Methodology

openFDA NDC Directory (US FDA) - drugs, labelers, product registrations, package NDCs, and marketing start/end dates. Ingested weekly (Mondays). The FDA source itself refreshes daily.

CMS NADAC (National Average Drug Acquisition Cost) - per-unit US price benchmarks keyed to 11-digit NDCs, published weekly by CMS. Ingested weekly (Thursdays, after the midweek CMS release).

CMS Medicaid State Drug Utilization Data (SDUD) - quarterly prescriptions and units dispensed under Medicaid, keyed to 11-digit NDCs, published annually by CMS. This is the dispensing-volume layer. Re-checked monthly; the source itself is released each spring for the prior year.

Raw source snapshots are archived before parsing, and every ingest anomaly is logged rather than silently corrected. Coverage today is US-only; UK, Canada, and Australia are planned.

A drug page represents a drug concept: the combination of generic name, dosage form, and route of administration (e.g. metoprolol succinate · extended-release tablet · oral). Different strengths and package sizes of the same concept are grouped on one page; different dosage forms are separate pages. Mapping to RxNorm concepts and ATC therapeutic classes is in progress and will refine this grouping.

FDA labeler names are free text, so the same company appears under many spellings (“Walgreens”, “Walgreens Co”, “Walgreen's Corporation”). Counting those as separate suppliers would overstate competition - and hide sole-supplier risk. Pillstats resolves name variants into canonical entities in two layers:

Automatic (conservative): names that are identical after normalizing case, punctuation, and trailing legal suffixes (Inc, LLC, Corp…) are linked automatically on every ingest.

Curated:fuzzier matches - typos, word-order differences, “doing business as” filings - are machine-proposed but human-reviewed before merging. Rejected candidates are kept on record so demonstrably distinct companies with similar names (they exist!) are never merged.

Original labeler strings are never discarded: each profile lists the names an entity files under, and supplier counts across the site are canonical-entity counts. Distinct legal entities under common ownership (e.g. numbered distribution subsidiaries) are kept separate; parent-company mapping is a planned, separate layer.

A manufacturer is counted as an active supplierof a drug when it holds at least one US product registration without a marketing end-date. “Sole supplier” means exactly one active canonical entity holds registrations for that drug concept. A registration lapsing is a signal, not proof, of a supply exit - labelers sometimes re-register.

The labeler named on an NDC is the company responsible for the product's marketing - which may differ from the company that physically manufactures it. Facility-level data (FDA establishment registrations and inspections) is on the roadmap.

NADAC measures what US retail pharmacies pay to acquire a drug, surveyed by CMS - it is a benchmark of one specific point in the supply chain. It is not a retail price, not what patients pay, not a manufacturer list price, and not a quote.

Trend charts average per-unit NADAC across all NDCs of a drug concept, carrying each NDC's last-known price forward between repricings so weekly averages aren't skewed by whichever packages happened to reprice that week.

The Medicaid dispensing volume shown on drug and market pages comes from CMS State Drug Utilization Data: prescriptions (and units) reimbursed under Medicaid, both fee-for-service and managed care, summed across all states into a national quarterly figure per drug concept. It is the closest free proxy to sales volume - true all-payer retail volume is proprietary and sold commercially (IQVIA, Symphony), and Pillstats does not use paid data.

Two things this means in practice. First, it is Medicaid only - a large but partial slice of the US market, so a drug used mostly outside Medicaid will look smaller here than its true footprint. Second, CMS suppresses small-count rows for privacy; suppressed state-quarters are omitted from the sum rather than treated as zero, so national totals are honest lower bounds and a quarter can read “n/a” when everything in it was suppressed. Prescriptions are used for cross-drug comparison (always comparable); unit totals are only compared within a single drug concept, never across dosage forms.

Rx and OTC products are both covered and labeled; supply-concentration framing is most meaningful for prescription drugs. Data is only as current as the last weekly ingest. Entity resolution is conservative by design, so some name variants may not yet be merged. Nothing on this site is medical advice; clinical information lives at DailyMed, which drug pages link to.

Link the specific drug or manufacturer page you reference. Underlying data originates from US public sources (openFDA, CMS); Pillstats adds entity resolution, supply-concentration analysis, and price-trend context. For questions about the data, contact hello@pillstats.com.

← Back to search