Read our COVID-19 research and news.

Methodology for FDA adviser and staff investigations

For the “Hidden conflicts?” feature, Science examined potential conflicts of interest in the Food and Drug Administration approval process in three drug categories—cardiovascular/renal, psychopharmacologic, and arthritis. This centered on all 28 drugs in those categories recommended for approval by FDA advisory committees from 2008 to 2014 and later approved by the agency.

Science used available data on pharmaceutical company payments to advisers from two sources—the Centers for Medicare & Medicaid Services Open Payments website for mid-2013 through 2016, the only years available (2017 data were released too late for this report), and disclosures made in scholarly writings in publications available to the public without payment. By excluding disclosures protected behind journal paywalls, this approach likely understates the presence or degree of actual or possible conflicts.

The analysis considered payments from the maker of a drug that was voted on and its competitors. A company was designated a competitor if it markets or is developing a drug designed to treat the same conditions as the product reviewed and approved by the FDA. Subsidiaries were treated as part of the parent company.

Science also looked at the more recent 2015 and 2016 advisory committee meetings in the same three categories to spot check for apparent conflicts among physician advisers, but did not comprehensively analyze those cases because Open Payments data were only available through 2016—insufficient for a full assessment of after-the-vote payments.

The investigation considered all 107 voting advisory committee members who were physicians and not federal employees. We excluded consumer and patient representatives, and academic experts in nonclinical specialties because they were not present in the Open Payments data set. We excluded government employees who by law or policy cannot accept pharma funding.

For a secondary investigation, Science checked the employment status of all FDA staffers who were integrally involved in writing, overseeing, and approving medical reviews for the 28 drugs examined, to determine if these federal employees later moved to pharma companies. We checked FDA websites, LinkedIn, and other public sources. Science excluded staffers who appear to have retired after leaving the agency and FDA specialists who contributed tangentially to the medical review.

Download data for story and sidebar

Science removed a few columns that were not relevant to our examination, and combined a few others to reduce needless complexity. No competitive analysis was completed for Northera and Zontivity, due to the highly specific indications they were approved for.