Publishing “Invisible” and “Abandoned” Clinical Trials: A Commitment for CTS

“Search for the truth is the noblest occupation of man; its publication is a duty” Madame de Stael (1766–1817).1

Both governmental agencies and nongovernmental organizations have placed increasing pressures on pharmaceutical companies and other sponsors of clinical research to improve the ability of investigators to access clinical trial data. Indeed, in a July 15, 2013 editorial, the editors of the New York Times described the efforts of individuals from the Cochrane Collaborations, an international network of experts based in Oxford, England, to obtain data from eight studies of the anti‐flu drug Tamiflu from Roche pharmaceuticals. After Cochrane researchers reported that Tamiflu had not shown benefits and a report in the British Medical Journal reported that Roche had hired ghostwriters to author Tamiflu articles, Roche capitulated and agreed to release detailed clinical data to outside investigators upon request. The Times editors opined that data sharing between pharmaceutical companies and investigators should become a standard industry practice. However, efforts to insure the dissemination of timely and accurate data from industry‐sponsored clinical research studies have been fraught with roadblocks.

In 2010, a group of industry organizations including the International federation of Pharmaceutical Manufacturers and Associations, the European Federation of Pharmaceutical Industries and Associations, the Japan Pharmaceutical Manufacturers Association, and the Pharmaceutical Research Manufacturers of America published a joint position paper on the public health benefits associated with making clinical trial results widely available through submission for publication. They proposed that: all trials should be published regardless of outcome, submissions should be timely—within 12 months and no later than 18 months after trial completion, manuscripts should be submitted to peer‐reviewed journals that are indexed by online databases, authorship should be consistent with the principles embodied in the ICMJE Uniform Requirements for Manuscripts, and any conflicts of interest should be clearly noted. However, numerous clinical trials remain unpublished and discrepancies are common when comparing published results with data submitted to the Food and Drug Administration (FDA).2, 3, 4

Congress also recognized the need for clinical trials to be more transparent. The Food and Drug Administration Amendment Act of 2007 (FDAAA 801) required that sponsors and principal investigators of clinical studies that involved human subjects register their studies and submit adverse event information by September 2009 using the ClinicalTrials.gov Web‐ site. And, in October 2008, the General Assembly of the World Medical Association amended the Declaration of Helsinki‐Ethical Principles for Medical Research Involving Human Subjects to state that “Every clinical trial must be registered in a publicly accessible database before recruitment of the first subject” and that “Authors have a duty to make publicly available the results of their research on human subjects and are accountable for the completeness and accuracy of their reports. They should adhere to accepted guidelines for ethical reporting. Negative and inconclusive as well as positive results should be published or otherwise made publicly available.”

Even with mandatory reporting, many trials are not reported on time or are reported late and as many as 60–90% are not reported at all.5, 6, 7 More importantly, the Federal legislation left an enormous hole: the only clinical trials results that are required to be registered under FDAAA 801 are those that study drugs, biologics, and devices that are “approved, licensed, or cleared by FDA.” (http://www.ncats.nih.gov/research/cts/cts.html) Everyone is anxious to publish positive trials—investigators, sponsors, and journals. However, there is often little incentive to publish negative trials: journals worry about the adverse effects on their impact factors and authors find that they are only able to publish negative results in less prestigious journals. But it is the study sponsors who may have the most to lose when negative results are published—particularly, when negative results are disparate from earlier results that led to the approval of the drug or device or when negative results might lead physicians or regulatory agencies to question the efficacy of the drug in the treatment of patients with a different disease or a different indication. However, the failure of a company to publish the results of a clinical trial may not be for nefarious reasons. Small companies may not have the resources to support the publication of negative results and some small companies may not even be in existence if their lead compound failed to demonstrate the requisite effects in a Phase II or Phase III clinical trial. FDAAA 801 also does not require sponsors to register or report the results of Phase I or small pilot studies.

Doshi et al. from the Johns Hopkins University School of Medicine recently pointed out that there were two major flaws in reporting the results of clinical trials that may lead healthcare providers to draw invalid conclusions: invisibility and distortion.8 Invisible trials are those that remain unpublished for years after completion or were “abandoned.” By contrast, distortion occurs when publications in medical journals are biased or misleading in their descriptions of the design, conduct, or results of a trial. He and his colleagues call on institutions that funded and investigators who conducted abandoned trials or who published trials that misreported their data to publish abandoned trials or republish misleading results within 1 year. They go on to suggest that in the absence of the publication of the seminal trial data or a correction of published data, all available data should be considered “public access data” and therefore available for other investigators to publish. This concept of Restoring Invisible and Abandoned Trials has been called RIAT. Recent litigation and new freedom of information rules in Europe have facilitated the ability of investigators to access clinical study reports, thereby enabling restorative authorship. Indeed, the authors of the RIAT proposal provide a list of abandoned trials that includes drugs such as zanamivir, atorvastatin, gabapentin, and paroxetine—drugs that are used by millions of people worldwide. This list provides strong evidence that the practice of incomplete and selective reporting of clinical trial results is endemic in the pharmaceutical and device industry—despite the public espousal of industry leaders.

Although litigation and freedom of information laws have provided European investigators with access to here to for secret documents from the pharmaceutical and device industry, the laws in the United States have not substantially increased the number of clinical trial documents that are available in the public domain. Nonetheless, it is expected that the amount of clinical trials data will increase substantially as there is an increasing public demand for transparency in clinical research. At present, one of the best sources for information about clinical trials comes from the FDA. When new drug applications are submitted to the FDA, they undergo a review by a member of the FDA staff and the reviews are made public prior to their presentation at an advisory board meeting. These reviews are in the public domain, are carefully undertaken, include a statistical analysis of both prespecified and nonprespecified analysis, may analyze data that is not presented in the relevant manuscripts, and can be attributed to the specific FDA staff member who performed the review. The outcome of advisory committee meetings are also in the public domain and recordings of the presentations and panel deliberations are available. Unfortunately, many failed clinical trials do not come to FDA review and therefore this opportunity is only available for a handful of clinical trials.

Other groups have also taken substantive steps to insure that the results of all clinical trials are reported in a timely and transparent fashion. In May 2013, the Geisel School of Medicine at Dartmouth and the Dartmouth Institute for Health Policy and Clinical Practice became the first US academic institution to join the AllTrials Campaign, an effort that was launched this past year in the United Kingdom to support the concept that all clinical trials be registered and their results reported. The AllTrials group pointed out that there are four levels of information about trials: (1) knowledge that a trial has been conducted, from a clinical trials register; (2) a brief summary of a trial's results, in an academic journal article or regulatory summary; (3) longer details about the trial's methods and results, from a clinical study report where available; and (4) individual patient data—The AllTrials campaign calls only for the first three to be published.9 Dr. Elliott Fisher, director of the Dartmouth Institute noted that “biased, selective or nonpublication of clinical trials violates the public trust and undermines the ability to understand the effects of tests and treatments.”10

The concept of reporting the results of clinical trials that have been abandoned presents unique challenges for both the investigator and journal editors. Important questions must be answered: At what point does a sponsor lose the opportunity to publish the study results themselves? What documentation is required to insure that the publication of the study was abandoned? How do the authors document the need for publication? Clinical study reports may include thousands of pages of information: are the study authors responsible for providing public access to all of the source documents and to an audit record? Should all underlying data be available electronically? And, should only the prespecified endpoints be presented or should other endpoints be analyzed as well? Perhaps the thorniest issue that journals will face is whether publication of data by an author who neither participated in the trial design nor represents an organization that funded the study could be viewed as usurping the responsibility of the principal investigator. This concern may be less relevant when the principal investigator seeks to publish their own data without the support or the approval of the study sponsor although contracts between the sponsor and the clinical sites often preclude the ability of an investigator to publish the results of a multicenter study without the concurrence of the sponsor.

To their credit, the RIAT investigators have for the first time outlined a practical algorithm for restoring invisible and abandoned trials.8 They propose six steps: (1) Obtain clinical study reports and other study data; (2) Collect documentation of trial abandonment; (3) Issue a “call to action” by publicly registering the possession of data sufficient for publication; (4) Collect documentation of the need for restoration; (5) Presubmit an inquiry to a “RIAT friendly journal”; and (6) Prepare and submit a manuscript according to RIAT procedures. Although this algorithm has practical challenges and the potential for unforeseen consequences, it provides a platform that can facilitate the ability of investigators to move forward in publishing the results of clinical trials that have been misreported or abandoned by their authors or by sponsors.

With a commitment to improving the process of bringing new treatments, preventions, and diagnostics to patients faster and more economically, the Clinical and Translational Science Institutes can serve as important homes for investigators interested in restorative clinical research. As editors of Clinical and Translational Science, we endorse the proposal made by Dorsi and colleagues and commit to publishing the results of invisible and abandoned clinical trials as well as manuscripts that correct the record for trials that have been reported inaccurately. The RIAT algorithm will serve as a template for the submission of restorative clinical trial submissions; however, we also encourage investigators to submit manuscripts that describe other algorithms, statistical methods applicable to restorative studies, informatics programs for handling large clinical trial datasets, or policy or regulatory issues that arise from restoring invisible and abandoned trials. Sir Iain Chalmers summed up the issue best when he noted that: “Clinical trial transparency is important for moral, scientific, and economic reasons. First, many people volunteer to participate in trials to help increase knowledge, so the failure to report trials is a betrayal of their trust. Second, failing to report trials fully results in biased estimates of treatment effects and leads other researchers up blind alleys. Third, precious time and resources are wasted.”11