As a result of the AllTrials campaign considerable progress has been made in the last few years to improve trial transparency. This is urgently needed as it is estimated that only around half of the trials that get done go on to be reported.
The trial transparency priorities of AllTrials are in four areas:
- Prospective trial registration: this has been the simplest to implement and as a result the most progress has been made in this area. Leading medical journals now require trials to be registered before publication and in the UK registration is now a requirement of ethics committee approval. A number of trial registries exist, the largest of these being ClinicalTrials.gov.
- Summary results reported: some progress has been made in this area. AllTrials calls for a summary of results to be made publicly available where the trial was registered, within one year of completion of the trial. Currently, not all trial registries accept summary results and many trials still do not report within one year of completion. It has been estimated that it takes on average one person around 40 hours to upload a summary report to ClinicalTrials.gov.
- Trial materials available: Progress has been slow in this area. Trial protocols and other key materials, such as statistical analysis plans, participant consent forms are often required by other researchers for secondary analysis, but currently, only the ISRCTN registry allows the upload of such documents in PDF format.
- Full reports: AllTrials calls for full clinical study reports to be made publicly available when they have been prepared. Typically, full clinical study reports are prepared by trial sponsors as part of the submission process to regulators for the approval of a new drug. An example of a registry that includes full clinical study reports is the ClinicalStudyDataRequest.com. The European Commission as part of the implementation of the new EU Clinical Trial Regulation are currently building a new trials registry for the registration and posting of clinical study reports for all trials conducted in Europe.
Overall, in the last few years progress has been made in all four areas above with the likelihood that in the next few years simple solutions will be found to those problems that are currently getting in the way of further progress. As a result of AllTrials, we are approaching an era when all trials are registered and the results are reported.
Providing access to individual patient data is not a priority of AllTrials. In fact, the AllTrials campaign is explicit in “not calling for individual patient data to be made publicly available” with some commentators believing that recent efforts by some organisations to promote transparency by sharing individual patient data actually distracts from the other ways outlined above that are already implemented and making progress . Many technical challenges still need to be overcome before sharing individual patient data becomes as simple as the approaches described above to promote trial transparency.
Another major concern is that requiring the sharing of individual patient data makes it much harder to get randomised trials done in the first place. The major reason for this is the time and effort required at the end of a trial to upload listings of deidentified individual patient data. While uploading summary results is feasible, doing the same for individual patient data is not cost-effective and is beyond the resources available for most randomised trials. It has also not yet been shown that the individual patient data that is uploaded can be effectively used by other researchers for secondary analysis without further information, such as meta-data and analysis manuals also being provided by trialists. Further, those with experience of doing collaborative individual patient data meta-analysis have also raised questions as to whether secondary analysis can be undertaken without active collaboration with the original trialists. In addition, current proposals that individual patient data be made publicly available within 6 months of publication of the trial have raised concerns that this will act as a disincentive for researchers to actually do trials and, instead, they might wait for others to do a trial and for the data to become available for secondary analysis.
In summary, the current four priorities of AllTrials to promote trial transparency are being implemented and are resulting in more trials being registered and the results of them being made publicly available. AllTrials explicitly does not support the sharing of individual patient data and the current focus on ways to share individual patient data might actually distract and slowdown the more straightforward ways that are already having an impact. Finally, requiring trialists to share individual patient data will make it harder to do trials in the first place.