Undercoding Isn't Playing It Safe: The Compliance Problem Hiding in Conservative Coding

Here is the contrarian truth: choosing a lower code than the documentation supports is not automatically the safer choice. Accurate coding means reporting the services actually performed and supported, not defaulting to the lowest level to avoid scrutiny.

I’m Dr. Dreama Sloan-Kelly, MD/CCS/CPC, and I approach this as a physician executive focused on the connection between clinical documentation, coding, and revenue integrity. For surgical practices, the issue can reach beyond the procedure code itself: it may include a separately reportable service, a supported modifier, or a documented component of care that never makes it onto the claim.

The goal is not to code higher. It is to make the code match the record.

What undercoding actually means

Undercoding occurs when a claim reports a lower service level than the documentation supports, or leaves out a billable service, modifier, or supply that was performed and is reportable under the applicable rules. It can result from an oversight, limited formal coding education, or a well-intended belief that conservative claims attract less attention.

That last assumption is worth challenging. The AAPC’s discussion of undercoding explains that undercoding is not an audit-avoidance strategy. Inaccurate coding can create payment errors even when the result is a lower payment than the documentation supports. The direction of the error does not make the code accurate.

For E/M services, the level must be selected using the applicable criteria. Under current office and outpatient guidance, that generally means medical decision making (MDM) or total time on the date of the encounter. If your team needs a refresher, see our breakdown of EM coding in 2026 and the CMS Evaluation and Management Services guide.

For surgical claims, the same basic principle applies: code what the documentation supports. The operative report, associated services, and any modifiers should tell a consistent story. A modifier should not be added because it might increase payment, and it should not be omitted simply because the team worries it could prompt a review. For more, see our guidance on modifier usage: when to use it and when to leave it off.

Why “conservative coding” doesn’t protect you

Most people get this wrong: treating undercoding as the opposite of overcoding, and therefore as the harmless one. They are different directions of error, but both can signal a breakdown in the same process: translating the documented service into the correct code.

A payer may evaluate whether a claim’s coding pattern is consistent with the documentation and applicable rules. Patterns that differ from peers or from a practice’s own expected mix can prompt questions; an unusually low pattern is not automatically safer than an unusually high one. Neither a low code nor a high code is a substitute for applying the coding criteria to the record.

There is also a compliance dimension, and it is worth being precise about it. Undercoding is not automatically a regulatory violation. Coding inaccuracies and regulatory violations are different categories: a miscoded claim is a claim-integrity problem, while a violation arises from the facts, the payer or program requirements involved, and whether the pattern was knowing or systematic. That distinction matters because the practical response differs. The goal is not to accuse or overcorrect; it is to investigate systematically inaccurate coding and correct the documentation-to-code process so claims are supported either way. An inaccurate claim does not become accurate simply because it was paid at a lower amount, and systematic inaccuracies warrant investigation rather than dismissal as harmless conservatism.

An undercoded claim and an overcoded claim can both grow out of the same root: a documentation-to-code translation process that is not grounded in the record. The correction is not “always choose the lower level.” It is to make the code accurately reflect the supported service.

If your team is preparing for a payer review, our practical guide to surviving a payer audit before, during, and after can help you think through the process without making assumptions about the outcome.

The data feedback loop: today’s coding affects tomorrow’s picture

Codes do more than determine the payment on an individual claim. Submitted claims contribute to utilization and resource data used in reimbursement calculations. CMS’s Physician Fee Schedule resources explain how relative value units and other components factor into Medicare payment.

When services are systematically reported below what the documentation supports, the effect can reach beyond one encounter. Inaccurate claims data distort a practice’s own utilization reporting and internal revenue analysis: management is making staffing, service-line, and investment decisions based on a picture of the work that does not match the record. At the aggregate level, claims contribute to the utilization data sets that inform future relative-value and reimbursement calculations, but no single practice’s coding determines those valuations; the process depends on the combined accuracy of the broader data. One practice cannot control the aggregate, but it can avoid adding inaccurate information to it, and it can protect the accuracy of the numbers its own leaders rely on.

This does not mean coding decisions should be driven by RVUs or payment targets. They should be driven by the service performed, the documentation, and the applicable coding rules. Accurate reporting supports a more reliable picture of the work surgical teams actually provide.

What this looks like for a practice's revenue. Consider a surgical group whose E/M claims consistently report levels below what the documentation supports. Each individual difference may look small, but across a year of encounters the aggregate can be substantial: payment that was never captured for services that were furnished and documented. At the same time, the practice's internal reporting understates the true complexity of its work. Utilization reports show a mix weighted toward lower levels, service-line and staffing decisions get made against that incomplete picture, and when leadership compares its revenue per encounter against benchmarks, the comparison is distorted in both directions at once. Nothing in this analysis suggests coding higher; it suggests that what the practice bills, measures, and manages should match the work its documentation describes.

AI scribes, richer notes, and payer downcoding

AI documentation tools and scribes can capture more detail in the medical record. That can be helpful, but only if the coding workflow keeps pace. A longer, richer note does not automatically support a higher code, and more documentation should not be mistaken for more complexity. At the same time, when the note documents work that meets the criteria for a particular level, a coder or automated tool that routinely defaults lower can create systematic undercoding.

There is pressure in the other direction, too: some payers run E/M downcoding programs that reduce payment for a claim they consider above the supported level. The AMA’s guidance on payer E/M downcoding describes these programs and notes that an increasing number of payers downcode claims automatically using software algorithms, without first requesting and reviewing the medical record. The guide explains how to recognize when claims have been downcoded: staff should review remittance advice carefully and compare the payment received for each billed code against the anticipated amount.

The guide lists example codes whose language may indicate that a claim has been downcoded. For technical precision, these fall into two classifications: CO-150 pairs the group code CO (Contractual Obligation) with Claim Adjustment Reason Code (CARC) 150 — payer deems the information submitted does not support this level of service; M85 and N610 are Remittance Advice Remark Codes (RARCs) — M85: subjected to review of physician evaluation and management services; N610: payment based on appropriate level of care; and CARC 186 is a Claim Adjustment Reason Code: level of care change adjustment.

Two complications are worth knowing. First, downcoding takes two forms: some payers change the billed code to a lower-level service, while others adjust payment without changing the submitted procedure code at all — which makes the adjustment nearly invisible in coding audits. Second — and this is our professional guidance, not a statement from the AMA — the presence of one of these codes alone does not establish payer downcoding. These codes are clues, not a conclusion. Before drawing conclusions, review the full remittance advice, the payer's explanation, the claim details, and the applicable policies.

If a payer has reduced payment and you believe the record supports the billed level, the AMA recommends appealing with documentation that supports the level under MDM or total time, as applicable. Track recurring adjustments so your team can identify patterns by payer and service. The AMA also notes that certain claims are more likely to be subject to payer downcoding than others — pay special attention to claims indicating a high level of service (for example, 99204, 99205, 99214, and 99215) and complex claims, including those using modifier 25.

One more note on timing. In its CY 2027 Medicare Physician Fee Schedule final rule, CMS finalized a payment policy that touches directly on this discussion. Effective January 1, 2027, when a separately identifiable office/outpatient E/M visit (reported with modifier 25) is furnished on the same day as a procedure with a 0-day global period, the most expensive service is paid at 100% and the other surgical procedure(s) or office/outpatient E/M visit(s) furnished that day are paid at 80% — a 20% reduction of the lower-valued service. The policy applies when the E/M visit and procedure are furnished by the same physician or a physician in the same group practice; CMS limited the finalized policy to 0-day global procedures and excluded the 10- and 90-day global procedures included in the July 2026 proposal, and scaled the reduction back from the proposed 50%. The services the AMA flags as more likely to be downcoded — high-level E/M visits and claims using modifier 25 — may overlap with those affected by this 2027 payment policy, which is not limited to high-level E/M services; applicability depends on the policy's specific requirements involving office/outpatient E/M services, modifier 25, and 0-day global procedures. For coding and billing teams, the practical takeaway is that accurate modifier selection and documentation-supported E/M coding carry more payment weight in 2027, not less — which strengthens the case for coding the documented service rather than defaulting lower. We will address the broader fee-schedule changes in a separate article.

A common scenario: a surgeon’s note captures substantial work around an encounter, but an E/M code is selected by habit rather than by reviewing the documented MDM or time. Separately, the remittance later shows an adjustment that may indicate payer downcoding. In both situations, the practical response is the same: inspect the record and claim together, then determine whether the code and payment reflect the supported service.

A practical self-check for systematic undercoding

A brief review of your claims can help show whether a coding habit deserves a closer look. The following are prompts, not automatic evidence of an error:

• Review E/M level distribution. Look for a pattern that seems unusually concentrated at the lowest levels. Compare like services and specialties, and review documentation before drawing conclusions. The goal is not to match a target distribution.
• Check for missed modifiers. Confirm that staff can identify when a modifier is supported and when it is not. Review a sample of surgical claims alongside the operative documentation.
• Look for unbilled components. Consider whether documented, separately reportable services, supplies, or other claim elements are consistently omitted. Apply payer and coding rules; not every documented activity is separately billable.
• Compare documentation with submitted codes. Does the record describe work that meets the code’s criteria, while the claim repeatedly reports a lower level? Review examples with the coder and clinician to understand why.
• Review undercoding and overcoding together. An effective education process looks at errors in both directions and focuses on how the documentation-to-code decision is made.
• Follow up on payer adjustments. Compare remittance amounts with expected payment and investigate relevant remark or adjustment codes. Document the rationale and use the payer’s appeal process when appropriate.

The purpose of this review is to identify where training, workflow, or documentation guidance may help, not to encourage higher coding. As a physician consultant, I believe the most useful question is simple: Can your team explain how the service documented led to the code submitted?

Physician-led coding and documentation audits

If your practice wants a clear answer to whether undercoding, overcoding, missed reporting opportunities, or documentation gaps are affecting its claims, a physician-led audit is the place to start. Our physician-led audits combine clinical expertise with coding and documentation analysis to identify coding inaccuracies, missed reporting opportunities, and documentation gaps before they become larger reimbursement or compliance concerns.

Schedule a consultation to discuss where a physician-led audit would make the difference for your practice.

Code the service the documentation supports — neither lower out of fear nor higher for payment.

Don't miss the next update. Each month I send The DocsDoc Brief — clinical documentation, coding, compliance, and AI insights — straight to your inbox. 👉 Subscribe to The DocsDoc Brief here

Dr. Dreama Sloan-Kelly, MD, CCS, CPC
CEO/President, Dr. Sloan-Kelly Consulting LLC