How to Read Medicare Part D Prescriber Data on a Doctor's Profile

Where the data come from

CMS publishes the Medicare Part D Prescribers by Provider and Drug dataset every year. Each row describes one prescriber, identified by National Provider Identifier (NPI), and one drug, identified by brand name and generic name, for one calendar year.

CMS's methodological overview of the Part D prescriber data explains how the file is built:

• The source is Part D prescription drug event records: the records that Medicare drug plans submit to CMS for each filled prescription.
• It covers people enrolled in both stand-alone Prescription Drug Plans and Medicare Advantage plans with drug coverage.
• Over-the-counter drugs are excluded.
• Brand and generic names come from linking each drug's National Drug Code to a drug information database.
• Prescriber names and addresses come from the NPPES NPI registry.

Part D itself is Medicare's drug benefit. Medicare.gov's overview of the parts of Medicare explains that you get it through a separate drug plan or a Medicare Advantage plan with drug coverage. Our Part D guide explains how the benefit works for patients.

What each figure means

A profile on this site shows, for each data year we have loaded, a table with these columns. The definitions follow CMS's methodology.

Prescriptions

The number of Medicare Part D claims for the drug, which CMS counts as original prescriptions plus refills. A patient who fills the same prescription every month for a year adds about twelve claims.

Day supply

The total number of days of medicine dispensed across those claims. One 90-day fill and three 30-day fills both add 90 days.

30-day fills

Because some prescriptions are for 30 days and others for 90, CMS also computes standardized 30-day fills: the days supplied on each claim divided by 30. The methodology notes that values below 1 are counted as 1 and values above 12 as 12. Profiles show the yearly total as a summary card, so you can compare prescribers who use different fill lengths.

Total cost

The methodology defines total drug cost as the amount paid for all the claims, including ingredient cost, dispensing fee, sales tax and any vaccine administration fees, whether paid by the Part D plan, the patient, government subsidies or other payers. It notes that the figure does not reflect manufacturer rebates. It is the total for all patients, not what any one patient paid, and it is not your cost: your share depends on your plan's formulary, tier and coverage stage, as Medicare.gov's Part D costs page explains.

Patients

The number of different Medicare Part D patients with at least one claim for the drug. When that count is between 1 and 10, CMS suppresses it, and the profile shows a dash.

Why small counts are missing

To protect patient privacy, CMS's methodology says any record built from fewer than 11 claims is left out of the file, and other counts between 1 and 10 are suppressed. Two consequences matter when you read a profile:

• The list is not complete. A drug a clinician prescribed only a few times to Medicare patients will not appear. Adding up the rows underestimates the real total.
• No section does not mean no prescribing. Many clinicians who rarely prescribe, or who mostly treat patients without Part D, have no Part D data at all.

Brand and generic names

Each row shows the drug name and the generic name. CMS's methodology describes the drug name as the brand name for trademarked drugs and the generic name as the chemical ingredient. When the two match, the drug was dispensed as a generic. Medicare.gov's page on how drug plans work explains that generic drugs use the same active ingredients and must work the same way as the brand-name drug, and gives an example of plan tiers in which most generic drugs are on the lowest-cost tier.

A prescriber's mix of brand and generic drugs depends on their specialty and patients, and Medicare.gov notes that there may not be a generic version of some brand-name drugs. Comparing two clinicians only makes sense within the same specialty and for similar patients.

What the data do not show

CMS's methodology lists the main limitations, and they apply to every profile on this site:

• Not quality of care. The methodology says the information does not indicate the quality of care provided by individual clinicians.
• Only Part D patients. Clinicians treat many patients who do not have Medicare drug coverage, so the data may not represent a prescriber's whole prescribing pattern.
• Attribution is imperfect. Some prescriptions are listed under an organization's NPI instead of the individual's, and some may be attributed to the wrong prescriber because of errors when the prescription is entered at the pharmacy.
• Past years only. Each release describes one calendar year that has already ended.
• Not the reason for a prescription. The data do not include diagnoses, so they cannot show why a drug was prescribed or whether it was the right choice for a patient.

Why a high count is not a red flag by itself

A long list of prescriptions or a large drug cost usually reflects what a clinician treats and how many Medicare patients they see. A clinician who manages long-term conditions writes many refills of the same drugs, and a clinician whose specialty uses expensive medicines can show high total costs even with few patients. None of this says anything about appropriateness.

If something in the data concerns you, the right step is a conversation, not a conclusion. You can ask your clinician why a medicine was chosen, whether a generic or lower-cost option would work, and how long you should take it.

Part D data and industry payments

Profiles may also show Open Payments records: payments that drug and device companies reported to CMS. Seeing a payment from a company next to prescriptions of that company's drug does not show that the payment caused the prescribing. The two datasets are separate, and CMS's methodology notes that linking Part D data with Open Payments data has to be done carefully because the identifiers and time periods differ. You can look up the payment records yourself in the Open Payments search tool; our guide on looking up payments to your doctor explains how, and our Open Payments guide explains the program.

How T-MedInfo shows Part D data

We load the CMS file as published and show it without changes: the drug names, prescriptions, days' supply, total cost and patient counts for each data year we have loaded, with a tab for each year. Summary cards show the year's total prescriptions, standardized 30-day fills and total drug cost, and next to the table we state that CMS does not publish drugs with fewer than 11 claims for a prescriber and that prescribing data are not a measure of quality. We do not rank clinicians by prescribing, estimate missing values or add drugs that are not in the CMS file. Our methodology page explains how and when the data are loaded.