How the NADAC survey picks its pharmacies, and who answers

In August 2017, CMS told state Medicaid officials how its national drug price survey worked. A contractor mailed requests to about 2,500 randomly chosen retail pharmacies each month, and 450 to 600 of them sent back their invoices, according to the agency's webinar slides². Those responses produced 200,000 to 300,000 usable cost observations a month.

From that voluntary trickle comes the National Average Drug Acquisition Cost, or NADAC, the weekly benchmark of what pharmacies pay for drugs. CMS wrote in a 2023 proposed rule⁷ that "most States use this file" to help set what Medicaid pays pharmacies for a prescription.

How pharmacies are chosen, what they are asked to send, and what happens when too few answer all shape that number. CMS's December 2024 methodology¹ and other agency documents trace the survey from the sample draw to the help desk, and GxRadar's own NADAC records show what changed after CMS responded to swings in participation. For the averaging math itself, see how NADAC is calculated.

Who is in the pool

The survey covers retail community pharmacies, a term federal Medicaid rules define in 42 CFR 447.504⁸ as independent, chain, supermarket and mass-merchandiser pharmacies that dispense "to the general public at retail prices." Mail-order, nursing home, hospital, clinic and government pharmacies fall outside it.

CMS's NADAC questions and responses³ document defines a chain as a group of four or more pharmacies "under the same ownership" with the same name. Franchise pharmacies count as independent. The same document describes the sampling frame as roughly 64,000 pharmacies nationwide. GxRadar's piece on independent and chain pharmacies covers how that population has changed.

Two groups are screened out. Closed-door pharmacies, including self-designated specialty pharmacies that dispense mainly by mail, are excluded "at this time," the methodology says. Invoices showing 340B discount pricing are also removed, and the contractor may call a pharmacy to check when prices fall outside an expected range.

How a pharmacy gets picked

Each month the contractor draws a random sample of independent and chain stores from all states. The methodology says the sample "closely aligns" with the overall pharmacy population on four traits: independent, chain, rural and urban. A pharmacy that is selected is left out of the next three months of draws, which CMS says keeps the sample random because every remaining pharmacy has an equal chance.

Answering is voluntary. CMS's question-and-answer document acknowledges the consequence directly: "submission responses may not correspond to the sample composition." The final rate is a simple average of what arrives, with no weighting for chain or independent status, because CMS found the two groups paid similar prices.

What a pharmacy sends

Each request covers invoices for all covered outpatient drugs bought from all wholesalers and manufacturers during the month. Only four fields feed the calculation: the National Drug Code, unit price paid, invoice date and quantity. CMS estimates the job takes "no more than 30 minutes of non-pharmacist staff time," and a pharmacy can ask its wholesaler to send the data instead. In 2017, CMS said about 80 percent of submissions came electronically or by fax.

Several things are left out. Discounts or rebates that do not appear on the invoice line are not counted. Shipping, warehousing and administrative charges on an invoice are ignored. When a store reports the same drug more than once, only the latest purchase counts, and the lowest price wins if two purchases share a date.

CMS's most recent paperwork estimate, published in a Sept. 1, 2026, Federal Register notice⁶, lists 72,000 respondents a year and 36,000 annual burden hours, or 30 minutes each.

From invoices to a rate

For generics, a rate needs cost observations from the most recent six months, with "no less than five" used where costs are closely aligned. Observations more than two standard deviations from the mean are dropped by default. CMS reports that every generic rate has a margin of error under 10 percent of the mean at 95 percent confidence, with an average of 2.4 percent. In 2017, CMS said NADAC rates existed for about 96 percent of Medicaid claim submissions in a comparison with 2016 fee-for-service data. GxRadar's NADAC explainer covers what the benchmark leaves out.

When participation swings

CMS changed the method in late 2024 after rates bounced, a period GxRadar covers in the 2024 NADAC swings. The December 2024 methodology cites "recent generic NADAC rate volatility occurring as a result of fluctuations in survey participation" and, starting with that month's update, calculates each generic rate as a three-month moving average of monthly results. CMS called the step "temporary" and said it plans to seek stakeholder input on a permanent fix.

GxRadar's monthly NADAC series for 47 common generics, built from the CMS weekly files¹¹, show the change. The table counts how often a drug's monthly rate moved more than 10 percent from the month before.

Monthly figures are GxRadar's medians of the weekly per-unit rates for each drug. Large moves nearly tripled in the two years before the change and fell sharply after it. The data are consistent with CMS's account but cannot by themselves separate the moving average from other market changes. Individual drug histories are on the prices page.

Asking for a review

Pharmacies that think a rate lags their costs can contact the NADAC help desk, listed on CMS's Survey of Retail Prices page⁴. The request form⁵ asks for one drug per submission, the pharmacy's cost and claim details, and whether there is a supply problem. It warns that forms sent "without purchase record or invoice supporting your acquisition cost will not be considered for review."

The help desk checks the invoice against other pharmacies and market data. If a price change is confirmed, the new rate appears in the next weekly file and in the week-to-week comparison file marked "Help Desk." CMS's 2017 slides said such changes take effect on the weekly file date and are not backdated. The help desk does not handle questions about a specific state's claim payments; those go to the state Medicaid agency. GxRadar covers the payment side in how pharmacies get paid for a prescription.

Proposals to make answering mandatory

CMS's current methodology still describes participation as voluntary. A Senate bill introduced on March 11, 2025, S. 927⁹, would require retail pharmacies paid by Medicaid to respond if sampled, back that with civil money penalties, extend the survey to some non-retail pharmacies and require HHS to publish "the monthly response rate to the survey" and its sampling method.

Meanwhile, NADAC's reach is growing outside Medicaid. Under the Consolidated Appropriations Act, 2026¹⁰, pharmacy benefit managers working for Part D drug plans must report each drug's "average National Average Drug Acquisition Cost (NADAC)" to plan sponsors and HHS every year starting July 1, 2028. The benchmark built from voluntary pharmacy responses will then sit beside Medicare's own payment data. More on the benchmarks is in GxRadar's comparison of WAC, AWP, NADAC, MAC, FUL and ASP.