Calls for public scrutiny of clinical trial data have intensified. In the UK, clinical trial data held by public authorities (such as universities or NHS trusts) may have to be disclosed under the Freedom of Information Act 2000 (FOIA). In a recent decision, a tribunal hearing an appeal brought by Queen Mary University of London (QMUL) under the FOIA against the Information Commissioner’s Decision Notice was rejected. The tribunal held that QMUL’s clinical trial data had been sufficiently anonymised and thus was not exempt from disclosure on the grounds that it constituted personal data. The tribunal also found that other exemptions under the FOIA did not apply in this case. Notably, it rejected the university’s argument that complying with the FOIA request would prejudice the university’s commercial interests, such as obtaining research funding.
Just last week, researchers analysing published and unpublished clinical trial data reported that an average of only 46% of published material contained information about adverse events, compared with 95% of unpublished material relating to the same trials. Studies like this one have prompted calls for greater transparency in clinical research activities. It is in response to this campaign that in September the US Department of Health and Human Services issued more stringent requirements for publication of clinical trial data. Similarly, in March 2016 the European Medicines Agency (EMA) published detailed guidance for pharmaceutical companies on the publication of clinical data. This document provides instructions on anonymising data and the submission of clinical reports, including reporting on adverse reactions.
However there is still significant reluctance amongst private and public bodies to openly share clinical trial data. In this context, a recent decision where a UK university challenged an order to disclose such data under the Freedom of Information Act 2000 (FOIA) provides guidance.
The Queen Mary decision
The FOIA gives any member of the public the right to access any information held by public authorities (including universities and NHS Trusts) in England, Wales and Northern Ireland (and by UK-wide public authorities such as the NHS in Scotland), subject to certain exemptions. One such exemption covers information derived from a programme of research where the intention is to publish the research and disclosure of that information would be likely to prejudice the programme (s. 22A FOIA). The research exemption, however, will only apply before the research is published.
Whether the information remains exempt after the publication of research depends on the nature of the data, and therefore public authorities holding clinical trial data may also need to rely on other FOIA exemptions. These include exemptions related to disclosure of information prohibited by the Data Protection Act 1998, information provided in confidence, and information that would be likely to prejudice commercial interests. See §§ 40-43 of FOIA.
Each of the above exemptions was recently considered in the context of an appeal by Queen Mary University of London (QMUL) against the Information Commissioner’s decision to order the disclosure of clinical trial data (EA/2015/0269). The tribunal tasked with reviewing the decisions of the Information Commissioner upheld the decision to disclose the data.
Two issues addressed by the tribunal are of particular interest to those involved in clinical trials at public bodies:
1. Is the clinical trial data sufficiently anonymised to allow disclosure under the FOIA?
In relation to the first issue, QMUL argued that the individual patient-level data was pseudonymised (and not anonymised) and therefore was likely to constitute personal data exempt from disclosure under s.40 of the FOIA. QMUL argued that there was a significant risk of re-identification as:
(1) participants could self-identify;
(2) those with detailed prior knowledge of participants (such as friends, family or medical practitioners) could identify them; and
(3) motivated intruders could identify participants and link this data to other information previously released.
The majority of the tribunal members rejected this argument, holding that the data did not constitute personal data and thus was not exempt from disclosure. The tribunal noted that information requested did not contain any fixed or direct identifiers. Under the FOIA, anonymisation of personal data is not expected to be completely risk free; the risk must only be mitigated until it is remote. In this case, re-identification of patients would have required a great deal of work and would only have been possible on the assumption that the health workers would act outside their legal and ethical obligations.
QMUL also argued that the information was confidential under s. 41 of the FOIA and disclosure would contravene a legitimate expectation of confidentiality. The tribunal disagreed as the data was anonymised and therefore there was no breach of confidence.
In addition, the tribunal pointed to the fact that QMUL had previously supplied the data to independent scientists as part of normal research collaboration under formal confidentiality agreements thereby tacitly acknowledging that the data had been effectively anonymised. The tribunal made it clear that it was unacceptable for public authorities to cherry-pick who analysed their anonymised data.
2. Would disclosure of clinical trial data prejudice the university’s commercial interests, reputation and funding streams?
QMUL argued that it should be exempt from complying with the FOIA request as disclosure would prejudice its commercial interests in three ways. First, existing participants might withdraw consent; second, disclosure would deter current participants from participating in long-term follow-up studies; and third, it would deter other individuals from taking part in future research.
The majority did not accept QMUL’s commercial interest argument as there was little evidence of the withdrawal of consent and its occurrence had not been directly related to the issues raised here. The tribunal accepted the Commissioner’s argument that it was up to QMUL to allay participants’ fears by explaining to them that data was anonymised so that they could not be identified from it. The tribunal suggested that new confidentiality guarantees, which were possibly more explicit, could be given to new participants. The tribunal rejected the university’s argument that its ability to obtain future funding would be prejudiced by disclosure, as the university had already obtained funding for a new trial after Commissioner’s decision to disclose the data.
The tribunal unanimously held that such minimal risk of prejudice to QMUL’s research programs, reputation and funding streams did not outweigh the public interest in disclosure to assist legitimate academic debate. In this case, over 12,000 people including researchers, academics, patients and journalists had signed a petition calling for the retraction of questionable claims and disclosure of the data.
The research exemption from disclosure under the FOIA will apply to clinical trials undertaken by public authorities for the duration of the research programme (as long as there is an intention to publish). However, this exemption ceases to apply after publication.
At this point, unless one of the other exemptions applies, anonymised individual patient-level data may be disclosed under the FOIA as it no longer constitutes 'personal data'. Data that has been anonymised to a level where the risk of re-identification of participants is remote is not exempt from disclosure. The risk of re-identification, however, may be higher in the case of data that relates to very small populations or rare conditions. Such data sets are thus more likely to fall within the FOIA personal data exemption.
Public bodies and those conducting clinical trials with them would be wise to put in place a long-term strategy for data management before the start of the clinical trial and prepare at the outset for the possibility of an FOIA request and disclosure, particularly in light of the recent push for transparency. To address the concerns of patient volunteers who are considering joining a trial, those undertaking clinical trials should provide clear and detailed guidance on the anonymisation of data and assure the volunteers that they will not be identifiable if individual patient-level data becomes public in response to an FOIA request.
*Golder S, Loke Y K, Wright K et al. Reporting of adverse events in published and unpublished studies of health care interventions: a systematic review. PLoS Med 13(9):e1002127. doi: 10.1371/journal.pmed.1002127