Medicaid Drug Payment Spread

For each state and drug: what Medicaid reimbursed pharmacies (State Drug Utilization Data) against what pharmacies paid to acquire the drug (National Average Drug Acquisition Cost). The gap is the spread — the margin embedded in Medicaid’s drug reimbursement.

By Kevin Contino · September 2026 · SDUD, NADAC and MDRP data from data.medicaid.gov, 2025.

Spread by state and drug

The 24 highest-reimbursement drugs nationally, cross-tabulated against every state. Each cell is colored by how far that state’s spread sits from that drug’s own average across all states — a fixed percentage cutoff would compare a drug with a tight national range against one with a wide one on the same scale, which isn’t a fair comparison. White marks a state reimbursing far below what’s typical for that drug; red marks far above. Hover a cell for the dollar detail and the drug’s own average. Faded cells carried under $500K in reimbursement — read their percentages cautiously. Hatched cells have no matched reimbursement at all — that’s a gap in the data, not a value of zero.

States ranked by dollar-weighted spread

Summed across every drug reimbursed in that state in 2025, not just the top 24 above. A handful of high-cost specialty biologics (GLP-1s, injectable biologics) carry most of the dollars and are reimbursed close to — sometimes below — acquisition cost, which is why these totals sit far under the median claim-line spread above. Delaware’s −31% and Oregon’s −6% are driven by a small number of very-high-volume specialty NDCs reimbursed under acquisition cost, not by underpayment across the board — their own median claim-line spread is +85% and +48% respectively.

Method

SDUD 2025 annual file (fee-for-service plus managed care utilization combined), aggregated to state × NDC, joined to the NADAC rate effective closest to and on or before 2025-12-31 for that NDC. Drug names are the FDA product name from the MDRP product-list file — SDUD’s own name field truncates at about ten characters at the source. Package-size and dose-strength variants of the same brand are summed together. Spread = reimbursed dollars − (units × NADAC rate); percent spread is that gap over the acquisition-cost total.

• Unit basis assumed, not row-verified. SDUD “Units Reimbursed” and NADAC “Pricing Unit” are both NDC-billing-unit based (each / mL / gram) and should align per NDC by construction, but this was not checked row by row — a mismatch would inflate or deflate that NDC’s spread.
• Year-end NADAC snapshot, not quarterly-matched. NADAC rates move through the year; this uses each NDC’s last known 2025 rate as a single reference point rather than matching the rate to the SDUD quarter it was billed in.
• 91.3% of reimbursed units matched a NADAC rate by total volume; unmatched NDCs (compounds, some brand-only products, discontinued drugs) are excluded rather than estimated.
• Roughly half of matched SDUD dollars carry no state at all. The source file codes $83.1B of the $164.9B in NADAC-matched 2025 reimbursement to XX, not to a real state — same top-drug composition as the states shown, just unattributed. Every state-level figure and the national +7.1% figure above exclude it; summing the match before that exclusion comes out close to double. Territories (PR) are shown; PS is also excluded.

Source: Medicaid.gov SDUD, NADAC, and MDRP Drug Products files, downloaded 2026-09-08. In active development — the next step is the behavioral-health spending companion piece.