GSK - Avandia: The Lancet wades in!

The story of rosiglitazone is one of death, greed, and corruption, according to the Staff Report of the United States Senate Committee on Finance, released on Feb 20, 2010. The 2-year investigation by Senators Max Baucus, Chuck Grassley, and others, suggests that excess cardiovascular events in patients taking rosiglitazone appeared as early as 2004, but that the manufacturer, GlaxoSmithKline (GSK), intimidated researchers and manipulated the scientific process for commercial advantage. RECORD, one of the studies at the centre of this storm, was published by The Lancet in 2009.

The Staff Report claims that GSK unblinded the data 2 weeks before approaching the RECORD steering committee to suggest an interim analysis. GSK maintains that the company has been diligent in investigating rosiglitazone's safety and points to the fact that the drug is still licensed by the US Food and Drug Administration. Add to this controversy Steven Nissen's account, published on March 24 in the Journal of the American Medical Association, of a manuscript leaked by a peer-reviewer, indiscreet industry emails, and clandestine tape recordings, and one has the ingredients of a John Grisham novel.

If the facts are as presented in the Senate Committee's report, then at a time when some pharmaceutical firms have received record fines for misconduct, the saga of rosiglitazone tests the limits of tolerance. The implications of the Staff Report deserve attention for many reasons. First, such findings damage all who are involved in clinical research, and will likely make funding, ethical approval, and recruitment more onerous for future studies. Second, without confidence in the research process, the ability of findings to benefit policy and patients' care will be compromised. Third, the trust between doctor and patient, researcher and participant, or author and editor is undermined when the foundations on which evidence is built are treated with such casual contempt. The only winners are those who would promote market-led, anecdote-based medicine, without regard to effectiveness, safety, or cost.

All who use, receive, or pay for health-care interventions depend on guidance from reliable research findings and will want reassurance that medical research is credible. How can this common desire be accomplished most effectively, and in a manner that neither stifles innovation, nor delays the communication of findings?

At The Lancet, we believe that one answer is better-designed and more transparent protocols that are freely available for inspection. A protocol determines the structural integrity of an investigation, as an architect's drawing determines that of a building. Since 1997, we have introduced several initiatives to improve the quality, reporting, and transparency of research in general and randomised trials in particular, by emphasising the importance of protocols. We offer to review protocols to improve trial quality and reduce publication bias. We consider submissions of randomised trials only if registered and accompanied by a protocol, which is sent with the manuscript to peer-reviewers. We apply internal checklists to confirm fidelity between protocols and the submitted manuscript, and last year we offered to include links to the full protocol for any research that we publish. Frankly, the results of these efforts are disappointing. Generally, protocols are not respected. An audit by Sanaa Al-Marzouki and colleagues that compared accepted protocols with the subsequent Lancet publications found that adherence to protocols was ambiguous and selective. One reason may be that while there is excellent guidance to design a clinical question or report a study, there is little to inform the construction of, or enforce adherence to, a robust protocol.

To fill this gap, An-Wen Chan, Jennifer Tetzlaff, and CONSORT visionaries David Moher and Doug Altman, together with an international consortium of trialists, methodologists, editors (including from The Lancet), and other stakeholders have spent 3 years developing SPIRIT (Standard Protocol Items for Randomized Trials). With an evidence-based approach to define, refine, and validate a 33-item checklist, SPIRIT is designed to ensure that protocols include a minimum set of core scientific, ethical, and organisational matters related to randomised trials.

The SPIRIT statement and checklist are expected to be launched later this year. By showing in a transparent manner the thoroughness of the protocol, SPIRIT will be a useful aid for designing and monitoring research. While nothing can completely protect against scientific misconduct, SPIRIT—in addition to improving overall trial quality—will make deviations more difficult to execute and to disguise.

http://www.thelancet.com/journals/lancet/article/PIIS0140673610605235/fulltext?rss=yes