Leading UK medical universities are failing to publish the results of clinical trials on registries, making data harder to find, researchers claim.

TranspariMED, an organisation campaigning for more transparency in medical research, said that it had found the outcomes of only 206 (5.8 per cent) of the 3,540 clinical trials conducted by researchers at 16 top institutions on the main European and American clinical trial registries ¨C EudraCT and Clinicaltrials.gov ¨C where they were originally registered.

claims that the results of the remaining 3,334 trials were ¡°difficult or impossible to find¡±, raising the risk that repetition of experiments may waste resources or that the safety or effectiveness of a new drug or treatment may not be fully understood.

Universities have questioned the study¡¯s methodology and conclusions.

É«ºÐÖ±²¥

In the UK, universities are required to post the results of drugs trials on registries within 12 months of completion, but trials of medical devices and non-drug health interventions are not covered. Nevertheless, TranspariMED says global best practices recommend the posting of all trial results.

Of the institutions assessed,?Keele University?had the highest proportion of results of clinical trials registered, with results for just under a third (29.4 per cent) of its trials on the registries. The?University of Dundee?had 23.2 per cent of its trials (49 out of 211) registered.

É«ºÐÖ±²¥

But TranspariMED was unable to find results of 564 trials involving Imperial College London, meaning only 3.4 per cent of the institution¡¯s clinical trial results were on the registries. Other institutions with a large number of trials not found on registries by the researchers include the University of Oxford (544 trials not on the registry, 7.8 per cent registered) and University College London (466 trials not on the registry, 5.9 per cent registered).

Till Bruckner, founder of TranspariMED, said that while his ¡°guess would be that the majority of these trials have reported results somewhere in some form¡±, they were ¡°incredibly hard to find¡±.

There is a risk of producing ¡°drugs whose effectiveness and safety may not be fully understood¡±, and doctors giving ¡°mistaken health advice¡±, Dr Bruckner said, because they ¡°lack access to information that¡¯s really important for the decisions they take on how to treat an individual patient¡±.

Posting summary results on registries and putting trial numbers into the abstracts of research papers ¡°would solve 80 per cent of the problems¡±, Dr Bruckner added.

É«ºÐÖ±²¥

Universities have said that the TranspariMED study includes trials that are not due to report yet or had been published elsewhere.

An Oxford spokesman said that, while the report ¡°implies that the majority of results from Oxford researchers are not being reported¡±, this is ¡°simply not the case¡±.

¡°We thoroughly support the publishing of clinical trial results, whether positive or negative,¡± the spokesman said. ¡°It is standard practice for research teams to perform a full analysis of trial data before publishing, to look at all factors affecting the results, which can lead to a lag time between a trial finishing and a paper being published.¡±

An Imperial spokeswoman said that the institution was ¡°currently reviewing its standard operating procedures concerning the registration of clinical trial data on public databases¡± and ¡°will in future be advising its researchers that, in addition to registering trials before patients are enrolled, they should publish results in the registry once the trial has concluded¡±.

É«ºÐÖ±²¥

A UCL spokesman said the university ¡°actively encourages every trial to be registered on a public database, and facilitates the practice as part of operational processes¡±.

john.elmes@timeshighereducation.com