Thursday 10 April saw the publication of the Cochrane systematic review on oseltamivir and zanamivir, or Tamiflu (Roche) and Relenza (GlaxoSmithKline) to give them their better-known trade names. In short, the review found that Tamiflu doesn’t work quite as well as we thought; a finding that is the culmination of a four-and-a-half year battle for access to the raw data from the clinical trials.
The authors – Jefferson, Heneghan and colleagues – uncovered what they characterized as ‘multisystem failure’, with poorly-defined endpoints and confusion as to the authorship and contribution of the clinical trials. They also found that all studies were conducted against placebo, rather than against current best practice. Overall, the reviewers felt that the published studies were at high risk of bias, leading them to discount the literature as a source for the review and sparking the long fight for the clinical study reports.
Why register clinical trials?
The issues faced by the Cochrane Collaboration in this review are hardly new. As the World Health Organization states, “It is difficult to make informed decisions if publication bias and selective reporting are present.”
Indeed, this is part of the rationale for clinical trial registration. Making fundamental information about a clinical trial publicly available helps identify gaps in clinical trial research and enables researchers and health professionals to identify trials.
The movement towards registration of all clinical trials has been gaining momentum in recent years, with milestones such as the FDA Amendment Act of 2007 and the more recent EU Clinical Trials Regulation.
However, these are limited to new clinical trials for potential future drugs; it does not address the issue of registration of trials on the medications that are in use today. The AllTrials campaign was launched with the motto ‘All trials registered. All results reported’ supporting the retrospective registration of clinical trials, as encouraged by the ISRCTN trial registry, which we host.
Linking all publications regarding a single trial
Registering clinical trials is only the beginning though. Even for those trials that have been registered, it can be difficult to track down resulting publications. Most journals do not publish the clinical trial number (CTN) in the body of the article; therefore, although a results article may reference a published protocol, there is nothing to connect that article to subsequent publications. And nothing to link from the protocol to the results article.
Building on the concept of trial registration, the Threaded Publications initiative was proposed by Iain Hrynaszkiewicz in early 2011, with the ambition of adapting the existing CrossMark standard to centrally link all publications on an individual clinical trial together using their CTN.
Following on from a meeting we hosted earlier this year, an international, cross-publisher working group has been formed to govern the development of this project. Chaired by BioMed Central, the working group includes representatives from CrossRef, BMJ, The Cochrane Library, eLife, F1000, The Lancet, PLoS and Springer, and the ISRCTN registry. The project is now entering a pilot phase, and we hope there’ll be a working prototype developed over the coming months.
All results, not just ‘positive’ results
Of course, in order for an initiative such as Threaded Publications to reach its full potential, it’s important that all results from clinical trials are reported. While it’s often the ‘positive’ data—those that support carefully-constructed hypotheses—that get the most attention, non-confirmatory, so-called ‘negative’ data have an integral role to play in guiding future research.
Indeed, even if an experiment supports a hypothesis, there is always the possibility that it may be rejected by future experiments – a fact that is evidenced by cases such as Tamiflu.
We’ve long been outspoken advocates for the publication of ‘negative’ results. Our Journal of Negative Results in BioMedicine is now entering its thirteenth year and is ramping up its commitment to transparency with its recent move to an open peer-review policy.
Together with Trials, the journal has also publicly declared its support for initiatives, such as Restoring Invisible and Abandoned Trials (RIATs), with an open letter, as well as the ‘research note’ article type, a shortened article type that allows researchers to detail the ‘negative’ result obtained, in a simple, précised way.
And there’s still much more we can do…
All research needs a strong foundation in order to advance our understanding. While traditional results papers, whether confirmatory or not, often present the evidence for future steps, an equally important basis for research is pilot and feasibility studies.
Despite this, many pilot and feasibility studies are never published, and there is confusion as to the definitions and terms. To help address this and to provide a platform for the publication of this important aspect of clinical trials, we’re launching Pilot and Feasibility Studies, under the editorship of Dr Gillian Lancaster of the University of Lancaster. This will be the first journal devoted to this field, giving it the opportunity to influence the design and reporting of these studies. It is also supported by many members of the CONSORT pilot and feasibility studies group.
Reporting guidelines, form the backbone of the reporting of medical evidence, helping to define standards and ensure that all the necessary information relating to a study is reported in a clear and transparent way.
Our Journal of Medical Case Reports was one of seven journals to publish the CAse REport (CARE) Guidelines last September. I will be participating in the meeting of the main CONSORT Group in May, whose Executive Committee includes David Moher, co-Editor-in-Chief of Systematic Reviews, and Doug Altman, co-Editor-in-Chief for Trials. Similarly, many representatives from Systematic Reviews are involved in the production of the PRISMA Guideline extensions that are due to be published later this year.
We are also involved in initiatives to make the conducting of the trials themselves more efficient, such as the Trial Forge project, which is kicking off with a group meeting in July that I will be attending.
The failures in the Tamiflu case help to illustrate the driving need for full transparency around clinical trials; however, it is also an illustration of how medical evidence is built upon that transparency and its openness to criticism. After all, the saga of the Cochrane oseltamivir and zanamivir review began with a simple online comment.