Medicare Part D 2024: The Most-Prescribed and Costliest Drugs

What this data actually is

Each year, the Centers for Medicare & Medicaid Services publishes a near-complete ledger of what Medicare’s drug benefit — Part D — actually paid for. The file is called Medicare Part D Prescribers — by Provider and Drug, and it holds one row for every combination of a prescribing clinician and a specific drug: how many prescriptions they wrote for it, how many standardized 30-day fills that came to, and the total cost. The 2024 release is 28,023,892 of those rows, covering 1,139,455 prescribers and roughly 1,823 distinct active ingredients sold under 3,055 brand and generic labels.

This is the dataset behind every “most-prescribed drugs” and “what Medicare spends on drugs” headline — but the underlying file is enormous and unwieldy, so most coverage relies on a pre-digested CMS summary. Fonteum aggregates the row-level file directly, which is what lets this page rank drugs two completely different ways — by how often they are prescribed, and by how much they cost — and show how little those two lists overlap.

Two definitions matter before any number. “Total drug cost” is the gross amount Part D and its plans recorded — ingredient cost plus the dispensing fee and any sales tax — before the confidential manufacturer rebates that CMS does not publish at the drug level, so the real net cost of brand drugs is lower than the gross figures here. And Part D covers only Medicare beneficiaries — mostly people 65 and older or with disabilities — so this is the medicine cabinet of older America, not the whole country. Every figure on this page is an aggregate; no individual prescriber is named, ranked, or profiled.

A $226.74B year

In 2024, Medicare Part D recorded 1,479,628,807 prescription claims — counting refills — for a total drug cost of $226.74B. That works out to an average of $153 per prescription, but the average hides everything interesting: the typical generic fill costs a few dollars, while a single month of a brand-name biologic or cancer drug can cost thousands. The central finding of this study is that the prescriptions and the dollars live in almost completely different places.

The clearest illustration is one drug. Eliquis, an anticoagulant, was the costliest drug in Part D at $19.88B — yet it was only the 12th most-prescribed drug in the program. Each Eliquis prescription cost about $864; each prescription of Atorvastatin Calcium, the single most-prescribed drug, cost about $10 — a gap of roughly 84×. The rest of this page is that gap, drawn out across drugs, prescribing specialties, and states.

The most-prescribed drugs are cheap generics

Rank Part D by sheer volume and the list reads like a primary-care formulary. Atorvastatin Calcium — a cholesterol statin — leads with 72,474,420 claims, ahead of blood-pressure generics, levothyroxine for the thyroid, gabapentin, and the acid reducers omeprazole and pantoprazole. Almost every drug in the top 25 is a generic that costs Medicare only a few dollars per fill. These are the workhorses of chronic-disease management for older Americans: prescribed constantly, and individually cheap.

The costliest drugs are brand-name only

Re-rank the same file by dollars and a different program appears. The top of the cost list contains no generics at all. It is led by Eliquis ($19.88B), then a wall of brand-name diabetes drugs — the GLP-1 agonists Ozempic, Mounjaro, and Trulicity and the SGLT2 inhibitors Jardiance and Farxiga — the anticoagulant Xarelto, the COPD inhaler Trelegy, and a long tail of biologics and oral cancer drugs. Some of those cancer and immunology drugs appear with only tens of thousands of prescriptions but billions in cost, because a single fill can run into five figures.

Two prescribing economies in one program

Split every prescription into whether it was dispensed as a generic or a brand, and the divide is stark. Generic-dispensed drugs filled 1,125,888,795 prescriptions — 76.1% of the total — but accounted for only $22.39B, or 9.9% of the dollars. Brand-name drugs are the mirror image: 353,740,012 prescriptions, 23.9% of the volume, but $204.36B — 90.1% of the spending. The U.S. generic-substitution system works exactly as designed on volume; the cost pressure sits entirely in the brand book.

“Generic” = the brand name equals the generic name in the CMS file (the standard generic-dispensing convention); “brand” = the two differ. The two classes partition every 2024 row, so claims and cost sum exactly to the program totals.

The GLP-1 surge — and why Ozempic is here

No drug class better explains the cost story than the GLP-1 receptor agonists — Ozempic, Mounjaro, Trulicity, and their siblings. Together they accounted for $24.57B of Part D spending in 2024 across 19,591,489 prescriptions — about 10.8% of the entire program’s drug cost on one class of medicine. A common question follows: does Medicare cover Ozempic? It does — but only for type 2 diabetes. Federal law excludes drugs used for weight loss from Part D, so the same molecule is covered when it is prescribed to manage blood sugar and excluded when it is prescribed to lose weight. The dollars here are the diabetes book, and they are still enormous.

Who writes the prescriptions

Primary care dominates the volume: family practice and internal medicine top the list. But the most striking finding is who sits third. Nurse Practitioners are the 3rd highest specialty by prescription volume and the single highest by drug cost — $37.48B, ahead of every physician specialty including internal medicine and cardiology. The rise of advanced-practice clinicians as front-line prescribers, which the Medicare enrollment data shows is now the largest provider type in the program, is visible directly in what gets prescribed.

Where the prescriptions are written

Part D volume tracks where Medicare beneficiaries and their prescribers are, with CA leading at 129,565,592 claims and $21.26B in cost. Read this as a count, not a per-capita rate: populous states with more older residents naturally see more prescriptions. The geography is most useful as a denominator for local reporting — pair a state’s totals here with its enrolled-beneficiary count to ask whether spending per beneficiary is unusually high or low.

Methodology & reproducible SQL

Every figure is a direct aggregation over the cms_part_d_prescribers table — the operational projection of the CMS Medicare Part D Prescribers “by Provider and Drug” 2024 public-use file (28,023,892 prescriber-by-drug rows; RLS Pattern B, public read). Because the table is far too large to aggregate inside a single web request, the counts are pre-computed in five Postgres materialized views (one headline row plus most-prescribed, costliest, by-specialty, and by-state views); the page reads those bounded views and falls back to a committed snapshot of the same figures when the views have not yet been refreshed. No unpaginated row select is ever issued against the multi-million-row table. Method version part-d-prescribing/v1. The headline numbers reproduce with:

-- Headline: claims + cost, generic vs brand (one pass)
SELECT
  count(*)                                                              AS rows,
  sum(total_claims)                                                     AS claims,
  round(sum(total_drug_cost))                                          AS drug_cost,
  sum(total_claims) FILTER (WHERE lower(brand_name)=lower(generic_name)) AS generic_claims,
  round(sum(total_drug_cost) FILTER
        (WHERE lower(brand_name)=lower(generic_name)))                 AS generic_cost
FROM cms_part_d_prescribers
WHERE data_year = 2024;

-- Most-prescribed drugs (top 25 by claims)
SELECT brand_name, generic_name,
  sum(total_claims)               AS claims,
  round(sum(total_drug_cost))     AS cost
FROM cms_part_d_prescribers
WHERE data_year = 2024
GROUP BY brand_name, generic_name
ORDER BY claims DESC
LIMIT 25;

-- Costliest drugs (top 25 by drug cost)
SELECT brand_name, generic_name,
  sum(total_claims)               AS claims,
  round(sum(total_drug_cost))     AS cost
FROM cms_part_d_prescribers
WHERE data_year = 2024
GROUP BY brand_name, generic_name
ORDER BY cost DESC
LIMIT 25;

A few decisions shape the counts. We scope every aggregate to data_year = 2024. A drug is counted as generic-dispensed when CMS records the brand name equal to the generic name — the standard convention for a generically filled prescription — and brand otherwise. Drug rankings group on the published brand_name, generic_name pair, so a molecule sold under several brands appears as several rows. “Total drug cost” is gross of manufacturer rebates, which CMS does not publish at the drug level. Read how every Fonteum figure is sourced on the sources page.

Limitations

• One data year, Medicare only. All figures are CMS Part D data year 2024. Part D covers Medicare beneficiaries — mostly people 65+ or with disabilities — not the whole U.S. population, so “most-prescribed” means most-prescribed to seniors.
• Small-volume suppression. CMS removes any prescriber-drug row with fewer than 11 claims before publishing, and suppresses the beneficiary count where it is under 11. Totals therefore slightly understate the true universe, and rare drugs and very low-volume prescribers are absent by design.
• Brand vs generic is inferred. The file carries no explicit brand/generic flag; we infer it from whether the brand and generic names match, the standard CMS approach. A handful of edge cases (authorized generics, combination products) can land on the other side of the line.
• Loading in progress at capture. The 2024 release is being ingested into Fonteum’s database; the committed fallback figures are a preliminary capture and the live page reflects the final totals once the source views are refreshed. The relative findings — generics dominate volume, brands dominate cost — are stable throughout.
• Aggregate-only, no quality claim. Fonteum does not name, rank, or rate any prescriber here, and prescribing volume is not a measure of a clinician’s judgment or the appropriateness of any prescription.

Frequently asked questions

Related research

• Open Payments: what drug and device companies pay doctors — the money flowing the other way, from manufacturers to clinicians.
• Who is enrolled in Medicare (PECOS) — why nurse practitioners are now the largest provider type.
• Hospital ownership & margins — the financial backdrop of the Medicare program.
• NPI lookup — resolve any prescriber NPI to its NPPES, PECOS, and OIG record.
• Sources & methodology index.

Sources

1. U.S. Centers for Medicare & Medicaid Services — Medicare Part D Prescribers, by Provider and Drug (data year 2024). data.cms.gov
2. CMS — Medicare Part D Prescribers methodology and data dictionary. methodology (PDF)
3. CMS — Medicare Prescription Drug Coverage (Part D) program overview. cms.gov · Part D

Fonteum Research · Published 2026-06-12 · Reviewed by Jennifer Montecillo, MD, medical reviewer. Non-practicing medical reviewer. · Data source: CMS, public domain under 17 U.S.C. §105 · Published free under CC BY 4.0 · Method part-d-prescribing/v1.