How to Reduce Denial Rate in Orthopaedic Billing: A 2026 Guide to Reading Your CARC and RARC Data

The fastest way to reduce your denial rate is to stop reporting it wrong and start reading the codes your payers already send you. Every dollar a payer does not pay arrives in the 835 electronic remittance advice with a group code and a claim adjustment reason code attached, because the transaction has to balance. That means the root cause of every denial in your practice is already sitting in structured data.

Most orthopaedic practices never use it, and many inflate their own denial rate by counting contractual adjustments and patient responsibility as denials. This guide covers what the codes actually mean, which ones are not denials at all, how to map the rest to the root causes that matter in orthopaedics, and what to fix first.

The code sets, and who controls them

Three code sets travel together on every remittance.

‍Claim Adjustment Group Codes classify who is responsible for the adjustment. There are four current values, all effective May 20, 2018: CO for contractual obligation, OA for other adjustment, PI for payor-initiated reduction, and PR for patient responsibility. Medicare contractors are limited to CO, CR, OA, and PR, and PI is not used for Medicare.

‍Claim Adjustment Reason Codes, or CARCs, explain why a claim or service line was paid differently than billed. They are hosted at x12.org and maintained by a code maintenance group under X12.

‍Remittance Advice Remark Codes, or RARCs, convey additional explanation. They are maintained by CMS. Some pair with a CARC, and some stand alone as alerts.

CMS notes that updates publish three times per year, around March 1, August 1, and November 1, and that the official X12 site may not match CMS system release schedules.

One correction that will save you arguments with payers: group code prefixes are usage conventions, not code definitions. X12 attaches an explicit group code restriction to very few CARCs. Code 18 carries an OA note, code 45 carries a PR or CO note, and most of the rest carry none. Write and speak about "CARC 197, typically transmitted with group code CO" rather than treating CO-197 as the code itself.

Stop counting these as denials

This is the most common error in practice denial reporting, and it makes your denial rate look worse than it is while hiding the denials that matter.

‍CARC 45 is "Charge exceeds fee schedule/maximum allowable or contracted/legislated fee arrangement." That is your contractual write-off. It is the difference between your charge master and your negotiated rate. It is not a denial.

‍CARC 1, 2, and 3 are "Deductible Amount," "Coinsurance Amount," and "Co-payment Amount." That is patient responsibility. It is not a denial either. It is a collections problem, and a growing one: Kodiak Solutions reported patient responsibility rising from 6.8 to 7.3 percent of net revenue between 2024 and 2025 while the patient collection rate fell from 45.1 to 42.4 percent.

If your denial reporting includes CARC 45 and the PR series, rebuild it before you do anything else. You are measuring the wrong thing.

The codes that matter in an orthopaedic practice

Here are the CARCs worth building a report around, with their official X12 descriptions.

‍CARC 197: "Precertification/authorization/notification/pre-treatment absent." This is the prior authorization code and it is the one to watch most closely in orthopaedics.

‍CARC 288: "Referral absent." Distinct from 197 and frequently miscounted with it.

‍CARC 4: "The procedure code is inconsistent with the modifier used." Your modifier error signal. In orthopaedics, this is modifier 25, 59, 50, RT and LT, and the global period modifiers.

‍CARC 11: "The diagnosis is inconsistent with the procedure." Your diagnosis-to-procedure linkage problem, and a frequent finding when ICD-10 specificity is thin.

‍CARC 97: "The benefit for this service is included in the payment/allowance for another service/procedure that has already been adjudicated." This is the bundling code, driven by NCCI procedure-to-procedure edits. In orthopaedics it clusters on arthroscopy, fracture care, and injection claims.

‍CARC B15: "This service/procedure requires that a qualifying service/procedure be received and covered. The qualifying other service/procedure has not been received/adjudicated." This is the add-on code failure mode. If you bill an add-on code without its qualifying primary procedure, this is what you get back.

‍CARC 50: "These are non-covered services because this is not deemed a 'medical necessity' by the payer." Your medical necessity signal.

‍CARC 151: "Payment adjusted because the payer deems the information submitted does not support this many/frequency of services." Frequency and units, which matters for injections, casting supplies, and therapy.

‍CARC 252 indicates that an attachment or other documentation is required to adjudicate the claim.

‍CARC 29: "The time limit for filing has expired." Pure workflow failure. Any volume here is a process defect.

‍CARC 16: "Claim/service lacks information or has submission/billing error(s)." This one is uninformative alone. Its meaning lives in the paired remark code. X12 explicitly requires that at least one remark code accompany it.

‍CARC 185 covers a rendering provider not eligible to perform the billed service, and CARC 208 covers an unmatched national provider identifier. Together with RARC N290, "Missing/incomplete/invalid rendering provider primary identifier," and RARC N286, "Missing/incomplete/invalid referring provider primary identifier," these are your credentialing and provider enrollment signal. In a growing orthopaedic group adding surgeons and advanced practice providers, this cluster spikes predictably after every hire.

‍RARC MA130 is worth memorizing: "Claim contains incomplete and/or invalid information, and no appeal rights are afforded because the claim is unprocessable." When you see MA130, do not build an appeal. Correct the claim and resubmit. Practices waste real staff time appealing unappealable claims.

‍CARC 22 covers coordination of benefits, CARC 27 expenses after coverage terminated, CARC 24 services covered under a capitation agreement or managed care plan, and CARC 109 a claim sent to the wrong payer. That group is your registration and eligibility signal.

Map the codes to root causes

Once you have the codes, group them into the buckets you can actually assign to an owner.

‍Front-end registration and eligibility: CARC 22, 24, 27, and 109. This is your patient access team.

‍Prior authorization and referral: CARC 197 and 288. This is your authorization team and your surgery scheduling workflow.

‍Coding and modifier: CARC 4, 11, 97, and B15, plus CARC 16 when paired with data-element remark codes. This is your coding team, and it is where documentation quality drives everything.

‍Clinical documentation and medical necessity: CARC 50, 151, and 252. This is a surgeon documentation problem that presents as a billing problem.

Then add two operational buckets: credentialing, meaning CARC 185, 208, and the N290 and N286 remark codes, and timely filing, meaning CARC 29.

The reason this mapping matters is accountability. "Our denial rate is 9 percent" is not actionable. "Forty percent of our denied dollars carry CARC 197, concentrated in two Medicare Advantage plans, on spine and total joint cases" assigns itself.

What the benchmark data says about where denials come from

The best publicly available breakdown of denial causes remains Optum's 2024 Revenue Cycle Denials Index, built on roughly 124 million hospital claim remits across more than 1,400 hospitals for calendar year 2023. It found a national average denial rate of 12 percent, up from 9 percent in 2016, with causes breaking down as registration and eligibility at 24.33 percent, missing or invalid claim data at 15.89 percent, authorization and precertification at 12.80 percent, medical documentation requested at 12.08 percent, service not covered at 9.67 percent, and medical necessity at 6.76 percent. Forty-four percent originated front-end. Eighty-four percent were characterized as potentially avoidable, and of those, 40 percent were nonrecoverable.

Note that this is hospital data and the most recent edition covers 2023.

For medical groups specifically, MGMA DataDive reported a single-specialty aggregate first-pass denial rate of 8 percent, and MGMA Stat polling in March 2024 across 235 respondents found 60 percent of leaders reporting rising denial rates, with incorrect modifier use, particularly modifier 25, among the top named causes. Experian Health's third annual State of Claims survey, published September 2025 across 250 finance and claims professionals, found 41 percent reporting denial rates of 10 percent or higher, with missing or inaccurate data cited by 50 percent, authorizations by 35 percent, and incomplete registration data by 32 percent.

HFMA's published KPI guidance sets a clean claim rate target of 98 percent, a denial rate range of 5 to 10 percent with under 5 percent described as optimal, resolution of 85 percent of denials within 30 days, and days in accounts receivable of 30 to 40.

Two orthopaedic-specific findings are worth carrying into your prior authorization workflow. Premier's analysis of 2023 data across 280 hospitals found that 10.4 percent of denied claims had already been pre-approved through prior authorization, up from 3.2 percent in 2022.

Having the authorization is no longer sufficient evidence that you will be paid. And at the AAOS 2025 Annual Meeting, a study of 3,922 commercially insured total hip arthroplasty patients found that 72.4 percent required prior authorization, that any form of denial ran 4.8 percent in the authorized cohort versus 3.0 percent without, and that authorization added 2.1 days to time to surgery without producing cost savings.

The new leverage: your payers now publish their own denial rates

This is the most underused development in denial management.

Under the CMS Interoperability and Prior Authorization Final Rule, CMS-0057-F, impacted payers, meaning Medicare Advantage organizations, Medicaid and CHIP fee-for-service programs and managed care plans, and qualified health plan issuers on the federally facilitated exchanges, must post annual prior authorization metrics publicly, including the percentage of requests approved, denied, and approved after appeal, and the average time from submission to decision. The first reports were due March 31, 2026, covering CY2025.

KFF's August 2026 analysis of those disclosures found standard-request denial rates of roughly 12 percent for Medicare Advantage, 14 percent for Medicaid managed care, and 18 percent for Marketplace plans, with appeal overturn rates of 67, 47, and 43 percent respectively. Forvis Mazars, analyzing the same disclosures, reported a median Medicare Advantage denial rate of 7.2 percent, that only 7.3 percent of denials were appealed, and that 52.6 percent of appealed denials were overturned, with 92 percent of required payers having published by July 2026. The two analyses use different denominators, so cite each with its source rather than blending them.

The practical use is immediate. Pull the published metrics for every impacted payer in your contract portfolio. Compare your CARC 197 denial rate against what that payer reports nationally. If you are materially worse, the problem is upstream in your own workflow. If you are in line and the payer's rate is simply high, you have the payer's own published number for your next contract conversation.

What to fix first

Rebuild the denial report. Exclude CARC 45 and the PR series. Report denied dollars, not denied claims, grouped by root-cause bucket, then by payer, then by service line. Denied dollars is the number that ranks your work correctly, because a denied total knee and a denied office visit are not the same event.

‍Attack CARC 197 with the payers' own data. It is the largest clinical denial driver, it is concentrated in Medicare Advantage, and appeal overturn rates are high. Set a policy that every clinically supported authorization denial gets appealed.

‍Separate MA130 from the appeal queue. Those claims are unprocessable and carry no appeal rights. Correct and resubmit.

‍Audit your CARC 97 volume against NCCI. Bundling denials in orthopaedics cluster on arthroscopy, fracture care, and injections. Some are correct edits you should stop billing against. Some are payer policy denials that are winnable on appeal. You cannot tell which without pulling the operative notes.

‍Watch CARC 4 as your coding quality metric. Modifier and procedure inconsistency is the cleanest signal that coding accuracy is improving, and it responds quickly to intervention.

‍Build the credentialing cluster into your onboarding checklist. CARC 185, 208, and remark codes N290 and N286 spike after every new surgeon or advanced practice provider starts. That is predictable and therefore preventable.

Frequently asked questions

What denial rate should an orthopaedic practice target?

HFMA's published guidance puts the industry range at 5 to 10 percent, with under 5 percent described as optimal, and a clean claim rate target of 98 percent. MGMA DataDive reported a single-specialty aggregate first-pass denial rate of 8 percent. There is no published, credible orthopaedics-specific denial benchmark in the public domain, and vendor claims of a specific ortho denial rate generally have no disclosed methodology. Measure your own baseline, excluding contractual adjustments and patient responsibility, and improve against it.

‍What is the difference between CARC 16 and MA130?

CARC 16 says the claim lacks information or has a submission error, and X12 requires at least one remark code to accompany it, because 16 alone tells you nothing. MA130 is a remark code stating the claim is unprocessable and no appeal rights are afforded. When you see MA130, correct and resubmit rather than appealing.

‍Why did we get CARC 97 when the procedures were clearly separate?

CARC 97 is the bundling code, driven by NCCI procedure-to-procedure edits. In orthopaedics the common cause is an edit that cannot be bypassed at all. For example, the 2026 NCCI Medicare Policy Manual states that shoulder arthroscopy procedures include limited debridement even when performed in a different area of the same shoulder, and that ipsilateral shoulder edit pairs generally cannot be bypassed with a modifier. Check the edit before appealing, because appealing a correct edit costs staff time and creates a pattern.

‍How much does it cost to rework a denied claim?

MGMA Stat polling with Change Healthcare found an average of 25.20 dollars per claim in a practice setting, with 86 percent of denials potentially avoidable and 48 percent of avoidable denials not recoverable. Premier's hospital-based analysis found the cost to appeal rising from 43.84 dollars in 2022 to 57.23 dollars in 2023. Practice-side rework is cheaper than hospital-side, but the ratio that matters is your rework cost against your average denied dollar amount, which for surgical claims usually strongly favors appealing.

‍Should we appeal everything?

No, but most practices appeal far too little. KFF found that only 11.5 percent of Medicare Advantage prior authorization denials were appealed in 2024 while 80.7 percent of appeals were partially or fully overturned. Kodiak reported a provider appeal success rate of 42.1 percent in 2025 across all denial types. The rule that works is: appeal every clinically supported denial above a dollar threshold you set, never appeal an MA130, and never appeal a correct NCCI edit.

‍Where do CARC and RARC codes actually come from on a remittance?

They arrive in the 835 electronic remittance advice, produced in the HIPAA-compliant ASC X12 835 format. CARCs and their dollar amounts sit in CAS segments at both claim and service-line level, each carrying a group code, a reason code, and an amount. RARCs arrive in LQ, MOA, and MIA segments. Because the 835 must balance at the service, claim, and transaction levels, every unpaid dollar carries a reason code. That is why denial data is complete in the 835 in a way it never is on a paper explanation of benefits.

See how Maia is helping orthopaedic practices. Book a demo at usemaia.com.