Abstract
Competing Interest Disclosures Compared With Industry Payments Reporting Among Highly Cited Authors in Clinical Medicine
Daniel M. Cook,1 Kyle Kaminski2
Objective
Medical journals seek to minimize bias and enhance research integrity by requiring authors to disclose competing interests. The International Committee of Medical Journal Editors guidelines require disclosing related financial interactions from the 36 months prior to manuscript submission. Recently, the Affordable Care Act required the health care industry (eg, drug and device manufacturers) to report payments to physicians. This created the Physician Open Payments Database maintained by the Centers for Medicare and Medicaid Services. We compared conflict of interest information from the Physician Open Payments Database with that disclosed by article authors.
Design
Cross-sectional data were extracted from Open Payments and from medical journals and qualitatively coded. We started with the “Most Highly Cited Authors in Clinical Medicine for 2015” list from the Web of Science (n=375). For the authors with institutional affiliation in the United States (n=208), Open Payments data for 2014 were searched. For those authors with reported payments (n=121), we obtained 3 research articles published by them in 2015 or later. We examined the competing interest disclosure statements in the articles for concordance with the Open Payments report. Each article was coded as 1 of 4 categories: (1) full disclosure (total concordance with Open Payments); (2) partial disclosure (some financial ties found in Open Payments declared but not all, eg, Merck declared but Pfizer omitted); (3) declared different relationships (listed industry ties not in Open Payments); or (4) declared nothing to disclose.
Results
A total of 363 articles were coded for 121 authors with 3 articles per author. One hundred sixty articles (44%) declared no competing author interests; 124 articles (34%) had partial (incomplete) author disclosure; 39 articles (11%) disclosed different financial ties than those found in Open Payments; and 40 articles (11%) were coded as having full concordance with the Open Payments. From among the 121 authors, 4 had all 3 sampled articles in full concordance with Open Payments. All 3 articles from 27 authors (22%) found in Open Payments claimed to have nothing to declare.
Conclusions
Most of the highly cited authors sampled have not fully disclosed payments from industry. Our findings are consistent with those of other studies of scientist self-disclosures among conference presenters, clinical guidelines authors, and publications within specific specialties. Competing interest disclosures rely on trust and common understanding about the purpose. Authors may not perceive some industry payments as relevant to a particular article. The new source of payments data allows verification of submitted disclosures, and therefore improves assessment of the medical literature.
1School of Community Health Sciences, University of Nevada, Reno, Reno, NV, USA, dmcook@unr.edu; 2Interdisciplinary PhD Program in Social Psychology, University of Nevada, Reno, NV, USA
Conflict of Interest Disclosures:
None reported.