Regulators argue for compromise on raw data release

Key European Union regulators and their counterparts from the UK, France and The Netherlands have called for greater transparency over access to clinical trials data.

But in a paper in the journal PLoS Medicine Hans-Georg Eichler, the European Medicines Agency's (EMA's) senior medical officer, and the agency’s executive director Guido Rasi, stop short of calling for uncontrolled release of the trials data.

Their proposal, outlined in the paper and supported by officials from the national pharmaceutical regulators in the three countries, is based on fears of the consequences of publishing full clinical trials reports of authorised drugs to enable independent re-analysis of drugs' benefits and risks.

Among other objections, they warn that scientists might seek out problems that might advance their professional careers, to the detriment of balance.

“We submit that analyses by sponsor-independent scientists are not generated in a [conflict of interest]-free zone and, more often than not, ego trumps money. Independent analyses may therefore also be ‘vulnerable to distortion’.”

They also warn that patient confidentiality could be breached.

The way forward, they suggest, is to lay down standards for protection of personal data and quality standards for meta-analyses, and to establish rules of engagement.

A US-based website sponsored by the National Institutes of Health, other governmental agencies and private industry, has details of studies from 179 countries totalling 123,184 trials receiving more than 50 million views a month and 65,000 visitors daily.

Fewer than a quarter of the clinical trials listed with the federal database have had their results posted in the database within a year of publication.

Objections to cautious approach

There have been objections to the cautious approach advocated in the paper.

Professor Ole Petter Ottersen, rector of Oslo University and a professor of biomedicine, told University World News: “Does this not make the research community look suspicious, and the opposite of increased transparency? It is one thing to keep the data close to your chest before publication, but after that the data should be publicly available.”

But Eichler responded: “I agree. This is what we argue – that researchers should be willing to share their original data. In most cases this will be after publication. In some cases, for example, in the case of an urgent public health issue, it might be even before that.”

Asked if he had come across concrete cases of scientists using "multiple exploratory re-runs of datasets looking for a problem that might help make their name, sacrificing balance", Eichler declined to name cases.

Peter Doshi, a researcher at Johns Hopkins University's school of medicine and author of an article setting out “the imperative to share clinical study reports”, to which Eichler and his colleagues responded, told University World News:

“I think Dr Eichler and colleagues make many excellent points in their PLoS article. Academics are said to be among the poorest in trial outcome reporting to http://ClinicalTrials.gov, so there is room for improvement all round.

“And within academia, a desire to promote one's career is certainly a powerful motivator that can encourage all kinds of behaviour, including things like data secrecy and at times data fabrication."

Doshi and his colleagues base their argument on their research into Tamiflu, which they say was hampered by the manufacturer’s refusal to share the results of clinical trials.

He said: “The European Medicines Agency released more than 25,000 pages of Tamiflu trial reports to our group in response to our freedom of information requests, and did so promptly and with little redaction. So I am not pessimistic about EMA’s intentions, given how prompt and forthcoming they have been with data to us.”

Differing views

Professor Matthias Kaiser, who was for nearly two decades until 2010 secretary-general of the Norwegian National Committee for Research Ethics and is now director of the University of Bergen Centre for the Study of the Sciences and the Humanities, told University World News:

“Openness in data sharing, the public availability of data, is considered a good thing by most active researchers, even in medical research.

“Now, scientists from drug-regulating bodies issue a warning: also scientists who are not paid by the pharmaceutical industry may be guided by other interests than mere truth, consequently they may not be fully objective.

"The implication is, of course, that this would not apply to the regulatory bodies themselves, but we are not told why regulatory bodies are free of any conflict of interest," Kaiser said.

“Be that as it may, they fear that the data may, for example, be used to issue public health scares (similar to the food scares with which we have become familiar). The public may get confused by different and differing voices from science. They may not know what is in their best interest .

“The situation they fear is that medical research becomes like what most other fields of research already are today. When it comes to hard policy decisions, with high-value stakes and great uncertainties, it is public trust that matters the most. And trust is strengthened by public data sharing, not by controlling access.”

Professor AJS Davies, former deputy director of the Institute of Cancer Research in London, said: “It seems to me that raw datasets concerning clinical trials should, like income tax returns, be kept for a period of time relevant to the duration of the trial. For income tax it is seven years in the UK.

“They should be available for inspection but not freely and only when there seems a good reason for them to be gone over. There should be no reason for the full array of trial data to be made available at the time of publication of the trial outcome," said Davies.

"If someone questions the statistical methods applied to estimate the weight of the results then that could be a reason for going back to the original information. If there are other, conflicting trial results then again that could be a reason for going back to the original results.”

He said that that the authors of the Tamiflu trial should have made their results available – there should have been some means of obtaining disclosure because clearly there is considerable doubt that the drug actually works.