MPs told about practical difficulties of full data disclosureBMJ 2013; 346 doi: http://dx.doi.org/10.1136/bmj.f3200 (Published 16 May 2013) Cite this as: BMJ 2013;346:f3200
MPs examining the topic of clinical trials and data disclosure heard that there may be operational practicalities in the compulsory publication of clinical study reports and raw patient data.
Kent Woods, chief executive of the Medicines and Healthcare Products Regulatory Agency, told the science and technology select committee that he agreed that all trials should be registered and a summary report published. But he added that publishing more detailed information, such as clinical study reports or raw patient data, was more difficult.
Woods said: “They would need to be carefully scrutinised in detail and may need to be redacted in some places, which could be a very labour intensive process.” He added: “We need to give serious consideration [to] how it can be done so it can be useful and operationally feasible and protect patient confidentiality.” He said that there was a danger it could swamp the end user with a level of detail that was not useful and make the process of carrying out research much more burdensome.
Peter Knight, deputy director and head of research information and intelligence at the Department of Health, agreed that protecting patient data was important. “Anonymisation is an issue because there is a risk that different datasets could be meshed together,” he said.
In January, the AllTrials campaign was launched, calling for all clinical trials to be registered and at least basic results published (www.alltrials.net).1 Earlier in 2013, the European Commission proposed changes to its clinical trials directive including measures to ensure comprehensive registration of trials.2
In March, Glenis Willmot, leader of Labour MEPs and spokesperson on health, proposed that these changes did not go far enough, and called for full clinical study reports to be put on the database.3
Janet Wisely, chief executive of the Health Research Authority, said: “There is no excuse for not registering all clinical trials and the publication of basic results.” She told the committee that from September 2013, registration of a clinical trial would be a condition of ethics approval by the Health Research Authority.4
Tracey Brown, managing director of Sense About Science and cofounder of the AllTrials campaign, told MPs: “We need to get past this issue that secrecy is OK.” She added: “Patients who have taken part in trials are shocked when they find out that all the information is not shared.”
Brown told the committee that research data could be divided into four levels. Level 1 was basic registration of the trial, while level 2 was a basic summary including primary outcomes and protocols. Level 3 was full clinical study reports, or their equivalent, which may contain a lot of individual patient data including adverse reactions. Level 4 was the raw patient data. She said that the aim of the AllTrials campaign was to publish at levels 1 and 2. “It is shocking that this has not been published already,” she said. But for levels 3 and 4, Brown agreed that there were still problems that needed to be resolved.
Brown said that 300 organisations had already signed up to the AllTrials campaign in the few months since it was launched. She said that GlaxoSmithKline has signed up and was committed to releasing level 3 data.
She told the committee that AllTrials were calling for the publication of trial data to be retrospective from 1990. This was because the majority of medicines prescribed in the United Kingdom have received their marketing authorisation since 1990. She warned against making publication of trial data prospective: “We have to be careful about continually kicking the football up the pitch,” she said.
Simon Denegri, chair of INVOLVE, an advisory group that supports greater public involvement in research, told the committee that one of the commonest complaints from patients who had taken part in a trial was that they were never told the results of a trial and often not even thanked for taking part.
Cite this as: BMJ 2013;346:f3200