Medicare DMEPOS NPWT Payment Estimates in 2024–2025: the Evidence Buyers Still Need

What the two CMS rows actually measure

The Comprehensive Error Rate Testing program, CERT, estimates improper payments in Medicare fee-for-service. Tables G2 and K4 report DMEPOS policy groups; this article examines their NPWT rows. The methodology projects reviewed claims to the relevant payment universe using sampling weights and expenditure information. The improper-payment rate is a payment-dollar measure. It is not a clinical complication rate, a device malfunction rate or a simple count of denied claims divided by submitted claims. ²

That distinction changes how the NPWT figures should be read. Applying 11.1% to the 45 reviewed claims would confuse a projected dollar ratio with an unweighted count of improper claims. Claim values and sampling weights matter. Nor does the projected improper-payment amount mean CMS examined that many dollars in the small NPWT sample or assessed that amount against one supplier.

The reporting year is also not the same as the underlying claims-submission period. The 2024 report's footnote identifies claims submitted from 1 July 2022 through 30 June 2023. The 2025 report identifies 1 July 2023 through 30 June 2024. Describe these as historical annual-report estimates, with their respective claim periods, when using them in October 2026. ^{1 2}

Source: 2024 and 2025 CMS Medicare FFS supplemental reports, Table G2, DMEPOS NPWT policy group, printed page 46 in both reports. Values are reproduced at the precision CMS publishes. ^{1 2}

The last row is particularly easy to misuse in a presentation. Both reports display 0.0% as NPWT's share of overall improper payments, while the same rows show positive projected dollar amounts. The positive projected dollar amount resolves the apparent zero created by rounding. The correct sentence is that NPWT represents a small displayed share at this precision, not that CMS found no payment problem.

Interpreting the reduction from about $12.1 million to $6.5 million as lost NPWT market revenue, changed device sales or a new entrant's reimbursement exposure would substitute a different measure for projected improper payments. Those questions have different denominators and require different evidence.

Separate the lower estimate from explanations of the change

Subtracting the published rates gives 11.1% − 17.0% = −5.9 percentage points. It is reasonable to report that lower point estimate. Attributing it to improved documentation or a smaller reviewed sample requires evidence identifying that cause. Supplier behavior, product mix, documentation systems and enforcement remain possible questions for investigation, beyond the causal reach of two aggregate rows. ^{1 2}

The sample count fell by 65 claims, or approximately 59.1%. That describes the reviewed sample. National NPWT claim volume is a separate measure. A sample-size change can affect precision through the sampling design, weights and variability; those features and the observations determine the estimate, beyond sample size alone. CMS's Appendix N explains the stratified sampling and weighted estimation framework that makes a simple unweighted interpretation inappropriate. ²

The published confidence intervals make the limited precision visible. Their widths are 19.3 and 19.5 percentage points, calculated from the displayed endpoints. The 2025 interval is therefore not narrower simply because the point estimate is lower. These widths are descriptions of rounded published intervals, not an independent re-estimation of the survey's variance. ^{1 2}

The intervals overlap, but that fact is not itself a formal test of the difference between years. Overlap of two individual 95% intervals is not equivalent to a confidence interval for their difference; non-overlap is not a necessary condition for a statistically significant difference. A proper comparison needs the relevant variance information and an appropriate treatment of any dependence between estimates. The Census Bureau explains this statistical distinction in general terms, separate from an NPWT-specific analysis. A formal year-to-year NPWT difference test would require additional statistical information beyond the two Table G2 rows. ³

CMS published a lower estimate accompanied by broad uncertainty. Its significance, persistence and cause remain unresolved by these figures alone.

For a fixed underlying rate, the 95% describes the confidence procedure's long-run coverage, rather than a probability assigned to the particular interval after it has been calculated. In the usual repeated-sampling interpretation, the interval-producing procedure has the stated coverage under its assumptions. NIST explains that general interpretation using a mean; CMS's survey-specific estimation method remains the relevant method for these payment estimates. ⁴

The upcoding estimate is a signal with a different scope

The 2025 report's Table K4 lists DMEPOS NPWT projected improper payments attributed to upcoding of $501,364, an estimated rate of 0.9%, and a 95% interval displayed as (0.8%)–2.5%. The rate's numerator is projected upcoding improper payments and its denominator is projected payments for the DMEPOS NPWT policy group—not all NPWT improper payments. CMS uses parentheses here for the negative lower endpoint. Keep that category's $501,364 distinct from the $6,504,562 total in Table G2. ²

An interval extending below zero reflects the statistical construction of the interval; negative upcoding has no substantive interpretation. Nor does an interval including zero erase the errors found in the sample that underpin a positive estimate. Read this as an imprecise estimate, preserving both the observed category estimate and uncertainty about its magnitude.

An absent NPWT row in another year's displayed upcoding table leaves that comparison unresolved. Treating “not shown” as “none” would create an unsupported baseline. The 2025 table supports a report-year estimate and a reason to check product-to-code fit; establishing when upcoding began or its contribution relative to other errors requires additional evidence. ^{1 2}

For a distributor, the useful question is whether the product description, the code used and the evidence supporting that code match. CMS's NPWT policy article distinguishes an integrated system of pump, dressing sets and separate collection canister from other wound-suction arrangements. It states that products billed as E2402 must have a written Coding Verification Review by the Pricing, Data Analysis and Coding contractor, PDAC, and subsequent publication on the appropriate Product Classification List. The buyer therefore needs evidence for the exact model, not just an internal assessment or a competitor's code. ⁵

The article also assigns disposable wound-suction pumps and their related supplies to A9272 and states that they are statutorily noncovered because they do not meet the DME benefit. This is a DME-benefit distinction, not a finding that every Medicare setting or payment pathway excludes disposable NPWT, or that disposable systems are clinically inferior. Determine the applicable branch from the actual configuration, beyond a marketing label such as “portable NPWT.” No listing or coding status is inferred here for a VEMERIX or third-party model. ⁵

Convert policy requirements into evidence requests

The payment estimate is background. For a US Medicare transaction, the current coverage and documentation sources are more directly actionable. Local Coverage Determination L33821 and its related policy article A52511 describe conditions for NPWT pump coverage, continued coverage, supplies, orders and documentation. They should be checked for the relevant date of service and product configuration rather than treated as a universal policy for every payer or every NPWT setting. ^{5 6}

The ordering distinction is worth preserving. The LCD says a Standard Written Order must reach the supplier before a claim is submitted. Its general language separately addresses items that require a Written Order Prior to Delivery. The current policy article directs readers to the periodically updated list of specified codes subject to face-to-face and WOPD requirements. Check the exact code and date of service against that list; neither a universal pre-delivery rule for all NPWT products nor a blanket exemption follows from the generic text. Other applicable ordering and documentation requirements still need to be met. ^{5 6}

For continued coverage, the sources emphasize evidence of wound evaluation and progress. The policy article calls for at least monthly quantitative documentation of wound characteristics, including length, width, depth and exudate amount, and for month-to-month comparisons using like measurements. A supplier must obtain an assessment of healing progress from the treating clinician; the article distinguishes that communication from a requirement that the supplier personally view the entire medical record for each bill. The records may still be requested to corroborate the claim. ⁵

The four-month boundary is also conditional rather than a casual renewal date. The LCD includes four months of pump use for the most recent wound, including prior inpatient use, among the conditions ending coverage. The policy article describes seeking individual consideration for one additional month at a time through appeals, supported by contemporaneous medical-record information explaining the special circumstances. It does not grant automatic coverage beyond four months because a wound has not completely healed. ^{5 6}

The LCD sets ordinary supply limits, including up to 15 dressing kits per wound per month and up to 10 canister sets per month, with specified documentation for high-volume drainage and additional canisters. These are coverage conditions within this policy, not a treatment schedule for an individual patient and not a universal packaging specification for every market. Clinical use follows the treating team's judgment, applicable labeling and care plan. ⁶

An evidence request can retain these distinctions:

This is a procurement and compliance evidence matrix derived from the distinctions in the cited policies. Apply the underlying policies and obtain the patient-specific coverage determination separately. ^{5 6}

Use CMS examples without turning them into prevalence claims

CMS's NPWT compliance education page provides a concrete documentation example. A supplier submits a claim for an NPWT electrical pump for a patient with a neuropathic diabetic foot ulcer, with records lacking evidence of the specified nutritional evaluation, comprehensive diabetic management and pressure-reduction measures. The example ends in an insufficient-documentation error and recoupment. It illustrates how a claim can lack support despite identifying a wound and supplying a medical record. ⁷

The example is useful for designing a check: does the documentation substantiate the relevant coverage conditions, or merely name the diagnosis? The example supplies a concrete documentation issue, while its frequency in the 45 reviewed NPWT claims remains unspecified. Nor does it establish that insufficient documentation explains a particular share of all NPWT errors. Keep a teaching case separate from a population denominator.

A distributor evaluating a US billing partner could ask how that partner obtains missing supporting records, distinguishes initial from continued coverage and handles requests for additional documentation. The partner's answer should identify responsibilities and evidence, not promise that a particular pump eliminates reimbursement risk. A manufacturer may supply product documentation and training within its scope; the treating clinician's patient record must come from the care actually provided.

The same logic applies to supplier performance data. If a partner offers a low “denial rate,” ask what the denominator includes: initial submissions, final paid claims, appealed claims, payment dollars or claim counts. Ask about the period, payer mix, exclusions and whether corrected resubmissions are counted separately. Those are proposed due-diligence questions, not findings about any named supplier. They are necessary before comparing the partner's metric with a CERT payment-dollar estimate.

Where VEMERIX fits and where due diligence still begins

VEMERIX's portfolio includes an NPWT pump and NPWT dressings. For a distributor or procurement team, the next step is the technical and regulatory documentation for the exact products and intended destination. US authorization, Medicare coding eligibility, supplier enrollment and patient-specific reimbursement each require separate verification beyond portfolio inclusion.

An enquiry should identify the target country, care setting, expected configuration, annual volume, required consumables and the evidence needed for the buyer's regulatory and reimbursement assessment. Request the current labeling, applicable registration or authorization documents, technical specifications, compatibility evidence and relevant quality records through the product and documentation discussion. Review the quality information as one part of supplier qualification, not as a substitute for product-specific or market-specific evidence.

For non-US buyers, the CMS figures are an example of why reimbursement evidence must be separated from device evidence. Another country's payment forecast requires that health system's own evidence.

The decision for the next supplier review

Use the 2024 and 2025 CERT rows with their reporting periods, payment denominators and uncertainty intact. State that the 2025 point estimate is lower. Keep that observation separate from causal explanations, claims of no upcoding and product-level endorsement.

Then ask for the evidence that can actually change the purchasing decision: product and configuration identity, destination-market authorization, code fit, the billing partner's documentation workflow and its own transparently defined performance data. Where the decision concerns Medicare, test that workflow against the current LCD and policy article rather than a simplified sales summary.

For an E2402 purchasing assumption, one concrete stop/go check is whether the exact model has the required PDAC review and listing evidence. Holding that assumption pending verification is our procurement recommendation; the underlying coding condition comes from CMS policy.