Why Your Billing Company Doesn't Catch Undercoding

If you're an urgent care operations leader, you've probably assumed your RCM vendor or billing company has coding accuracy covered. They process claims, handle denials, manage the revenue cycle. Coding is their job.

It is — but not the part that's costing you money.

The distinction between claim processing and code accuracy

Revenue cycle management platforms are built to move claims through payer adjudication as efficiently as possible. They scrub claims for formatting errors, missing modifiers, invalid ICD-10 combinations, and other technical issues that trigger denials. They are extremely good at this.

What they do not do is open the chart, read the clinical documentation, evaluate the three elements of medical decision-making, and determine whether the billed E/M code matches the complexity that was actually documented.

That is a fundamentally different function. It requires clinical judgment, knowledge of the 2021+ MDM framework, and the ability to distinguish — for example — whether a documented prescription for doxycycline with a discussion of adverse effects constitutes "prescription drug management" under the Risk element of MDM. A billing platform doesn't evaluate that. It sees a 99213 come through, confirms the claim is clean, and sends it to the payer.

Why payers don't flag it either

Payers have no reason to tell you that you're billing below what your documentation supports. Undercoding benefits them. They pay less for a 99213 than a 99214 — roughly $50 less per visit, depending on payer and region. If your providers are consistently documenting Level 4 complexity but submitting Level 3 codes, the payer receives a lower claim and adjudicates it without objection.

Payers do flag overcoding. They audit claims that look high relative to specialty benchmarks, they request documentation for 99215 visits, and they recoup overpayments. The incentive structure is asymmetric: there is a built-in mechanism to catch overbilling and zero mechanism to catch underbilling.

Where the undercoding actually happens

In urgent care, the most common undercoding pattern is a Level 3 visit (99213) that should be a Level 4 (99214) based on the documented medical decision-making. This doesn't happen because providers are doing anything wrong — it happens because the E/M code is often selected before the full clinical picture is reflected in the documentation, or because the provider defaults to a comfortable code level out of habit or risk aversion.

The MDM drivers most commonly missed are not exotic. They are routine clinical activities that providers perform and document every day:

Prescription drug management — any new prescription or dose change that requires documented risk-benefit consideration. This is one of the most common moderate-Risk drivers in urgent care and it shows up in a large percentage of visits.

Acute illness with systemic symptoms — presentations where the documented symptoms extend beyond the primary site of complaint into constitutional or multi-system involvement. The distinction between an uncomplicated acute illness (low complexity) and one with systemic symptoms (moderate complexity) is one of the most consequential in urgent care coding.

External records reviewed — any time the provider documents reviewing prior imaging, outside lab results, or records from another facility. This satisfies a Data element that can elevate MDM to moderate complexity.

These drivers are present in the chart. They are simply not being reflected in the code selection.

What your billing company would need to do differently

To catch undercoding, your billing company would need to assign a clinically trained auditor to open each chart, read the provider's documentation, evaluate the three MDM elements against the 2021+ framework, and compare the supported code to the billed code. This is a time-intensive, high-skill function. It is not part of a standard RCM contract, and it would be cost-prohibitive to perform on every encounter.

That is exactly why independent coding audits exist. A structured sample — typically 25 charts per provider — is enough to identify whether a systematic pattern exists, which providers are most affected, and which specific MDM drivers are being missed. It's a diagnostic tool, not an ongoing service.