It is “neither desirable nor realistic to maintain the status quo” of clinical trial data being available on a limited basis, a group of leading European drug regulators believes.
The officials from the European Medicines Agency (EMA), France’s Agence Française de Sécurité Sanitaire des Produits de Santé (AFSSAPS), the UK’s Medicines and Healthcare products Regulatory Agency (MHRA) and the Dutch Medicines Evaluation Board (CBG-MEB) have pitched into the debate over selective publication of clinical-trial data in an article for the open-access journal PloS Medicine.
They were responding to a piece in PloS Medicine by Peter Doshi and colleagues about the difficulty of securing access to reliable clinical-trial data on Roche’s antiviral Tamiflu (osteltamivir).
Doshi et al argued that full clinical trial reports on licensed drugs should be made publicly available to enable independent re-analysis of these products’ risks and benefits.
Not commercially confidential
The regulators agree with Doshi et al. that clinical trial data should not be regarded as commercial confidential information.
Most patients who enrol in clinical trials do so with the assumption that they are contributing to medical knowledge, they observe.
Non-disclosure of comprehensive trial results undermines that philanthropic sentiment, write EMA executive director Guido Rasi; Hans-Georg Eichler, senior medical officer at the EMA; Eric Abadie, the former scientific adviser to AFSSAPS who has just resigned as chair of the EMA’s Committee for Medicinal Products for Human Use; Alasdair Breckenridge, chair of the MHRA; and CBG-MEB chair Hubert Leufkens.
By clinical trials results, Rasi et al. mean “not just the protocol, summary tables, and figures of (mostly) randomised controlled trials (RCTs), but the full ‘raw’ data set, including data at the patient level”. The discussion in PloS Medicine refers only to data on drugs where the regulatory benefit-risk assessment has been completed.
The potential benefits for public health of independent (re-)analysis of clinical-trial data “are not disputed”, and in an open society this should not be the sole preserve of trial sponsors and regulators, Rasi et al. write.
All the same, “the different responsibilities of regulators and independent analysts have to be acknowledged. Regulators, unlike academicians, are legally obliged to take timely decisions on the availability of drugs for patients, even under conditions of uncertainty”.
Other, “potentially more important” benefits that could arise from public disclosure of raw trial data include developing predictive models for patient selection to appropriate treatments; and, eventually, “machine learning systems that will allow clinicians to match a patient’s electronic health record directly to RCT and observational study datasets for better, individualised therapeutic decisions”, the regulators suggest.
There are, however, drawbacks in providing easy and unrestricted access to full RCT data, the authors caution.
Foremost among these is the question of personal data protection or patient confidentiality, “a concept that is very different from commercial confidentiality”.
There is a “small risk” that personal data could inadvertently be publicised or that individual patients could be identified from anonymised datasets, Rasi et al acknowledge – for example, in the case of clinical trials for ultra-rare diseases.
“Achieving an adequate standard of personal data protection is not an insurmountable obstacle, though, and proposals for best practice for publishing raw data are available,” they add.
Nonetheless, “implementation is not straightforward, standards will need to be agreed upon up front, and data redaction may in a few cases be resource intensive”.
Conflicts of interest
A second and “likely more contentious” caveat is the implied assumption in the debate over sponsor-independent analyses of clinical trial data that these analyses are necessarily free from conflicts of interest (CoIs).
“We beg to differ,” the authors comment. “Personal advancement in academia, confirmation of previously defended positions, or simply raising one's own visibility within the scientific community may be powerful motivators.”
Analyses by sponsor-independent scientists “are not generated in a CoI-free zone and, more often than not, ego trumps money”, Rasi et al. suggest.
As such, the regulators are “concerned that unrestricted availability of full datasets may in some cases facilitate the publication of papers containing misleading results, which in turn lead to urgent calls for regulatory action”.
In the worst case, this would “give rise to unfounded health scares with negative public health consequences such as patients refusing vaccinations or discontinuing drug treatment”, they warn.
Nor is independent analysis per se a guarantee of high quality, the authors add. “The regulatory community has been confronted with meta-analyses that were later contradicted by additional evidence or found to be flawed.”
Rasi et al. purpose a three-pronged approach to resolve these potential conflicts:
1. Developing and agreeing on adequate standards for the protection of personal data when publicizing RCT datasets. “We emphasise adequate standards because excessive demands and unrealistically high standards may in effect become an ‘anti-commons’ and frustrate important public health gains,” the authors comment.
2. Ensuring the general adoption of established quality standards for meta-analyses and other types of (confirmatory) data re-analysis that may warrant regulatory action.
3. Establishing “rules of engagement”. For example, the European Network of Centres for Pharmacoepidemiology and Pharmacovigilance has recently published guidance for raw data-sharing in the field of observational studies based on healthcare databases.
“These rules of engagement follow the principle of maximum transparency whilst respecting the need to guarantee data privacy and to avert the potential for misuse.”
The authors “take it as self-evident” that the same standard of openness should apply “to all (drug) trial data, whether sponsored by industry, investigator-initiated, or sponsored by public grant-giving bodies”.
Likewise, they add, the same standard of third-party scrutiny should be applicable to all secondary data analyses.