MFP Effectuation for Manufacturers: The 14-Day Refund Operation Behind Medicare's Negotiated Prices

What MFP effectuation means for a manufacturer

MFP effectuation is the manufacturer's legal duty to make sure every pharmacy that dispenses a selected drug to a Medicare beneficiary actually gets the maximum fair price, either by selling at that price or, far more often, by paying a refund within 14 days of receiving the claim data from CMS's Medicare Transaction Facilitator (MTF). In practice it is a claim-by-claim data operation: receive a file, decide the refund for each line, identify 340B claims, pay or authorise payment, return a response for every claim and be able to prove all of it later.

The first ten drugs have been running this process since 1 January 2026. On 1 January 2027 the 15 drugs from the second negotiation cycle join them, including Ozempic, Rybelsus and Wegovy, Trelegy Ellipta, Janumet, Tradjenta and Linzess. CMS has already selected 15 more for 2028, the first cycle to include Part B drugs. For a growing number of trade, gross to net and market access teams, MFP effectuation is moving from a policy topic to a weekly operating rhythm.

The legal summaries on page one explain the rules. The actuarial analyses explain the pharmacy economics. This guide covers the part neither does: the data work inside the manufacturer, with a worked example, a readiness checklist for the 2027 cohort and the reporting leadership will ask for. We are a data and analytics firm, not a law firm, so confirm legal positions with your counsel.

MFP effectuation in numbers

How maximum fair price effectuation works through the MTF

The flow is the same for every selected drug, and each step creates a record the manufacturer has to keep. Based on CMS guidance as summarised by Covington and the CMS contractor agreements for the MTF:

1. The pharmacy dispenses and bills the Part D plan at a rate built on the MFP, so it is often paid less than it spent buying the drug at or near wholesale acquisition cost (WAC).
2. The plan's claim flows through CMS to the MTF Data Module (MTF DM). Milliman notes the plan side has about 7 days to submit, which is why pharmacies experience a total wait of around 21 days even when everyone is on time.
3. The MTF DM sends the manufacturer a file of claim-level data elements, together with empty payment element fields for the manufacturer to complete. Receipt of this file starts the 14-day prompt MFP payment window.
4. The manufacturer decides each claim. For every line it returns payment elements stating how the pharmacy got access to the MFP: a refund paid, the MFP made available by another route such as a sale at MFP, or no refund required in the manufacturer's view, with the amount where one is due.
5. Money moves. Manufacturers using the MTF Payment Module (MTF PM) authorise the MTF to pay the dispensing entity by day 14. Manufacturers paying another way must make sure the refund reaches the pharmacy by day 14 themselves.
6. Adjustments follow. Reversals, corrections and 340B claims identified after payment are handled later, and the MTF PM keeps a ledger that offsets future refunds to the same dispensing entity.

Two rules shape everything else. First, the clock does not stop for uncertainty: Covington's summary of CMS guidance is explicit that a manufacturer may not delay a refund because a claim's 340B status is unclear. Second, CMS judges compliance case by case using the pharmacy's invoice amount, the manufacturer's documentation and the refund actually paid, and starts with a Notification of Potential Noncompliance. Both rules reward a manufacturer that can reproduce any refund decision from its own data on demand.

Choosing the refund amount: standard default or actual acquisition cost

Standard default

The MTF DM calculates the SDRA as the drug's WAC minus the MFP, and CMS treats a manufacturer paying it as having refunded enough. It is simple to run because the amount per unit is known in advance. Milliman's October 2026 analysis points out the trade-off: it refunds the full list-price spread whatever the pharmacy actually paid, so pharmacies that buy below WAC gain margin and the manufacturer pays more than the strict minimum.

Acquisition cost

Manufacturers may instead refund the difference between the pharmacy's actual acquisition cost and the MFP. That needs purchase prices by pharmacy, product and date: wholesaler sales data, chargeback history and contract prices, joined to each claim. It lowers the refund, but every line must be defensible if a pharmacy complains, and Milliman warns that pharmacies did not know which method each manufacturer used until the first payments arrived in late January 2026.

Commercial effect

Milliman argues the refund method drives pharmacy margin more than the negotiated price itself, that community pharmacies may gain under SDRA while long-term care and health systems can be squeezed, and that one manufacturer's choice can shift dispensing incentives across a whole therapeutic class. Its September 2025 survey data had nearly one in five pharmacists saying they would not stock MFP drugs over cash flow. Model the method before you choose it, by channel, not just in total.

The data layer an MFP effectuation team needs

The MTF moves files and money. It does not run your gross to net forecast, reconcile refunds to your ledger or explain to the head of trade why liability spiked in one channel. That work sits with the manufacturer, usually across trade operations, gross to net, government pricing and market access.

1. An MTF claims table with full status history. Every line from every MTF DM file lands with the file receipt date, which starts the 14-day clock, and every later event: payment element returned, refund authorised, paid, reversed, adjusted through the ledger, disputed. Store the file as received as well as the parsed table, so any response can be traced to its source.

2. A price table by NDC-11 and date. WAC and MFP with effective dates, so any refund can be recomputed exactly as paid. A WAC change during a quarter changes the SDRA, and the refund must use the price in force for that claim.

3. Purchase data by dispensing entity. Needed if you refund on actual acquisition cost, and useful even under SDRA: wholesaler sales, chargebacks and contract prices joined to the pharmacy on the claim. This is the same data your chargeback management process already holds, so reuse it rather than building a second copy.

4. 340B identification. CMS asks manufacturers to describe how they identify 340B claims and to keep documentation for each one. Covered entity registrations, claims flagged by pharmacies, contract pharmacy relationships and, from 1 January 2027 for pilot drugs, 340B rebate pilot claims all feed one match. Because a refund cannot wait for this answer, the table must also track claims refunded first and identified as 340B later, and the ledger offset that recovers the duplicate.

5. Reason codes for every non-refund. A short fixed list, agreed with counsel, for "MFP made available another way" and "refund not required". Free-text notes cannot be counted, reported or defended.

6. Links to finance. Refund liability by product and month, paid against accrued, feeding the gross to net waterfall and the government price reporting team, since Covington flags the effect of MFP sales on government price reporting, including ASP.

If you already run a pharma commercial data warehouse, this is a new subject area inside it with its own access controls. If you do not, MFP effectuation is a contained, high-stakes place to start one.

A worked example: one month of MFP refunds

The figures below are illustrative, chosen to show the arithmetic. They are not the real WAC or MFP of any selected drug.

Suppose a drug has a WAC of $1,000 per 30-day pack and an MFP of $400. The standard default refund amount is $600 per pack. In one month the MTF DM delivers files covering 50,000 claims, one pack each, so the gross refund exposure at SDRA is $30 million.

Now change one assumption. If the manufacturer refunds on acquisition cost and the average pharmacy bought at 3 percent below WAC, $970, the refund per pack falls to $570 and the month's refunds to $26.39 million, about $1.39 million less. That saving only exists if the purchase price for every one of 46,300 claims can be shown on request. It is also money that leaves pharmacy margins, which is exactly the channel effect Milliman warns about.

Leadership will ask three questions about this month: how many claims missed the 14-day window (the answer should be zero, listed by name if not), how much refund liability is accrued but not yet paid at month end, and how much has been recovered through the ledger for late 340B identification and reversals. Each answer comes straight from the tables above.

MFP effectuation readiness checklist for the 2027 cohort

The MFP effectuation dashboard leadership will ask for

• 14-day compliance

  Claims received, decided and paid by day, with days remaining on each open batch and any breach listed by claim, never averaged away.

  • Clock from file receipt
  • Oldest open claims first
  • Breaches named

• Refund liability

  Paid, authorised and accrued refunds by product and month, tied to the ledger and the gross to net forecast.

  • Paid vs accrued
  • True-up history
  • Feeds month-end close

• 340B nonduplication

  Claims identified as 340B before and after refund, the evidence behind each, and amounts recovered through the MTF ledger.

  • Identified once
  • Evidence stored
  • Recoveries tracked

• Non-refunds by reason

  Every claim without a refund, by coded reason, product and dispensing entity, so a pattern is visible before a pharmacy complains.

  • Coded reasons only
  • By channel
  • Disputes to outcome

• Channel and access signals

  Claims volume by pharmacy type and region against expected demand, as an early warning if pharmacies stop stocking the drug.

  • Community vs LTC vs health system
  • Month on month
  • Linked to access KPIs

• Audit pack

  For any claim or period, the file as received, the price used, the decision, the payment and any later adjustment, produced from the same tables.

  • Reproducible on demand
  • One source of truth
  • Retention rules applied

Common mistakes, and when not to build anything new

These are the failure points we would expect in any new claims-level feed, and the public reporting on 2026 suggests they are real:

• Treating the MTF as the system of record. It routes data and money. Your finance, compliance and dispute answers still have to come from data you control and can recompute.
• Holding refunds for 340B certainty. CMS guidance does not allow it. Refund on time and recover through the ledger with documentation.
• Pricing on the wrong date. A WAC change mid-quarter alters the SDRA. Use the price in force for the claim, and keep the history.
• Choosing a refund method on total cost alone. Milliman's analysis shows the method moves margin between pharmacy channels and can affect stocking. Model it by channel before deciding.
• Letting the data drift. MTF claims data is for effectuation. Keep it in its own governed zone with role-based access, away from targeting and segmentation models, and confirm the permitted uses with counsel.
• Building separate pipes for 340B and MFP. The two programmes test the same unit. One model, one match, one evidence trail.

When not to build: if none of your products is selected and none is a likely candidate in the next cycles, you do not need an effectuation layer yet. If you use the MTF PM, refund at SDRA and already have a vendor that delivers reconciled MTF, price and 340B data into your warehouse, add the dashboard rather than rebuild the pipes.

Greenwolf builds this kind of data layer for pharma commercial, trade and market access teams: pipelines, matching logic, the warehouse model and dashboards, on the cloud you already run. Our pharma commercial analytics service page explains how an engagement runs, the market access analytics guide covers the payer data next to it, and the specialty pharmacy analytics hub covers dispense and patient services data.

Questions about MFP effectuation