How we compute things

Source data

Everything here comes from the CMS Open Payments publication released on June 30, 2026. That release covers program years 2019-2025. We do not merge in other datasets, and we do not accept data submissions from companies or clinicians.

The reports are filed by the paying companies, not by doctors. That is why we say "as reported to CMS" wherever figures appear.

The three payment categories

Open Payments separates transfers of value into three categories, and so do we:

• General payments: meals, travel and lodging, consulting fees, speaking fees, honoraria, education, gifts, and royalties.
• Research payments: funding connected to research, often institution-directed and tied to clinical trials. A large research figure usually reflects a study budget, not personal income.
• Ownership and investment interests: equity or ownership stakes a clinician or a family member holds in a reporting company.

These categories are never blended on GlassScript. When a page shows a single headline figure, its category is labeled, and research is always presented separately from general payments.

Why product detail covers 2023-2025

Summary totals cover all of 2019-2025. Product-level and nature-of-payment detail (which drug or device a payment was associated with, and whether it was a meal or a consulting fee) covers 2023-2025. The record-level files behind that detail are very large, so we process them for the three most recent program years, where the detail matters most for decisions being made now. Pages that show this detail carry a note saying so.

Association, not allocation

When a company reports a payment, it can list up to five products the payment was related to. CMS records an association, not an allocation: a $1,000 consulting fee associated with two drugs does not mean $500 for each, and it does not mean the payment was made to promote either one. That is why product figures on GlassScript always say payments "associated with" a product, never payments "for" it.

Percentiles and benchmarks

For each clinician with at least one reported payment in the latest program year (2025), we compute where their total falls among other recipients, three ways:

• within their specialty, nationally
• within their state, across all specialties
• nationally, across all specialties

Two things matter about this method. First, the comparison group is recipients: clinicians with no reported payments are not part of the ranking, so "median" means the typical recipient, not the typical doctor. Second, we phrase percentiles as comparisons ("more than 80% of endocrinology providers"), never as leaderboards. Appearing high on a legal, public list is not an accusation.

Specialty and state medians shown on browse pages come from the same recipient population for the same year.

Limitations

• Name matching. CMS assembles clinician profiles from company reports. Mismatches happen, especially with common names, name changes, and clinicians who practice in several states.
• Disputes in progress. A record can be under dispute at the time of publication and still appear in the data until it is resolved.
• Publication delay. CMS publishes about six months after a program year closes. The most recent year here is 2025.
• Corrections cycles. CMS republishes corrected data on a regular cycle, so figures can change between our refreshes.

Accuracy and corrections

If a record about you looks wrong. The underlying records belong to CMS, not to us. Clinicians can dispute a record through the official CMS review-and-dispute process. When CMS publishes corrected data, our next refresh picks it up.

If we rendered something wrong. If GlassScript shows a figure that does not match the official publication, that is our bug and we want to know: hello@glasscript.com. We fix rendering errors promptly.