The European Medicines Agency (EMA) is committed to the publication of trial data. The question is how to do it? Today, the EMA held a stakeholder workshop which followed the Agency’s decision to ‘proactively publish clinical-trial data and enable access to full data sets by interested parties’.
There is momentum gathering that it is not just a promise of making data available that is required, but that the data actually needs to be made available. No contracts no ties, just available.
Peter Gotzsche, of the Nordic Cochrane Center, spoke at the meeting and declared, ‘there should be open access for all; exactly the reasons why individuals participate agree to participate in trials in the first place.’
Industry talked of better collaboration, better availability of data, better sharing. But within these promises lies an inherent problem. Nobody, not even the manufacturers, knows exactly what they might find. For companies this would be commercially worrying, a point reiterated throughout the meeting: investigative journalists and epidemiologists could find new harms, maybe new indications; but in the short term the worry is openness may undermine profits.
‘Problems were not solved by regulators, but by many eyes, to search and analyse drugs that failed,’ said Ben Goldacre.
Industry representative’s stated they would like openness on a case by case basis and protection of intellectual property rights. Such an approach will potentially stall the process. Any aspect of trial data can be seen as commercially confidential, it is a simple argument to make, and therefore under the guise of confidentiality, data can legitimately be withheld.
There is no doubt that making clinical data available will have a substantial impact on human health. But there are current concerns from an industry perspective that will potentially stall the process.
Ginny Barbour, editor of PLOS, asked for ‘research to be reliable and reproducible.’ In terms of transparency, data should be made available at the time of approval, allowing reproducibility, and be freely available under a creative commons license. Passionaltely, Barbour voiced the need to ‘not allow the initiative to be mired in endless negotiations. If we locked the room now, we could sort this out in 48 hours.’
One smokescreen is the issue of confidentiality and the need for ethical consent to share data. The idea that access to individual patient data (IPD) requires a special consent procedure is wrong. I have been involved in a number of IPDs, and in the end all of the trialists freely gave their data. Ethics committee were consulted in about half the cases, it was not needed. The main issue for open data is anonymisation, once it is, which it should be at the point of data collection, then the problem goes away.
The European Agency may make data available under a restricted access policy. This position is fundamentally flawed, as the freedom of information (FOI) gives rights to anyone to access official, recorded information held by public authorities.
In terms of the Cochrane Neuraminidase Groups, our current policy, in terms of data acquisition, revolves around the principles of openness and transparency.
‘All of our research will include data sharing agreements based on the principles of open access, data sharing and transparency. This is in the interest that all our work can be independently verified now and in the future by third parties. ‘An attainable minimum standard is “reproducibility,” which calls for data sets and software to be made available for verifying published findings and conducting alternative analyses’. Sharing allows others to ‘access to raw numbers, analyses, facts, ideas, and images that do not make it into published articles and registries’.
Therefore, our paln is to publish data files, associated with any published article, that facilitate data sharing, as well as other files important to the article. All data deposited will be released under the Creative Commons Zero waiver, which releases all legal restrictions on reuse of the data.
The key issue is whoever analyses the data, at some point in time has to make it available, because the results need to be transparently analysed, reproduced, and scrutinized. The good news for now is that the EMA is moving forward with the issue and in the UK, Members of Parliament are set to investigate drug companies.
The next steps for the EMA include forming an advisory group around five related themes: protecting patient confidentiality, clinical trial data formats; rules of engagement; god analysis practice and legal aspects. Nominations for membership of the five panels, limited to one member per organization, are open as of December and registration will be open until the 21ST of the same month.