of accrued interest for failing to pay it automatically. Texas imposes up to 18 percent of the contracted amount annually and its penalties cannot be waived by contract. New York pays the greater of the Tax Law rate or 12 percent. Self funded ERISA plans are exempt from state prompt pay laws entirely.

‍Why do orthopaedic practices have more underpayment exposure than other specialties?‍

Because expected reimbursement on an orthopaedic surgical claim is the fee schedule after multiple procedure ranking, bilateral adjustment and assistant surgeon percentage, each of which the payer computes with its own logic. A variance model that does not replicate that logic produces false positives on every multi procedure claim and misses the real errors. Implant carve outs and unlisted codes add two categories that no fee schedule model can evaluate at all.

‍Is it worth chasing small dollar underpayments?‍

Judge it against the MGMA benchmark of $25.20 as the average cost to rework a claim, and remember that underpayments are usually systematic rather than random. A single misapplied MPPR indicator or bilateral rule affects every claim in that pattern, so the correct unit of analysis is the pattern and the contract, not the individual claim.

See how Maia supports orthopaedic practices today. Book a demo at usemaia.com.