The CO-97 Bundling Appeals Playbook: Recover the Revenue Most Billers Write Off

CO-97 fires when a payer determines that a service billed on a claim was bundled into the payment for another procedure billed on the same date of service. The denial means the payer paid for one of the two codes (the column 1 code, sometimes called the comprehensive code) and considers the other code (the column 2 code, the component code) to be already included in that payment. The bundling rule can come from one of three sources: the National Correct Coding Initiative (NCCI) Procedure-to-Procedure (PTP) edit table maintained by CMS and updated quarterly, payer-specific bundling rules that go beyond NCCI (UnitedHealthcare, Anthem, and other major commercial payers maintain extensive proprietary Reimbursement Policy libraries), or the global surgical package rules that bundle related E/M and minor procedures into a base surgical code's 0, 10, or 90 day global period. CO-97 does NOT mean the service was billed in error. It means the payer applied a bundling rule. Whether that rule was correctly applied to your specific clinical scenario is the question every CO-97 should trigger before deciding to appeal or write off.

Every NCCI PTP edit pair carries a modifier indicator: 0, 1, or 9. This indicator determines whether the edit can be bypassed by appending a modifier on resubmission. Modifier indicator 0 means the edit cannot be bypassed under any circumstance. No modifier will allow separate payment. Do not appeal a CO-97 with indicator 0 on a clinical basis. The bundle is absolute regardless of what the documentation shows. Either accept the bundled payment or, if the bundling decision was clinically inappropriate for the case (rare), submit a corrected claim with only the column 1 code. Modifier indicator 1 means a modifier MAY bypass the edit when the clinical circumstance supports a distinct, separately identifiable service. This is the most common indicator, and this is where the recoverable revenue lives. Appropriate modifiers include 59 (distinct procedural service), the more specific X-modifiers (XE, XS, XP, XU), modifier 25 when an E/M is involved, and others depending on the clinical scenario. Modifier indicator 9 means not applicable; the edit pair has no modifier provision documented in the current NCCI file. Always run the NCCI quarterly update lookup before submitting an appeal so you know which indicator applies to your specific code pair on the specific date of service.

Modifier 59 is the historical workhorse for bypassing NCCI edits. It still works. But CMS introduced the X-modifiers (XE, XS, XP, XU) in 2015 specifically to add precision to modifier 59 use because audit data showed it was being applied without consistent rigor. Major payers including Medicare Administrative Contractors and several commercial payers now prefer the X-modifiers over modifier 59 and apply less audit scrutiny when the more specific modifier is used. XE means Separate Encounter: the two services were performed during different patient encounters on the same date. Use when the procedures occurred at clearly different times of day or at different facility visits. XS means Separate Structure: the two services were performed on different anatomic sites or organ systems. Use when the column 1 procedure was on one site and the column 2 procedure was on a clinically distinct site. XP means Separate Practitioner: the two services were performed by different physicians or other qualified healthcare providers. Use when documentation clearly establishes that two different rendering providers performed the work. XU means Unusual Non-Overlapping Service: the service does not overlap usual components of the main service. Use when the service is genuinely distinct but does not fit the other categories. Practical playbook: when an indicator-1 edit fires, ask which X-modifier most precisely describes why the services were distinct. Use that one. Default to modifier 59 only when no X-modifier accurately fits.

When CO-97 fires on an E/M code billed alongside a minor procedure on the same date, the right modifier is usually 25 on the E/M code, not 59 on the procedure. Modifier 25 specifically indicates that a significant, separately identifiable E/M service was provided on the same date as a procedure with a 0 or 10 day global period. The E/M must address something beyond the routine pre-procedure evaluation that would normally be bundled into the procedure payment. Documentation requirements: a distinct chief complaint or HPI section addressing problems not directly related to the procedure indication, a separate assessment and plan element for the E/M-related problems, and ideally a clear narrative separator in the note ('After completing the above E/M evaluation, the patient consented to and underwent [procedure]'). When the E/M and procedure documentation visually flow into each other, audit teams flag the modifier 25 use as unsupported. When they are visually separate, the modifier 25 survives. Common winning scenarios: a patient comes in for a planned joint injection but during the visit also discusses a new complaint requiring evaluation. A patient presents for a screening colonoscopy and during the pre-procedure evaluation reports new symptoms requiring assessment. A patient has a wound check (planned) and also reports new chest pain requiring workup.

Step 1: Read the EOB carefully. Identify the column 1 (paid) code and the column 2 (denied) code. Note the date of service and the modifier indicator if the EOB provides it. Some payer EOBs include the indicator; others do not, in which case you look it up in the current NCCI quarterly file. Step 2: Pull the chart. Read the documentation for the date of service. Determine whether the two services were genuinely distinct (separate site, separate session, separate practitioner, distinct clinical scenario) or whether the bundling rule was correctly applied. If the bundling was correct, write off the column 2 line and move on. If the services were genuinely distinct, proceed. Step 3: Determine the right modifier. If E/M is involved, modifier 25 is usually the path. If two procedures are involved, identify the X-modifier (XE, XS, XP, XU) that most precisely describes why they were distinct. Use modifier 59 only if no X-modifier accurately fits. Step 4: Submit a corrected claim with the modifier appended on the column 2 code. A corrected claim is faster than a written appeal for indicator-1 edits. Submit through your clearinghouse with the appropriate corrected-claim type code (typically frequency code 7 in the 837 transaction). Step 5: If the corrected claim also denies, escalate to a written first-level appeal with the chart documentation attached, the NCCI quarterly file showing indicator 1, the modifier rationale tied to the specific clinical facts, and a request for reconsideration. Most indicator-1 CO-97 cases resolve at the corrected claim stage. The appeals stage is only needed when the corrected claim is rejected on additional grounds.

Three scenarios where appealing CO-97 is wasted effort. First, indicator 0 edits. The bundle is absolute. No modifier will bypass it. Some practices waste hours filing appeals on indicator 0 cases that have no chance. Stop. Look up the indicator before appealing. Second, payer-specific bundling beyond NCCI. UnitedHealthcare, Anthem, and several others maintain proprietary bundling rules that are NOT in the NCCI file. These appear as CO-97 denials but the modifier indicator question does not apply because the bundling source is the payer's medical policy, not NCCI. For these, you must work the appeal against the payer's specific medical policy text. Pull the policy, find the bundling rule cited, and demonstrate why the policy does not apply to your case. These appeals win less often than NCCI-based ones but are still worth working when the documentation supports a clearly distinct service. Third, global period bundling. If the column 2 code is being bundled because it falls within the 10 or 90 day global period of a previously billed major surgery, the appeal path is modifier 24 (unrelated E/M during global), modifier 79 (unrelated procedure during global), or modifier 78 (related return to the OR), not modifier 59 or X-modifiers. The modifier choice depends on the clinical relationship of the current service to the global procedure.

A four-physician cardiology practice in the Midwest had been writing off CO-97 denials at a rate of approximately 40 to 50 per month. Quarterly write-offs averaged 42,000 dollars before our team took over. The pattern: 93306 (transthoracic echocardiogram complete with Doppler) was being denied as bundled into 93000 (electrocardiogram complete) on encounters where both were performed. The denial was a payer-side application of an indicator-1 NCCI edit. The modifier rationale was straightforward: the EKG and echo were two distinct diagnostic studies addressing different clinical questions, performed on the same date as part of a comprehensive cardiovascular workup. Our team began appending modifier 59 (and later XU as the more precise option) on resubmission, with chart documentation showing the EKG read and the separate echo report. Recovery rate on submitted appeals: roughly 78 percent on first-level resubmission. Quarterly recovery jumped from zero to over 32,000 dollars within 90 days of process implementation, with the remaining 10,000 dollars recovered through second-level appeals or absorbed as legitimately bundled cases. The practice's overall CO-97 write-off rate dropped 75 percent. Total annualized recovery: roughly 130,000 dollars on what had been considered un-recoverable revenue.

Recovery is one half of CO-97 economics. Prevention is the other half. The prevention layer has three components. Component one: pre-submission scrubber. Every claim should run through an NCCI quarterly edit check before submission. Generic clearinghouse scrubbers catch many but not all PTP edits. Specialized scrubbers (NextGen, athenahealth, Waystar Edits, claimMD) catch more and update faster after each NCCI quarterly release. Verify that your scrubber is updated to the current quarter. Many practices run on stale edit files. Component two: payer-specific edit packs. Beyond NCCI, the major commercial payers publish or maintain proprietary bundling rules. Subscribing to payer-edit-pack updates from a clearinghouse or a service like Cotiviti reduces post-submission CO-97 denials by 30 to 50 percent within 60 days. Component three: provider documentation training. The most common upstream cause of CO-97 denials that fail appeal is documentation that does not clearly establish the distinctness of the services. Providers who reflexively document distinct anatomic sites, separate sessions, distinct practitioners, or unusual circumstances at the time of the encounter create the audit trail that supports the modifier strategy. Practices that combine all three components see CO-97 denial rates drop by half within a quarter while appeal-recovery rates climb because the underlying documentation is stronger.

CO-97 is a recurring revenue line for almost every practice we manage. Our process: every CO-97 denial that crosses our queue is reviewed by an AAPC-certified coder against the chart documentation within 24 hours. We determine the modifier indicator for the specific code pair on the specific date of service from the current NCCI file. We pick the precise modifier (X-modifier preferred over modifier 59, modifier 25 when E/M is involved, global-period modifiers when applicable). We file the corrected claim within 48 hours. We track resolution outcomes by payer and by code pair, building a payer-specific intelligence layer that improves modifier selection over time. We surface payer-specific bundling patterns to clients monthly so providers can adjust documentation in real time. Most CO-97 denials are recoverable revenue. Our average appeal-recovery rate on indicator-1 cases is in the 70 to 80 percent range. Pricing starts at 2.49 percent of net collections, no setup fees. For a practice with 40 to 50 CO-97 denials per month, the recovered revenue alone typically covers the entire billing service cost.