Following the publication of a controversial editorial in the New England Journal of Medicine, the Health Affairs Blog recently took a closer look at the risks and benefits of sharing clinical trial data.
In his blog post, Stuart Bock acknowledges the potential for data misuse, while maintaining the view that increasing transparency is an overwhelmingly positive practice. There is the possibility of researchers racing to scoop original trialists for high-tier journal publications, or of a trial being invalidated through an incorrect reanalysis. However, such events are unlikely and it is difficult to find proven instances of these activities. Stuart highlights that the most famous secondary uses of clinical trial data (for example reanalyses of paroxetine and zanamivir data) exposed shortcomings in the original reporting, providing information of direct benefit to patients.
Importantly there are a number of safeguards that can be put in place that prevent the misuse of data. Stuart provides the example of the Yale Open Data Access project (YODA), which is typical of other data sharing portals. Here, if researchers want to gain access to data from partnering data holders, they must complete a training module and submit a research protocol and statistical analysis plan, conflicts of interest and additional information. Proposals are reviewed for scientific merit and those applicants that gain access must agree to having their identity publically disclosed alongside their scientific aims.
In addition to highlighting efforts to prevent data misuse, Stuart also proposes ways to combat fears over the loss of publication opportunities. As well as being evaluated based on the number of self-authored publications, he suggests that researchers are also rewarded for creating and making public a data set that is useful to others in the field. For example, by acknowledging the creator of a dataset that is used for 20 publications as if they were an author on those publications. Similarly, he advocates awarding grants to applicants with a track record of making data available in trusted repositories. As a result, it will be in researcher’s own interests to share data. Such incentives, alongside measures to prevent data misuse, will ensure that increased data transparency benefits all.