Cochrane researchers continue to face challenges over access to data on flu drugs

BMJ 2013; 346 doi: (Published 16 May 2013) Cite this as: BMJ 2013;346:f3190
  1. Zosia Kmietowicz
  1. 1BMJ

Cochrane Collaboration researchers have said that they may be unable to analyse clinical trial data given to them by GlaxoSmithKline (GSK) on the antiviral drug zanamivir (marketed as Relenza) because nearly two thirds of the reports provided have been redacted in different ways.

The researchers have been trying to get access to data held by drug companies on neuraminidase inhibitors for more than three years so that they can review the drugs’ safety and effectiveness.

GSK was the first company to hand over clinical study reports, giving researchers 30 reports at the end of April. The reports hold vast amounts of information on trials’ conduct, adverse events, and outcomes.

The Cochrane researchers remain positive about GSK’s attempt to share the detailed data it holds on zanamivir but have written to the company asking it to clarify why certain information had been redacted.1 This information includes descriptions of serious adverse events, the dates of the start and end of studies, and codes for study sites.

The letter to GSK from the Cochrane investigators Tom Jefferson and Peter Doshi said, “If we do not have a complete CSR [clinical study report], we have to figure out what we are missing and why. This is not something we like doing, as we are neither auditors nor regulators.”

Jefferson and Doshi asked 19 questions of GSK about the data and the purpose of releasing them to researchers, including, “What type of independent research does GSK think this kind of release enables?”

The researchers are trying to define a minimum dataset for use in evidence synthesis but acknowledge that the learning curve for them and GSK is steep. They hope that a discussion will lead to a more fruitful analysis of the data.

In February GSK became the first drug company to sign up to the AllTrials initiative (, set up in January by the charity Sense About Science, the BMJ, and other supporters of transparency in research.2 It is calling for the registration of all clinical trials and for full study results and full clinical study reports to be made publicly available.

But the cat and mouse game of hiding data and promising disclosure continues.

On 2 April the drug company Roche told the Cochrane researchers that it would give them 74 clinical study reports on oseltamivir (Tamiflu) in a series over the next few months.3 But the researchers have not yet received any of the studies.

Jefferson told the BMJ, “We have already had one promise from Roche in December 2009 in the BMJ. We are still waiting for the ‘full study reports’ that were promised. So far there have been promises, declarations, and statements, but we have not seen anything.”

On 10 May Jefferson wrote to Roche’s “life cycle leader” on oseltamivir after analysing the clinical study reports on two trials that the company sponsored.4 He called for one 14 year old study to be published “for the whole world to benefit from the data contained in it.” The second study, which was published in JAMA in 2001,5 needed to be retracted or amended because many of the statements and conclusions it drew were “not supported by the design of the trial or the data presented,” he said.

That study has played and continues to play a role “in the clinical recommendations for the use and the stockpiling of Tamiflu at huge taxpayers’ expense,” he wrote.

Campaigners for full transparency of clinical data are also frustrated by an interim ruling by the General Court of the European Union ordering the European Medicines Agency not to release documents it holds on the safety and efficacy of drugs made by the US companies AbbVie and InterMune, which took out injunctions against the agency after other companies requested the information.6 7

Campaigners condemned the ruling, saying that there was no justification for withholding information that would allow patients to make fully informed decisions about the drugs they take.


Cite this as: BMJ 2013;346:f3190