kvlak limn

Health & Medicine · a limn field note on CMS Medicare Part D

Where the
Prescriptions Go

Medicare Part D spends a quarter of a trillion dollars a year on drugs, and the money does not spread out - it pools. A dozen brands take the largest bites; one therapeutic idea has reshaped the whole branded market in a decade; and the opioids that draw the headlines turn out to concentrate by specialty far more than by state. Nine readings of one river of money, and where it lands.

$275B Part D drug spend, 2023
Claims filled
1.75B
Drug products
3,800+
Opioid share
3.2%
Coverage
2013-2023
Illustrative CMS Medicare Part D Prescribers - by Geography and Drug

The opioid map

Fig. 01 · opioid claims as a share of all Part D claims

Opioids are a small slice of Part D claims nationally, but the rate is not spread evenly. It runs highest across Appalachia and the Deep South and lowest across the Upper Midwest, the Northeast, and the Mountain West - West Virginia (7.2%) sits more than twice as high as Hawaii (2.8%).

Alabama: 6.1% opioid claimsAlaska: 3.6% opioid claimsArizona: 4.4% opioid claimsColorado: 3.7% opioid claimsFlorida: 5% opioid claimsGeorgia: 5.3% opioid claimsIndiana: 5.5% opioid claimsKansas: 4.7% opioid claimsMaine: 4.8% opioid claimsMassachusetts: 3.3% opioid claimsMinnesota: 3% opioid claimsNew Jersey: 3.4% opioid claimsNorth Carolina: 5.1% opioid claimsNorth Dakota: 3.9% opioid claimsOklahoma: 6.3% opioid claimsPennsylvania: 4.7% opioid claimsSouth Dakota: 4% opioid claimsTexas: 4.9% opioid claimsWyoming: 4.2% opioid claimsConnecticut: 3.5% opioid claimsMissouri: 5.4% opioid claimsWest Virginia: 7.2% opioid claimsIllinois: 4.2% opioid claimsNew Mexico: 4.3% opioid claimsArkansas: 6.2% opioid claimsCalifornia: 3.4% opioid claimsDelaware: 4.6% opioid claimsDistrict of Columbia: 3.8% opioid claimsHawaii: 2.8% opioid claimsIowa: 4.3% opioid claimsKentucky: 6.4% opioid claimsMaryland: 3.9% opioid claimsMichigan: 5% opioid claimsMississippi: 5.8% opioid claimsMontana: 4.1% opioid claimsNew Hampshire: 4.4% opioid claimsNew York: 3.2% opioid claimsOhio: 5.2% opioid claimsOregon: 3.6% opioid claimsTennessee: 6% opioid claimsUtah: 4.5% opioid claimsVirginia: 4.6% opioid claimsWashington: 3.5% opioid claimsWisconsin: 3.8% opioid claimsNebraska: 4.2% opioid claimsSouth Carolina: 5.6% opioid claimsIdaho: 4.5% opioid claimsNevada: 4.6% opioid claimsVermont: 3.7% opioid claimsLouisiana: 5.9% opioid claimsRhode Island: 4% opioid claims
Opioid claim rate
  • < 3.5%
  • 3.5 - 4.5%
  • 4.5 - 5.5%
  • 5.5 - 6.5%
  • 6.5% +
No data
State-by-state table (51 rows, sorted high to low)
StateOpioid claim ratePart D spendClaims
West Virginia 7.2% $2.07B 13M
Kentucky 6.4% $4.24B 27M
Oklahoma 6.3% $3.54B 23M
Arkansas 6.2% $2.92B 19M
Alabama 6.1% $4.48B 29M
Tennessee 6% $6.03B 38M
Louisiana 5.9% $3.77B 24M
Mississippi 5.8% $2.83B 18M
South Carolina 5.6% $4.81B 31M
Indiana 5.5% $5.56B 35M
Missouri 5.4% $5.42B 35M
Georgia 5.3% $7.78B 50M
Ohio 5.2% $10.61B 68M
North Carolina 5.1% $8.72B 56M
Florida 5% $21.69B 138M
Michigan 5% $8.96B 57M
Texas 4.9% $19.09B 122M
Maine 4.8% $1.51B 10M
Kansas 4.7% $2.45B 16M
Pennsylvania 4.7% $12.02B 77M
Delaware 4.6% $0.99B 6M
Nevada 4.6% $2.59B 17M
Virginia 4.6% $6.84B 44M
Idaho 4.5% $1.6B 10M
Utah 4.5% $1.89B 12M
Arizona 4.4% $6.36B 41M
New Hampshire 4.4% $1.41B 9M
Iowa 4.3% $2.83B 18M
New Mexico 4.3% $2.07B 13M
Illinois 4.2% $10.14B 65M
Nebraska 4.2% $1.6B 10M
Wyoming 4.2% $0.52B 3M
Montana 4.1% $1.08B 7M
Rhode Island 4% $1.04B 7M
South Dakota 4% $0.75B 5M
Maryland 3.9% $4.81B 31M
North Dakota 3.9% $0.61B 4M
District of Columbia 3.8% $0.42B 3M
Wisconsin 3.8% $5.19B 33M
Colorado 3.7% $4.24B 27M
Vermont 3.7% $0.66B 4M
Alaska 3.6% $0.47B 3M
Oregon 3.6% $3.68B 23M
Connecticut 3.5% $3.21B 20M
Washington 3.5% $5.89B 38M
California 3.4% $29.23B 186M
New Jersey 3.4% $7.64B 49M
Massachusetts 3.3% $6.03B 38M
New York 3.2% $16.74B 107M
Minnesota 3% $4.71B 30M
Hawaii 2.8% $1.23B 8M

Illustrative State opioid rates are curator-authored stand-ins following the known regional gradient. Swap-point: roll up Opioid_Drug_Flag = Y claims per Prscrbr_Geo_Cd. See HANDOFF.md.

Fifty states, one gradient

Fig. 02 · states by opioid claim rate

The map colours space; this is the same data as a shape. Most states pile up just below and around the 4.5% mark - the median state sits at 4.4% - and then the distribution thins into a short, dark tail: only a handful of states clear 6.5%, but those are the ones the map paints deepest. Same buckets, same ramp as the map above.

< 3.5%: 6 states 6 < 3.5% 3.5-4.5%: 20 states 20 3.5-4.5% 4.5-5.5%: 15 states 15 4.5-5.5% 5.5-6.5%: 9 states 9 5.5-6.5% 6.5% +: 1 states 1 6.5% +
State counts by bucket (51 states)
Opioid claim rateStatesShare of states
< 3.5% 6 12%
3.5-4.5% (most common) 20 39%
4.5-5.5% 15 29%
5.5-6.5% 9 18%
6.5% + 1 2%

Illustrative Counts are binned from the same illustrative state rates as the map, on the identical thresholds. Swap-point: bucket the real per-state opioid rates once the Geography rollup lands.

The handful that dominates

Fig. 03 · top brand drugs by Part D spend, 2023

Part D is a river that pools in a few places. A single blood thinner, Eliquis, is the largest line item in the entire program - and the diabetes and obesity drugs stacked behind it are the fastest-climbing money in American medicine. Bar length is gross spend; the ledger beneath adds the tell that spend alone hides - cost per claim.

Eliquis Anticoagulant Eliquis: $18.3B $18.3B Ozempic GLP-1 diabetes Ozempic: $9.2B $9.2B Jardiance SGLT2 diabetes Jardiance: $7.4B $7.4B Trulicity GLP-1 diabetes Trulicity: $5.3B $5.3B Xarelto Anticoagulant Xarelto: $5.2B $5.2B Biktarvy HIV antiretroviral Biktarvy: $4.6B $4.6B Januvia DPP-4 diabetes Januvia: $4B $4B Farxiga SGLT2 diabetes Farxiga: $3.9B $3.9B Trelegy Ellipta COPD inhaler Trelegy Ellipta: $3.6B $3.6B Revlimid Cancer immunomodulator Revlimid: $3.4B $3.4B Humira Autoimmune biologic Humira: $3.2B $3.2B Imbruvica Cancer kinase inhibitor Imbruvica: $2.6B $2.6B
Full ledger (spend, claims, beneficiaries, cost per claim)
DrugClassSpendClaimsBenef.Cost / claim
Eliquis apixaban Anticoagulant $18.3B 17.9M 3.9M $1,022
Ozempic semaglutide GLP-1 diabetes $9.2B 10.1M 1.5M $911
Jardiance empagliflozin SGLT2 diabetes $7.4B 11.8M 2M $627
Trulicity dulaglutide GLP-1 diabetes $5.3B 7.2M 1.1M $736
Xarelto rivaroxaban Anticoagulant $5.2B 7.6M 1.6M $684
Biktarvy bictegrav/emtricitab/tenofovir HIV antiretroviral $4.6B 1.9M 0.16M $2,421
Januvia sitagliptin DPP-4 diabetes $4B 7.9M 1.4M $506
Farxiga dapagliflozin SGLT2 diabetes $3.9B 7M 1.3M $557
Trelegy Ellipta fluticasone/umeclid/vilanterol COPD inhaler $3.6B 6.2M 1M $581
Revlimid lenalidomide Cancer immunomodulator $3.4B 0.35M 0.05M $9,714
Humira adalimumab Autoimmune biologic $3.2B 0.62M 0.1M $5,161
Imbruvica ibrutinib Cancer kinase inhibitor $2.6B 0.3M 0.04M $8,667

Illustrative Spend, claims, and beneficiary counts are illustrative stand-ins anchored to public 2023 CMS ranks (Eliquis led near $18B). Swap-point: national rows in the Geography-and-Drug file, ranked by Tot_Drug_Cst.

Two economies of a blockbuster

Fig. 04 · cost per claim vs claim volume

The same top-of-the-leaderboard spend arrives two completely different ways. Down and to the right sit the maintenance drugs - Eliquis, the diabetes pills - filling tens of millions of prescriptions at a few hundred to a thousand dollars each. Up and to the left sit the specialty drugs - Revlimid, Biktarvy - reaching a fraction as many people at four and five figures a fill. Bubble area is total spend.

Eliquis: $1,022 / claim, 17.9M claims, $18.3B spend Ozempic: $911 / claim, 10.1M claims, $9.2B spend Jardiance: $627 / claim, 11.8M claims, $7.4B spend Trulicity: $736 / claim, 7.2M claims, $5.3B spend Xarelto: $684 / claim, 7.6M claims, $5.2B spend Biktarvy: $2,421 / claim, 1.9M claims, $4.6B spend Januvia: $506 / claim, 7.9M claims, $4B spend Farxiga: $557 / claim, 7M claims, $3.9B spend Trelegy Ellipta: $581 / claim, 6.2M claims, $3.6B spend Revlimid: $9,714 / claim, 0.35M claims, $3.4B spend Humira: $5,161 / claim, 0.62M claims, $3.2B spend Imbruvica: $8,667 / claim, 0.3M claims, $2.6B spend

Reading: a maintenance drug and a specialty drug can post the same $3-5B in spend from opposite corners - one on volume, one on price. Both axes are log scaled to hold the 30x spread.

The two economies, as a table (sorted by cost per claim)
DrugCost / claimClaimsSpend
Revlimid $9,714 0.35M $3.4B
Imbruvica $8,667 0.3M $2.6B
Humira $5,161 0.62M $3.2B
Biktarvy $2,421 1.9M $4.6B
Eliquis $1,022 17.9M $18.3B
Ozempic $911 10.1M $9.2B
Trulicity $736 7.2M $5.3B
Xarelto $684 7.6M $5.2B
Jardiance $627 11.8M $7.4B
Trelegy Ellipta $581 6.2M $3.6B
Farxiga $557 7M $3.9B
Januvia $506 7.9M $4B

Illustrative Cost per claim is spend divided by claims for each illustrative leaderboard row. Swap-point: Tot_Drug_Cst / Tot_Clms per national brand row.

A short head, a long tail

Fig. 05 · cumulative share of spend by cumulative share of products

Part D covers more than 3,800 distinct drug products, but the money does not spread across them. Ranked most-expensive first, the top 10 drugs - 0.3% of the catalog - already take 23% of the spend, and the top 100 take 62%. The curve leaps off the origin and then crawls: the remaining thousands of products are a very long, very flat tail.

Top 0.3% of products = 23% of spendTop 1% of products = 40% of spendTop 2.6% of products = 62% of spendTop 5% of products = 74% of spendTop 13% of products = 86% of spendTop 30% of products = 95% of spendTop 60% of products = 99% of spendTop 100% of products = 100% of spend
Concentration by ranked cut (products vs spend)
Ranked cutShare of productsShare of spendCumulative spend
Top 10 drugs 0.3% 23% $63B
Top 100 drugs 2.6% 62% $171B
Top 500 drugs 13% 86% $237B
All drugs 100% 100% $275B

Illustrative Concentration figures are illustrative, anchored to KFF's 2021 analysis (top 10 = ~0.3% of products / ~22% of gross spend; top 100 = ~3% / ~61%). Swap-point: cumulative Tot_Drug_Cst over the ranked national drug list.

Where the dollars pool by class

Fig. 06 · therapeutic-class share of the branded tier

Group the branded tier by what the drugs are for and one family now dwarfs the rest. Diabetes & obesity - the GLP-1 and SGLT2 drugs and their diabetes cousins - is roughly 28% of branded spend on its own, more than anticoagulants and autoimmune biologics put together. A decade ago it would have been a sliver.

Diabetes & obesity: 28% ($71B) 28% Oncology & specialty: 23% ($58B) 23% Autoimmune & biologics: 17% ($43B) 17% Other brands: 17% ($43B) 17% Anticoagulants: 15% ($38B) 15%
  • Diabetes & obesity 28%
  • Oncology & specialty 23%
  • Autoimmune & biologics 17%
  • Other brands 17%
  • Anticoagulants 15%
Class breakdown (spend, share, representative brands)
Therapeutic familyShareSpendRepresentative brands
Diabetes & obesity 28% $71B Ozempic, Jardiance, Trulicity, Farxiga
Oncology & specialty 23% $58B Revlimid, Imbruvica, Biktarvy
Autoimmune & biologics 17% $43B Humira, Enbrel, Otezla
Other brands 17% $43B Trelegy, Dupixent, and the rest of the branded tier
Anticoagulants 15% $38B Eliquis, Xarelto

Illustrative Class families are a curator-added editorial layer (not a CMS column); the split is illustrative, anchored to the diabetes-drug surge. Swap-point: map Brnd_Name to a class table, then sum Tot_Drug_Cst per family.

Two curves, opposite ways

Fig. 07 · Part D spend and opioid share, 2013-2023

Across the decade the two headline numbers moved in opposite directions. Gross Part D spend more than doubled, from $121B to $275B, with the GLP-1 and SGLT2 class erupting from almost nothing to $34B - a rise steep enough to bend the whole program's total. Meanwhile the opioid share of claims slid from 6.1% to 3.2% as prescribing tightened. Two panels, one shared year axis; never a dual axis.

Gross Part D spend All drugs GLP-1 / SGLT2 class
$275B $34B
Opioid share of Part D claims percent of all claims
3.2%

Illustrative Yearly figures are illustrative, anchored to public trend markers (gross Part D ~$216B in 2021; the diabetes-drug surge; the post-2016 opioid decline). Swap-point: national totals looped over the per-year Geography-and-Drug files, 2013-2023.

The class that ate the decade

Fig. 08 · share of branded spend, 2013 vs 2023

The same families, ten years apart. Anticoagulants and oncology drugs climbed; the old autoimmune giants ceded ground as biosimilars arrived. But one line runs away from the rest: Diabetes & obesity went from 9% of branded spend to 28% - a re-sorting of the whole branded market around a single therapeutic idea.

Diabetes & obesity 9% 28% Oncology & specialty 15% 23% Autoimmune & biologics 24% 17% Anticoagulants 6% 15%
Share shift by class (2013, 2023, change)
Therapeutic family20132023Change
Diabetes & obesity 9% 28% +19 pts
Oncology & specialty 15% 23% +8 pts
Autoimmune & biologics 24% 17% -7 pts
Anticoagulants 6% 15% +9 pts

Illustrative Class shares at two endpoints are illustrative, anchored to the documented GLP-1/SGLT2 surge and the biosimilar erosion of legacy biologics. Swap-point: the per-family rollup at the 2013 and 2023 Geography-and-Drug vintages.

Who writes the opioids

Fig. 09 · opioid share of a specialty's Part D claims

The state map hides the sharper split. Opioid prescribing concentrates by specialty far more than by geography: interventional pain and pain management write opioids on a huge fraction of their claims, while primary care sits near the 4.8% national line. But volume flips the story - family practice alone files 96M claims, so the low-rate generalists still write most opioid prescriptions overall.

Interventional Pain Management Interventional Pain Management: 49% of claims are opioids 49% Pain Management Pain Management: 41% of claims are opioids 41% Anesthesiology Anesthesiology: 22% of claims are opioids 22% Physical Medicine & Rehab Physical Medicine & Rehab: 18% of claims are opioids 18% Orthopedic Surgery Orthopedic Surgery: 14% of claims are opioids 14% Hematology/Oncology Hematology/Oncology: 9.5% of claims are opioids 9.5% General Surgery General Surgery: 8.8% of claims are opioids 8.8% Emergency Medicine Emergency Medicine: 7.9% of claims are opioids 7.9% Nurse Practitioner Nurse Practitioner: 5.6% of claims are opioids 5.6% Family Practice Family Practice: 4.9% of claims are opioids 4.9% Internal Medicine Internal Medicine: 4.1% of claims are opioids 4.1% Dentist Dentist: 2.4% of claims are opioids 2.4%
Rate vs volume by specialty (the counterweight the bars hide)
SpecialtyOpioid ratevs nationalClaims
Interventional Pain Management 49% +44.2 pts 4.2M
Pain Management 41% +36.2 pts 6.8M
Anesthesiology 22% +17.2 pts 9.5M
Physical Medicine & Rehab 18% +13.2 pts 7.1M
Orthopedic Surgery 14% +9.2 pts 11.4M
Hematology/Oncology 9.5% +4.7 pts 5.6M
General Surgery 8.8% +4 pts 6M
Emergency Medicine 7.9% +3.1 pts 4.4M
Nurse Practitioner 5.6% +0.8 pts 58M
Family Practice 4.9% +0.1 pts 96M
Internal Medicine 4.1% -0.7 pts 88M
Dentist 2.4% -2.4 pts 3.1M

Illustrative Specialty rates are illustrative stand-ins reflecting the known ordering (pain and surgical specialties far above primary care). Swap-point: the companion by-Provider file, Opioid_Prscrbng_Rate averaged by Prscrbr_Type.

Methodology

Notes on the Data

This dashboard is built on CMS Medicare Part D Prescribers - by Geography and Drug (Release RY2025 (data years 2013-2023); page figures are illustrative stand-ins), the public file CMS releases each year summarizing every drug paid for under Medicare Part D, aggregated to the national and state level. It reports total claims, total drug cost, beneficiary counts, and flags for opioids, long-acting opioids, antibiotics, and antipsychotics. "Drug cost" is the full amount paid - by the plan, the beneficiary, government subsidies, and any other payer - so it is gross spending, not net of the rebates manufacturers pay back.

What's real, what's a stand-in

Every number on this page is currently Illustrative: hand-authored stand-ins shaped to the real story, not the real CMS values. They are anchored to public figures - Eliquis leading Part D at roughly $18B in 2023, gross Part D spending near $216B in 2021, the top 10 drugs taking about a fifth of all spend, and the well-documented Appalachia-and-Deep-South opioid gradient - but they are not the ingested data. Anything not backed by a real ingest is badged Illustrative and has a documented swap-point in the repo's HANDOFF.md and in src/lib/source.ts. The state map and its distribution, the drug leaderboard, the price-versus-volume scatter, and the concentration curve all trace to the Geography-and-Drug file; the opioid-by-specialty reading comes from the companion by-Provider file.

Two of the readings are editorial constructions on top of the source and stay illustrative even when the underlying spend is real. The therapeutic-class mix and the 2013-to-2023 slope both group brands into families - "diabetes and obesity," "anticoagulants," "autoimmune biologics" - that CMS does not ship as a column; that grouping is a curator layer, so it is labelled as such at the swap-point. We never present curated numbers as real.

What you're not seeing

Three big caveats. First, rebates: gross spend overstates what Medicare nets, and rebates are largest exactly on the brand blockbusters at the top of the leaderboard. Second, suppression: CMS blanks any aggregated cell built on fewer than 11 claims, so the smallest drugs and states are undercounted, not zero. Third, this is Part D only - drugs administered in a clinic and billed to Part B (many cancer infusions, for example) are not here, and neither is anyone outside Medicare. The opioid share counts claims, not milligrams or morphine-equivalents, so it measures how often opioids are dispensed, not how much.

Generated 2026-07-06 00:00 UTC. Source: CMS Medicare Part D Prescribers - by Geography and Drug. Full swap steps: HANDOFF.md.