New study: European clinical trial registry fills significant medical evidence gaps

In this guest blog, Nicholas DeVito discusses his recent research showing that many clinical trial results are only known to patients and the scientific community thanks to the European EudraCT registry.

Background

[A paper we recently] published in BMJ Medicine builds on our prior work examining the EU Clinical Trials Register (EUCTR). The EUCTR has grown substantially as a repository for results of clinical trials since we first started our EU TrialsTracker in 2018. This work sparked hearings in the UK and an international advocacy effort to improve compliance with EU trial reporting guidelines which has paid dividends.

As of this month the EU TrialsTracker shows just under 85% of all trials we can assess as “due” with results. We therefore thought it would be useful to quantify the utility of the EUCTR as a results repository.

After all, even with the arrival of the new CTIS registry, the EUCTR archive will contain registrations for nearly 44,000 trials of which 23,000 currently have a result attached to them. It would be useful for anyone with an interest in efficiently searching, and collecting, the results of clinical trials, such as researchers conducting evidence synthesis, to know if the EUCTR is worth checking.

Prior work on ClinicalTrials.gov had shown that for many trials registered there, the results submitted to the registry are the only public version of results that exist.

Methodology

To do this we identified completed trials, and trials that look like they should have completed but had incomplete data on the EUCTR. This left us with around 25,000 trials. We then sampled 500 trials and began searching for results. We checked the EUCTR, ClinicalTrials.gov, the ISRCTN, and the literature of results using a comprehensive search strategy. If we found evidence a trial never happened, we replaced it in the sample.

Results

We found that 383 trials (77%) had results somewhere we looked.

Results appeared on the EUCTR (53.2%) at about the same rate as they appear in the literature (58.6%). Among those 383 trials with any result, 55 (14%) were only available on the EUCTR, while 28.2% were only available in a journal publication. For those trials also registered on ClinicalTrials.gov, only 3 only had results available on that registry. Results also generally appeared on the EUCTR as quickly as the literature.

Overall, the EUCTR seemed to reliably contain the timely results of European-based clinical trials at rates commensurate, or better than, other common dissemination routes. So what does this mean?

Discussion

We believe this is good evidence that the EUCTR should be specifically recommended in guidance to systematic reviewers, such as the Cochrane Handbook. The database will not only contain some results not available elsewhere, but also may increase the efficiency of searches for results compared to standard literature searches.

However, given the amount of trials with no results (n=117) or with results elsewhere but not on the EUCTR (n=117), this is a clear indication that things could be even better.

The European Medicines Agency, and EU/EEA member states, should take care to ensure the new CTIS registry improves on the standard set by the EUCTR and take full advantage of the powers granted to them in the new EU trial regulations to penalise noncompliance.

This blog by Nicholas J. DeVito was originally published under the title "What is the utility of the EUCTR?" on the website of the Bennett Institute for Applied Data Science. The full paper is available here. TranspariMED was not involved with the research project described above.