Transparency in clinical trial reportingBMJ 2018; 363 doi: https://doi.org/10.1136/bmj.k4224 (Published 09 October 2018) Cite this as: BMJ 2018;363:k4224
- Paula A Rochon, professor of medicine and health policy management & evaluation, and vice president of research1 2,
- Nathan M Stall, geriatrician and research fellow1 2,
- Rachel D Savage, epidemiologist and post-doctoral fellow1,
- An-Wen Chan, Phelan scientist and associate professor1 3
- 1Women’s College Research Institute, Women’s College Hospital, Toronto, ON, Canada
- 2Division of Geriatric Medicine, Department of Medicine, University of Toronto, ON, Canada
- 3Division of Dermatology, Department of Medicine, University of Toronto, Toronto, ON, Canada
- Correspondence to: P A Rochon
Patients, clinicians, regulators, and policy makers rely on clinical trials to evaluate the efficacy and safety of new drugs, vaccines, and devices.1 Since the 1990s, most clinical trials have had industry funders who collaborate with academics for clinical and methodological expertise, recruitment of study participants, and access to scholarly publications.23 More recently, industry has increasingly turned to contract research organizations for efficient oversight of clinical trials. These for-profit entities employ physician-scientists, pharmacists, biostatisticians, managers, and medical writers who take control of the design, conduct, analysis, and reporting of trials.34 This commercialization of the oversight of clinical trials has seen academic physicians shifted to comparatively minor roles.3
In a recent paper in the BMJ (doi:10.1136/bmj.k3654), Rasmussen and colleagues report a cross sectional analysis of 200 phase III and IV trials of vaccines, drugs, and devices published in high impact medical journals between 2014 and 2017 with full industry funding and at least one academic author.5 The investigators also surveyed the lead academic authors to determine their experiences with industry funder collaborations. The study reported that in most trials, both the funder and academic authors were involved in the design, conduct, and reporting. In around half of these trials, data analysis was conducted by the funder or contract research organization without involvement of an academic author, who sometimes had limited or no access to trial data. Of note, only eight (4%) trials had no direct involvement of the funder.
Although the survey of lead academic authors had a low response rate (40%), it had some important findings. Most authors found collaboration with industry beneficial, usually because of the funding, and many (79%) reported signing a trial agreement with a funder. Surveyed authors commented that the contract gave the funder the right to review, comment on, and even approve reporting of the clinical trial. This was of particular concern, as there was evidence of ghost authorship (17/80 studies) and some authors (11%) reported disagreements with the funder about trial design and reporting. Furthermore, several authors (11%) did not have access to the full dataset and others (4%) experienced delayed publication. These responses likely underestimate these challenges to academic freedom, as the survey was limited to successfully published trials.
Medical journals have required disclosure of conflicts of interest since the 1980s, yet its ability to effectively mitigate bias remains very uncertain.6 Experts argue that disclosure simply shifts the problem from “secrecy of bias” to “openness of bias,” leaving readers overwhelmed by the sheer “signal to noise ratio.”7 In the study by Rasmussen and colleagues,5 87% of trials had one or more co-authors who were employees of the funder, and the lead academic author had conflicts of interest with the trial funder in 83% of trials.
Transparent reporting of the roles of funders and investigators could be improved in several ways. Firstly, medical journals should require authors to independently identify their roles in each area of the study, including who bears final responsibility for the design, conduct, and reporting. A comprehensive tool such as the Financial Conflicts of Interest Checklist 2010 could complement the current International Committee of Medical Journals Editors (ICMJE) disclosure form, which is limited to information about financial activities.89
Secondly, medical journals should require the submission and public availability of full trial protocols, including any amendments, covering all elements recommended by the Standard Protocol Items: Recommended for Interventional Trials (SPIRIT) statement.1011 The SPIRIT statement asks for protocols to detail the roles and responsibilities of the study sponsor and funders; access to data, including disclosure of agreements that limit such access for investigators; and a dissemination policy including authorship eligibility guidelines and the intended use of professional writers.12
Thirdly, medical journals should do more to enforce the ICJME’s recommendations for the conduct, reporting, editing, and publication of scholarly work in medical journals.12 Currently, journals require that manuscripts conform to these recommendations, but the study by Rasmussen and colleagues highlights violations. In particular, the ICJME recommendations state that authors should avoid entering into agreements with sponsors that limit their access to data or interfere with their ability to analyze and interpret the data and independently prepare and publish manuscripts. Medical journals could reserve the right to inspect agreements in confidence to ensure that they strictly comply with the ICJME recommendations.
The study by Rasmussen and colleagues highlights challenges to the academic freedom and accountability of clinical trialists in their collaborations with industry. Medical journals must be impartial umpires in the reporting of clinical trials. Without more transparency, we risk undermining the integrity of some of medicine’s most influential research.
Competing interests: We have read and understood the BMJ Group policy on declaration of interests and declare the following interests: none.
Provenance and peer review: Commissioned; not externally peer reviewed.