Erlend Faltinsen,1,2 Adnan Todorovac,3 Isabelle Boutron,4 Lesley Stewart,5 Asbjørn Hróbjartsson,1,2 Andreas Lundh1,2,6
In Cochrane reviews, trial authors’ conflicts of interest are often not reported and trial funding information is sometimes missing.1,2 This study assessed (1) the proportion of Cochrane reviews reporting trial funding and authors’ conflicts of interest and (2) whether accessing the main trial publication and searching other information sources could identify additional information on funding, conflicts of interest, and the role of funders and authors.
In a cross-sectional study, 1 index trial was randomly included from the primary meta-analyses of 100 Cochrane reviews (October to December 2020). Two authors independently extracted trial characteristics, funding and conflict of interest information, and the role of funders and authors from the reviews and main trial publications, including conflict of interest disclosure forms. Other sources (eg, trial protocols and registry information) were also searched to retrieve additional information, and the time this took was noted. The proportion of Cochrane reviews and main trial publications reporting trial funding and conflict of interest information and role of funders and authors were calculated, as was the proportion of trials in which additional information was found by searching other sources.
The included trials were published from 1975 to 2020 (median: 2011), and 47 were drug trials. Sixty-eight reviews reported trial funding, and 25 reported trial authors’ conflicts of interest. Accessing the main trial publication led to identification of funding in 16 additional trials and conflict of interest information in 38 additional trials. In trials in which funders or trial authors had conflicts of interest, their roles were sufficiently reported in 20 of 36 (56%) and 20 of 30 (67%) main trial publications, respectively. It took approximately 9 minutes (range, 2-28 minutes) per trial publication to extract information. When searching other sources, additional information on funding was found in 2 trials and authors’ conflicts of interest in 13 trials, and it took approximately 22 minutes (range, 4-97 minutes) per trial to extract information. Trial registries and other publications by trial authors were the information sources that most frequently contained additional information.
One-third of recent Cochrane reviews did not report funding of a randomly selected included trial, and three-quarters did not report trial authors’ conflicts of interest despite the information often being reported in the main trial publication. Review authors should systematically access and read main trial publications and disclosure statements and consider searching other information sources.
1. Turner K, Carboni-Jiménez, Benea C, et al. Reporting of drug trial funding sources and author financial conflicts of interest in Cochrane and non-Cochrane meta-analyses: a cross-sectional study. BMJ Open. 2020;10:e035633. doi:10.1136/bmjopen-2019-035633
2. Roseman M, Turner EH, Lexchin J, et al. Reporting of conflicts of interest from drug trials in Cochrane reviews: cross sectional study. BMJ. 2012;345:e5155. doi:10.1136/bmj.e5155
1Centre for Evidence-Based Medicine Odense and Cochrane Denmark, Department of Clinical Research, University of Southern Denmark, Odense, Denmark, email@example.com; 2Open Patient Data Explorative Network, Odense University Hospital, Odense, Denmark; 3Psychiatric Hospital Esbjerg, Region of Southern Denmark, Esbjerg, Denmark; 4Université de Paris, CRESS, Inserm, INRA, Paris, France; 5Centre for Reviews and Dissemination, University of York, York, UK; 6Department of Infectious Diseases, Hvidovre Hospital, Hvidovre, Denmark
Conflict of Interest Disclosures
Isabelle Boutron, Lesley Stewart, Asbjørn Hróbjartsson, and Andreas Lundh are members of the TACIT Steering Group involved in developing a Tool for Addressing Conflicts of Interest in Trials. Isabelle Boutron is a member of the Peer Review Congress Advisory Board but was not involved in the review or decision for this abstract. No other disclosures were reported.
Andreas Lundh is a co–corresponding author.