Blog – Availability of clinical trial data for research purposes

19 Apr 2017

By Kevin Wing.

I recently attended a session entitled EMA’s Policy on Clinical Publication: Opportunities, Challenges and Measuring Success at the Drug Information Association annual conference in Glasgow, (still just about) UK, focusing on progress of the European Medicines Agency’s (EMA) new policy for making complete clinical trial data available for access by the public and researchers (Clinical Data Publication, also known as policy 70) at the website: I am about to start work on an electronic health records project that relies on the use of individual patient data from a clinical trial that I will be requesting from the website, and wanted to understand similarities/differences between each site.

The EMA’s Anne-Sophie Henry-Eude (Head of document access and publications) firstly gave an overview of the policy and website. Active for 5 months, trials relating to 13 regulatory applications are currently available online. The data provided include the actual clinical study reports and summary documents/overviews relating to the reports, as well as an anonymisation report, detailing the extent of redacting that has been performed in order to remove confidential information (such as participant and clinician identity) prior to release. Initial usage of the site shows that there have been 1600 general user accounts and 400 academic accounts created since launce.

The pharmaceutical industry perspective was then provided by Susan Forda, VP of International Regulatory Affairs at Eli Lilly & Company Ltd, UK. Susan highlighted that the EMA initiative was part of the evolution of industry transparency for clinical trial information in the EU but said that for research use, might be a better source of clinical trial data (although did not provide details).  Susan emphasised throughout her presentation that there was a large additional cost and resource burden being placed on pharmaceutical companies in order to make the full clinical trial reports ready for the EMA site.

Finally, the view from academia was provided by Tom Jefferson, Senior Associate Tutor at the Centre for Evidence Based Medicine (Oxford, UK), and editor for the Cochrane Acute Respiratory Infections (ARI) group. Tom’s focus was on how much more useful the full study reports would be for preparation of Cochrane systematic reviews than journal articles reporting clinical trials. He presented results of a survey of 31,901 Cochrane authors (around only 2000 of whom are “active”). Of the 156(!) who replied, 80% agreed that (full clinical trial reports from) regulatory data should be used when preparing Cochrane reports, but 67% identified barriers to using the data, with 58% unaware of how to access the data. In concluding remarks, Tom stated that he thought there should be a complete paradigm shift from the use of publications for the synthesis of evidence for reviews to the use of full clinical trial data. Tom’s final slides included a pointer to the Complex Reviews module of the MSc in Evidence-Based Health at Oxford University (, and a  number of additional clinical trial data sharing website references – the International Standard Randomised Controlled Trial Number (ISRCTN),,  the WHO International Clinical Trials Registry Platform (ICTRP), the +AllTrials campaign ( as well as

Although the session did not provide the detail I need to answer the question about the difference between and, my overall impression is that is likely to be a more straightforward way of obtaining individual patient data from the trials for research, while the EMA’s site focuses more on making the clinical data to support a drug licencing application as transparent as possible so that different stakeholders (i.e. general public, clinicians, researchers) will be able to understand how the EMA reached a final decision on whether to approve the application or not. This provides information that is likely to be very useful for the preparation of Cochrane reviews (such as the full trial protocol), but may not be as immediately useful for researchers aiming to attempt to create populations within observational research data that are as similar to the trial populations as possible for testing the ability of observational methods to replicate trial results. I will soon have some experience of obtaining data from and will hopefully blog again within the next few months re: how easy it all was!