What drug and device companies pay doctors

The national picture

1,693,971

Doctors and practitioners paid

People CMS holds a payment for. Not a count of the profession.

$17.3 billion

General payments

$17,307,111,322

Paid outside a research agreement, 2019–⁠2025. CMS’s record lists the kind of each.

$610.6 million

Research payments

$610,582,922

Money for studies. A different thing, never added to the figure above.

$2.2 billion

Ownership interests

$2,201,126,553

Invested by doctors or their immediate families in the companies themselves.

87,393,359

Payments recorded

The count is what separates one large payment from three hundred small ones.

2,808

Payments disputed

Contested by 843 recipients. CMS publishes a disputed payment either way and counts it in the totals.

3,063 of the profiles above are ones CMS believes it holds more than once for the same person, so a count of doctors is a count of records. Figures come from the Open Payments file taken on September 25, 2026.

What stands out

Seven findings. From the Open Payments and Medicare Part D files, each one assembled from values a query returned; both are in the source block at the foot of this page.

• Brand-name drugs were 12.8% of Medicare Part D claims in 2024 and 71.7% of the $288.4 billion Part D paid for drugs. Medicare Part D →

  Both shares are of the whole file. Claims CMS classes as neither brand nor generic, and counts it withholds, are in the totals and in neither numerator.

• The 100 clinicians paid most by drug and device makers received $3.0 billion in general payments over 2019–2025, 17.2% of the $17.3 billion paid to 1,693,971 clinicians. The 100 clinicians paid most →

  Each total is one clinician across every program year in the file. A payment is a record that money moved, not a finding about why.

• Physician assistants and advanced practice nurses were 31.0% of the clinicians paid by drug and device makers over 2019–2025 and received 5.7% of the money; physicians received 85.0%. Physician assistants and nurse practitioners →

  Shares of everyone paid and of every dollar, including the recipients CMS records with no specialty, who are in the totals and in no group.

• Opioids were 5.2% of Medicare Part D claims written by prescribers in Alabama in 2024, the highest share of any state, against 3.5% nationally. Alabama →

  A share of claims, not of patients or of doses. It says what was dispensed, not whether it should have been.

• General payments from drug and device makers to clinicians fell 30.7% in 2020, to $1.7 billion, and were back above 2019's $2.4 billion by 2022, at $2.7 billion. Program year 2020 →

  The file records that the amount fell and when it recovered. It does not record why.

• Clinicians disputed 2,808 of the 93,978,144 payment records reported about them over 2019–2025, about 3 in every 100,000; 843 clinicians disputed at least one. Disputed records →

  A dispute is a clinician saying a record is wrong. An undisputed record is one nobody disputed, which is not the same as one somebody checked.

• CMS withholds the count of antipsychotic claims for patients 65 and over for 538,626 prescribers in the 2024 file, 38.0% of the 1,416,883. What the file withholds →

  CMS suppresses small counts, and counts that could be worked out from ones it publishes, and does not say which applies to a given row. Either way the measure cannot be ranked or summed for those prescribers, and this hub does not.

By program year

A program year is published the following summer and revised afterwards, so 2025 will change. The drop in 2020 is the pandemic year.

Where the doctors are

The state is the practice address CMS holds, which is not necessarily where the work happened. No figure here is divided by a population: a doctor is not distributed over one, and the ratio would mean nothing. 435 doctors carry no state at all and are in the national total above and on no state page.

By specialty

CMS's own grouping, which is broader than a specialty a patient would name — every physician and surgeon sits in one group. The finer classification is on each doctor's page.

The largest recipients

Each total is what manufacturers reported paying the doctor, and CMS’s record for the doctor, linked from each page, lists the kind of each payment. Nothing in this record says a payment was improper, and nothing on this site will say so for it.

Research, and ownership

A research payment is money for a study, which may pass through a hospital or a university and reach the named doctor in no part. An ownership interest is a stake in the company itself, held by the doctor or an immediate family member. Neither is added to the general payments above, because they are not the same kind of thing.

Analysis

Each brief is prose written around values a query returned. Every figure in one is re-checked against its query before it can merge, so a number in the text cannot drift from the data behind it.

• Five stars, by who owns the hospital: 44.6% of rated VA hospitals, 5.0% of for-profits

  Of the hospitals CMS rates, five stars go to 44.6% of VA hospitals, 13.2% of private nonprofits and 5.0% of for-profits.

  American Factbook Research · October 9, 2026

• Where Medicare Advantage choice is thin

  CMS lists Medicare Advantage plans in 3,162 counties for contract year 2027. The median county has 42 of them; Hillsborough, Florida, has the most, 124. At the other end, 346 counties have fewer than ten — 10.9% of the counties CMS lists.

  American Factbook Research · October 3, 2026

• Medicare's prescribing totals are complete; the breakdowns beneath them are not

  Prescribers billed Medicare Part D for 1,709,828,271 prescriptions and refills in 2024. There were 1,416,883 of them. Both of those figures are complete: CMS withholds a prescriber's total from nobody, and not one of the 1,416,883 rows carries no total.

  American Factbook Research · September 22, 2026

• CMS documents what its two suppression markers mean, and the counts do not fit

  Medicare's Part D file names every prescriber in it and then, for some of them, declines to say what they prescribed. The blanks are not errors and not zeroes. They are deliberate, they are marked, and CMS documents why they are there.

  American Factbook Research · September 22, 2026

• Doctors contest almost nothing, and the record says why that is not reassurance

  Drug and device manufacturers have reported 93,978,144 separate transactions with doctors since 2019 — general payments, research funding and ownership interests together, which is the population a dispute can be raised against. Doctors have contested 2,808 of them.

  American Factbook Research · September 22, 2026