Monsanto Roundup Lawsuit

Thursday, March 20, 2008

Former GlaxoSmithKline Employee Speaks out

Can we trust information from drug companies?

Elizabeth Wager, freelance publications consultant, Princes Risborough

Thursday, 20 March 2008


GlaxoSmithKline will not face prosecution following an investigation by the MHRA into whether the company suppressed safety data about its antidepressant Seroxat (paroxetine). However, the report suggests GSK escaped criminal hearings mainly because the regulations on safety data disclosure were not clear enough.1

The rules will be tightened to prevent similar problems in the future, but, although this change is to be welcomed, it does not ensure that prescribers and policy makers will have access to all the information they need. Drug regulation in the UK (as in most countries) is carried out under an extraordinary cloak of secrecy. Data supplied by companies to the regulators is regarded as highly confidential. An ironic reminder of this appears in the MHRA report itself, which notes that “legal constraints make it impossible to disclose any information or evidence gathered during the course of the investigation.”

This should be of major concern to all clinicians. Evidence-based medicine requires reliable and complete evidence but if prescribers can read only what drug companies choose to publish, and unfavourable findings are suppressed, then the robustness of the evidence base must be questioned.

The GSK case should come as no surprise, and they are certainly not the only company guilty of not publishing unfavourable findings, since evidence of systematic publication bias has been available for decades.2 The Good Publication Practice guidelines for drug companies originally developed in 1999 call on companies to publish the results of all trials of marketed products.3 Drug companies are probably not the only culprits, and journal editors have been criticised for their unwillingness to publish negative findings. One response to this problem has been the creation of trial registers. The announcement by several major journals that they would only publish studies that had been properly registered (ie with details of the design posted on a public website before patients were enrolled) has acted as a major spur.4

Trial registration and prompt posting of results will become a legal requirement in the USA under the FDA Amendment Act which comes into force in 2009. But the journal requirements and American law apply only to current and future trials and to new drugs being licensed by the FDA; they will not correct the gaps in the evidence for the majority of medicines that are already licensed and are being prescribed today. Ironically, the European regulators already have a register of all the trials submitted by drug companies to obtain licences and fulfil post-marketing safety requirements. This goes back much further and is therefore far more comprehensive than, the register created by the US National Institutes of Health. Yet the EudraCT database, which could be such a powerful public health tool, remains inaccessible, with valuable data locked away, because it is considered commercially sensitive and therefore confidential.

While it may be acceptable for companies to keep their development plans for new compounds confidential, there is surely a strong ethical responsibility to publish results of all studies relating to drugs that are on the market. Not only should companies commit to publishing all such studies, they should also commit to publishing all the clinically relevant outcomes. Several studies have shown that selective reporting of favourable outcomes may be an even greater problem than non-publication of entire trials.5

Results databases linked to trial registers offer hope for the future.6 Comprehensive registers will highlight unpublished studies so, for example, people preparing systematic reviews can contact the investigators or sponsors requesting results. Electronic posting should also mean that results can be published easily and quickly, avoiding the problems of getting negative studies accepted by traditional medical journals. Linking the results with the outcomes stated in the original protocol should also reduce selective reporting.

Next time you see a representative from a drug company, ask about the company’s policy on trial registration and results posting, but perhaps more importantly, ask whether they have a policy to publish all studies for all their current products. Most of the major companies are ensuring they will comply with US legislation and many have created websites for posting results. That means we may have more confidence for an unbiased evidence base for drugs in the future – but as the GSK case shows, prescribing decisions for current medicines may be based on partial, and skewed, evidence.

Author’s competing interests: EW is an author of the Good Publication Practice guidelines for pharmaceutical companies and a former employee of GlaxoSmithKline


1. MHRA Investigation into GlaxoSmithKline/Seroxat

2. Chalmers, I. Underreporting research is scientific misconduct. JAMA 1990;263:1405-8

3.Good Publication Practice guidelines

4. De Angelis C, Drazen JM, Frizelle FA, et al. Clinical trial registration: a statement from the International Committee of Medical Journal Editors. Lancet 2004;364:911-2

5. Chan A-W, Altman DG. Identifying outcome reporting bias in randomised trials on PubMed: review of publications and survey of authors. BMJ 2005;330:753.

6.Wager E. Publishing clinical trial results: the future beckons. PLoS Clinical Trials 2006;1(6):e31

Read the new book, The Evidence, However, Is Clear...The Seroxat Scandal

By Bob Fiddaman

ISBN: 978-1-84991-120-7