The European body which oversees the collection of information on side effects of drugs wants to have the right to censor reporting of independent analysis of this information if it doesn't agree with the analysis. The European Medicines Agency oversees the EudraVigilance database where reports of side effects from approved drugs in Europe are recorded. Researchers can request access to the raw data held on the database in order to perform their own analysis of it. On Monday the EMA published a proposed new policy to extend the information it shares with researchers to include individual case study reports of patient's experiences of side effects. The consultation document contains the conditions researchers have to agree to when given access to raw data and individual case study reports. These conditions include that:

The Agency has the right to view any publication resulting from EudraVigilance data before submission (maximum period for initial Agency review will be six weeks) including a privacy check as regards possible re-identification of patients. Any issues raised by the Agency concerning incorrect analyses, unsupported inferences, misleading statements or the protection of personal data must be addressed to the satisfaction of the Agency before submission for publication.

Dr Ben Goldacre, author and co-founder of AllTrials:

Protecting individual patients' privacy is necessary and good. But the EMA also says it should be allowed to suppress anything in an academic paper that it regards as "incorrect analyses, unsupported inferences, [or] misleading statements." This is a profoundly outdated world view. Simply because they are the body collecting this public data from EU patients, that does not give them the right to control how it is used. This seems to simply represent state censorship of scientific discussion and analysis of public health data.

If there are flawed analyses, or over-interpretation of risk signals, then that is a matter for public discussion and debate, not censorship by the medicines regulator. It also puts the regulators in a very conflicted position: it is likely that some analyses of this data may draw close attention to regulatory decisions made by the EMA itself, since these are often close calls (and many have been made in closed and unminuted discussions). It is quite wrong that the EMA should be given the right to censor analyses critical of its own analyses.

Lastly, it is worth noting that this is not a new phenomenon. Regulators have frequently argued against transparency, saying that clinicians may be confused by poor quality or contradictory analyses of patient data, but this concern is not proportionate to the true risk. There are 28,100 scientific journals in print today, with over a million articles published each year, and over 23 million papers indexed in PubMed to date. Work of poorer quality is routinely conducted and published already: it is managed - to a reasonable degree - in the academic ecosystem of evidence synthesis and critical appraisal, before it can impact on practice. Any harm that the EMA might suggest could theoretically arise from a fractional increase in the total quantity of weak academic publications must be balanced against the huge benefits of wider access to patient data.

Dr Virginia Barbour, Medicine & Biology Editorial Director, PLOS, PLOS and co-founder of AllTrials:

In 2011 a number of journal editors including me for PLOS, The BMJ, NEJM and a few others were consulted by EMA about whether we as editors would consider early release of safety data as prejudicial to subsequent journal publication and unanimously we said of course not. So it's troubling and puzzling that the EMA now says that researchers need to have their analyses approved before publication. It's essential for public health that there are no barriers to rapid public release of safety data; we need public scrutiny of these data, not decisions made behind the closed doors of the EMA.

Professor Carl Heneghan, Director, Centre for Evidence-Based Medicine and co-founder of AllTrials:

Information contained in serious adverse event reports is often quite detailed and does pose problems with identification of individuals. Therefore the protection of personal data is warranted. However, the fact the Agency considers it can stop publication because of concerns over unsupported inferences or misleading statements is worrying. Furthermore, the agency also considers it can stop publication because of what it considers "incorrect analyses," this means it will require a significant team of statistical reviewers and will find itself embroiled in numerous longwinded disputes. Indeed, to come to the conclusion that analyses are flawed, will require the agency to openly publish these analyses, for independent scientific scrutiny.

Consequently, what it is proposing is unworkable, and what is clear is regulators are using these measures to prevent more open scrutiny of its databases. This is unacceptable and will lead to delays and inappropriate use of resources. What is also needed is for the regulator to give examples of where incorrect analysis, unsupported inferences or misleading statements have been dealt with in the past and what they mean by these terms. Providing transparent and timely information should be a priority of regulators if the full benefits and harms of treatments are to be realised. I continue to be surprised and perplexed that the European Medicines Agency, as a public body, doesn't think this is a fundamental priority of its work.

The AllTrials campaign for all clinical trials to be registered and the results reported was launched on 9th January 2013 by BMJ, the Oxford Centre for Evidence-based Medicine, Bad Science, Sense About Science, the James Lind Initiative, the Cochrane Collaboration and PLOS. It is being led in the US by Dartmouth's Geisel School of Medicine. The petition has been signed by over 79,000 individuals and supported by more than 500 organisations from medicine, science, consumer rights and industry including more than 200 patient groups.