Date: 13 Jun 2013, Science, Technology and Culture
Question setter: Tim Edwards

Clinical trials of medicines

The European Medicines Agency has announced that it will publish clinical-trial data and enable access to full data sets by interested parties. A number of practical and policy issues need to be addressed before complex data sets can be made available, and the EMA has been consulting broadly. Do you think the EMA will come out in support of publishing on its web-site all clinical data submitted by trial sponsors when it publishes its opinion, expected by 1st July 2013?


Response:


Answer: No
Confidence level: 100%
Mean confidence level (all requests): 48.67%

Justification:
Although the EMA will be pushing to create greater transparency, it is unlikely to support publishing all the data. There are too many issues of patient confidentiality and commercial competition mitigating against that.
It is possible that all data could be made available to selected, accredited individuals with the expertise to make full use of it but drug companies will be fighting hard not to lose exclusivity on particular products and will be threatening to cut development if all their secrets are revealed. The EMA has recognised this point and will make some concessions. In the end they will be asking the world to trust the EMA's oversight and will try to publish enough of the data, for trials good and bad, to reassure everyone that they have enough access.

Outcome: No
Score: 100
Mean score (all respondents): 15.33

Expert opinion:


Answer: No

Selected Expert Answer from John Karslake:
This topic has generated a lot of heat as MPs, MEPs and journalists have accused "Big Pharma" of withholding adverse results from their drug trials in order to maximise sales and profits. However, regulators and senior doctors, reviewing the data, follow rigorous process and, in the case of the former, wield huge power. The data not being freely available does not indicate that the regulators have not seen it either. So, the issue is really one of trust and it is evident that the public lacks full faith in regulators or reviewing practitioners. There are many cases where problems become evident after launch, when patients are hugely more numerous than in any test group and these cases maintain the attitude of suspicion. The EMA is clearly aware of the need to improve trust by increasing transparency. However, they want to avoid suppressing research and innovation where companies fear the release of commercially sensitive information.
The result is likely to be a recommendation that much more, but not all, data is published for general consumption. Further data may be released to designated reviewers, who act as the public's guardians and are trusted not to disadvantage producers unfairly by revealing commercial secrets.

Answer: No

Selected Expert Answer from Mettletest Panel:
Strictly speaking, the answer is likely to be no because we do not think that publication of ALL clinical data will be recommended, certainly not to all parties anyway. However, the spirit of the recommendation is likely to be a yes. The pressure from respected, popular journalist Ben Goldacre and taken up by many politicians and their committees, will increase the EMA's inclination towards transparency. The EMA is likely to push the limits as far as possible without damaging introduction of innovative new medicines. They want doctors and the public to trust their process and the medicines that the regulators approve. They certainly do not want to be seen to be influenced by the big drug companies.
So, the EMA will support publishing all but the most commercially sensitive data and allow the public access to the good and bad results of trials.

Answer: No

Selected Expert Answer from Tim Edwards:
The EMA will not propose making all clinical data freely available. They support the idea of transparency in principle but recognize practical problems. They may segment data into past and future applications for marketing approval, and into successful or failed applications. They will not look backwards, instead publishing clinical and pre-clinical data (i) in an appropriate form (ii) to legitimate researchers, where that data is in support of successful marketing authorization applications made after 1 Jan 2014. Data on unapproved products will, in the short term, not be published. Data needs to be in an appropriate format. Patient confidentiality needs to be protected and all data anonymised. Relevant data is that gathered from both pre-clinical safety studies in animals, and human clinical trials, together showing that the drug is both safe and effective, and possibly how it compares to other treatments. A product’s scale-up or manufacturing data will not be publicly available, for the time being. Clinical data should be in an agreed standard format (designed by the EMA). Only legitimate researchers, accredited by the EMA, should have access to such published data - practising doctors, investigative journalists, relevant academics and government officials, and some clinical research staff from registered pharmaceutical and biotech companies.


Outcome: No

Comment on outcome from Tim Edwards:
The EMA published their Policy on 26 June 2013, 4 days ahead of the scheduled time. The Policy will come into effect on 1 January 2014. Data will be published for both positive and failed studies. However, study data will be categorised into three categories: Category 1 data contains commercially sensitive information and will not, in general, be made available. Category 2 data does not contain any matter requiring personal data protection and can be accessed reasonably freely under ‘open access’ rules. The data will be in prescribed PDF format for download at the time of the EMA’s decision about the marketing authorization applications (MAA). Category 3 data do require personal data protection, and these are subject to ‘controlled access’ by identified ‘requesters’ of the information, operating under specified guidelines, who are themselves subject to the same transparency rules. They will not be able to identify patients. The data will be in ‘de-identified’ PDF format, also for download at the time of the EMA’s decision about the MAA.