EMA calls for open access to clinical trial data after Tamiflu concerns

“The article’s authors, including the European Medicines Agency’s (EMA) executive director Guido Rasi and its senior medical officer Hans-Georg Eichler, said that clinical trial data should not be considered commercially confidential, with open access potentially leading to public health benefits. It was written in response to a separate article in PLoS Medicine by Peter Doshi et al that recommended improvements in the sharing of clinical trial data after concerns were raised about the validity of data to support the use of Roche’s Tamiflu as an effective antiviral to treat influenza. According to Doshi et al’s article, there were discrepancies between what different groups were claiming about Tamiflu, with the EMA and the US Centers for Disease Control and Prevention both holding positive opinions about Tamiflu’s ability to reduce complications for people with influenza. In contrast, the US Food and Drug Administration (FDA) concluded that Tamiflu had not been shown to reduce complications, and required a statement in the drug's label to reflect this. This would also contradict the addition of Tamiflu to the World Health Organization's (WHO) essential medicines list. Doshi et al claimed the discrepancies are down to a meta-analysis of ten randomised clinical trials involving Tamiflu published in 2003 that  became the source of opinion for all these organisations, apart from the FDA which used data from the original clinical trials themselves. In their own review, which began in 2010, Doshi et al used both the 2003 meta-analysis and a freedom of information request to the EMA to obtain additional sections of clinical study reports for Tamiflu, uncovering several details missed in the meta-analysis. Doshi et al used their article to call for ‘clinical study reports become available to such scrutiny’, claiming that Roche refused to provide additional data for the 10 studies in the meta-analysis, while the company’s reasons for doing so ‘kept changing, and none seemed credible’. In the response article, Eichler et al said they agreed with Doshi et al that clinical trial data should not be considered commercial confidential information, and should be available for independent analysis and the development of predictive models. This wouldn’t be without its own problems, however, and Eichler et al also warned of the risk of breaches of patient confidentiality in publication of full raw datasets, arguing for the development of standards to overcome these risks. General adoption of established quality standards of meta-analyses, and rules of engagement for sharing raw data following guidance from the European Network of Centres for Pharmacoepidemiology and Pharmacovigilance (ENCePP) would also be necessary in moving forward... “