How to reconcile COI disclosures against Open Payments

Your attestation records what each person reports. Open Payments records what companies reported. Putting the two side by side once a year turns a collection exercise into a checked record, and most of the work is mechanical.

What you need before you start

You need a roster keyed to 10-digit NPIs, the program year you are checking, and the three files CMS publishes for that year from the Open Payments data pages: general payments, research payments, and ownership or investment interests. Open Payments covers physicians, physician assistants, nurse practitioners, clinical nurse specialists, certified registered nurse anesthetists, certified nurse-midwives and teaching hospitals. In this rule, physician includes dentists, podiatrists, optometrists and chiropractors, by reference to section 1861(r) of the Social Security Act (42 CFR 403.902). Staff outside those types, such as most basic scientists, will not appear in it at all, and that is expected.

Step 1: match the roster by NPI, not by name

Each published file carries the recipient's NPI (the general and research files call the column Covered_Recipient_NPI). Match on that. Names fail in both directions: two clinicians share a name, or one clinician appears under a maiden name or a middle initial. The research file also names up to five principal investigators per payment, each with an NPI, so check those columns too. Set aside anyone you cannot match and resolve them by hand before going further.

Step 2: fix the program year and the publication

A program year is a calendar year of payments. CMS publishes it on or by June 30 of the following year and refreshes it in January with corrections made in the meantime (CMS User Guide for Covered Recipients, section 1.1d). Prior years are republished too, so a record can read differently in February than it did in July. Write down which publication you compared against. The dates your office plans around are on the 2027 calendar.

Step 3: total by company and category

Add up each person's records by reporting company and by nature of payment (consulting, speaking, food and beverage, travel, research, ownership and so on). Then map those categories to the ones your own policy uses. This is the step that makes the two records comparable. It is also where most spreadsheet versions go stale, because the mapping is rarely written down.

Step 4: compare, and sort the differences

Each person lands in one of four places:

• Matched. The disclosure and the record agree.
• On the record, not on the disclosure. A company reported something the person did not list. Unreconciled until the person explains it.
• On the disclosure, not on the record. Usually normal. Open Payments covers only drug, device, biological and medical supply makers and group purchasing organizations, and excludes small payments below a floor CMS sets each year (42 CFR 403.904(h)). An indirect payment is not reported when the company did not know who received it. Some research payments are reported on time but held back from publication until FDA approval or clearance, or four years after payment, whichever comes first (42 CFR 403.910).
• Different amounts. Same company, different total. Often a period or attribution question.

Why the two records differ for ordinary reasons

• The periods differ. Open Payments is a calendar year. Many COI policies, and the PHS research rule, look back twelve months from the date of disclosure.
• The research file names principal investigators even when the payment went to the institution or a research site. The person may not think of it as theirs.
• Payments a company made to a third party at the person's request are reported under the person's name.
• A payment made to a group is divided among the people it was for (42 CFR 403.904(c)(4)), so a person can carry a small amount from a shared meal they barely remember.
• Ownership interests held by an immediate family member are reported too (42 CFR 403.906).
• Your policy threshold may be higher or lower than the federal floor.

And sometimes the record is simply wrong. That is what the next step is for.

Step 5: resolve each difference with the person

Send each person their own differences, line by line, and ask for one of three answers: confirm it, explain it, or say the record is wrong. Keep the answer with a date and the person's name on it. If the record is wrong, the person can dispute it in the CMS Open Payments system during the review and dispute window (at least 45 days, April 1 to May 15 in the standing cycle), or later in the year with the correction reflected in a later publication (cms.gov). Only the covered recipient can file it, directly or through an authorized representative they nominate in the CMS system. The office's job is to put the record in front of them in time and note what happened.

Step 6: report to the committee

The committee needs three numbers per department, reconciled, explained and still open, with the names behind each one. Keep last year's comparison as it stood, with the publication date, because the federal record will move under it.

Doing it by hand, and when not to

All of this can be done with the CMS download and a spreadsheet, and for a small medical staff that is reasonable. It gets heavy as the roster grows: three files, a category mapping, a re-run after each refresh, and a dated answer from each person. If you are choosing software for it, the COI software checklist lists what to require. To see what the public record already holds for a hospital's medical staff, look it up in the hospital directory. For the wider monitoring program, see how a hospital compliance program should monitor Open Payments.