A Routine Alert With an Uncomfortable Pattern
On September 24, 2026, the Centers for Medicare & Medicaid Services used its weekly MLN Connects newsletter to flag a problem that has nothing to do with fraud and everything to do with paperwork: a 32.6% improper payment rate on claims for toileting aids — commodes, bed pans, and urinals — with a projected improper payment total of roughly $2.5 million. CMS’s guidance to providers was blunt: know the coverage criteria, and document to it, or expect the claim to come back.
It’s a small line item in a routine newsletter, but it’s also a useful data point. CMS publishes dozens of these code-specific improper-payment alerts every year, and the pattern behind them is remarkably consistent: the codes themselves are rarely in dispute, but the medical record supporting them almost never says enough. For coding and compliance teams, that consistency is the real story.
What CMS Actually Found
The MLN item is narrow by design — it covers HCPCS codes for commodes, bed pans, and urinals billed under the durable medical equipment benefit — but the underlying finding generalizes. CMS’s Comprehensive Error Rate Testing (CERT) program, which is what generates these improper-payment figures, doesn’t typically find that beneficiaries received equipment they didn’t need. It finds that the documentation in the chart doesn’t establish medical necessity the way the coverage policy requires: no clinical note tying the item to a specific functional limitation, no evidence the ordering clinician evaluated the patient for that need, or missing detail on the exact type of equipment supplied.
A Familiar Companion Finding
The timing lines up with a separate, higher-profile audit. On September 17, 2026, Healthcare Dive reported that the HHS Office of Inspector General found HumanaChoice and UnitedHealthcare of Wisconsin plans had claimed a combined $178 million in improper Medicare Advantage risk-adjustment payments for 2020 and 2021, after sampling found the majority of reviewed enrollee-years lacked medical record support for the diagnosis codes submitted. Different program, different dollar scale, same root cause: the code that was billed outran what the documentation could support.
Two audits in the same week, at opposite ends of the payment system, pointing at the same gap. That’s not a coincidence CMS engineered — it’s what happens when documentation habits lag behind coding requirements across an entire industry.
Why This Keeps Happening
Coders are trained to code to the documentation in front of them, not to guess at what a clinician probably meant. But CMS coverage policies for high-scrutiny categories — DME, incontinence supplies, ambulance transport, certain HCC-relevant diagnoses — often require a specific combination of facts to be present in the note before a code is even payable, and clinical documentation habits weren’t built around those requirements. The result is a structural mismatch: clinically appropriate care, correctly selected codes, and a claim that still fails an audit because the chart doesn’t say the right things in the right place.
The categories CMS and the OIG keep coming back to share a few traits:
- Medical necessity depends on a functional or clinical detail that’s easy to omit from a routine note (a specific limitation, a failed conservative treatment, a severity threshold)
- The physician order or face-to-face encounter documentation is a separate artifact from the clinical note, and one can exist without the other
- Diagnosis specificity affects payment directly, so an imprecise but clinically “close enough” code has real financial consequences
- The item or service is billed at high volume, so even a modest per-claim error rate produces a large aggregate improper-payment figure
- Coverage criteria are published separately from clinical workflow tools, so the people documenting the encounter rarely see the payer’s checklist
What It Means for Coding and CDI Teams
The Documentation-to-Code Link Has to Be Checked, Not Assumed
The practical takeaway isn’t “code more conservatively.” It’s that coding accuracy and documentation completeness are two different problems that have to be solved together, at the point of coding, rather than discovered later in an appeal or an audit response. A code can be technically correct for the diagnosis or procedure described and still fail an improper-payment review because the supporting elements CMS looks for — medical necessity language, order documentation, specificity — simply aren’t in the note.
That’s the case for building the check into the coding workflow itself: flagging, at the time a claim is coded, when a high-scrutiny code is being applied to documentation that’s missing one of the elements CMS or a payer is known to require, before the claim goes out rather than after it comes back. It’s a narrower, more mechanical task than clinical judgment, which is exactly why it’s a good candidate for automation layered on top of — not instead of — coder and CDI review.
That’s the gap Medikode’s automated medical coding platform is built to close: it applies coverage and documentation rules alongside code selection, so gaps like the ones CMS just flagged surface before submission instead of after a CERT review or a takeback letter.
Closing the Gap Before the Next Audit
None of this requires waiting for CMS to publish the next alert. The categories most likely to show up in a future MLN newsletter or OIG report are the ones where documentation and coverage policy have always been loosely coupled — and that list is knowable in advance. Toileting aids this week, a different high-volume HCPCS or diagnosis category next month; the fix is the same in every case: check the documentation against the coverage rule at the moment the code is selected, not after the remittance advice comes back.