Back in August 2014 I wrote a blog about the AllTrials Campaign, and that our host Trust had signed up.
As I explained, AllTrials calls “for all past and present clinical trials to be registered and their results reported. Today, Oxford University Hospitals NHS Trust joins the 79611 people and 502 organisations that have so far signed the campaign. Alongside this central commitment to transparency and openness, the Biomedical Research Centre and Unit are launching a project to look back at their trials to assess publication rate, and identify how best to ensure publication of all trials.”
Fast forward 18 months, and we have just published our report of this project in BMJ Open, under the snappy title “Getting our house in order: an audit of the registration and publication of clinical trials supported by the National Institute for Health Research Oxford Biomedical Research Centre and the Musculoskeletal Biomedical Research Unit” and you can read it here.
So, what did the study find? In brief, we identified 286 trials from within the BRC and BRU since they began, but could not find registration for 4 trials, although they had been completed and published. Of the trials with a registered completion date pre-January 2015, just over half had been published, and half of these, published within 12 months.
There are many ways to look at the paper, including as a detailed description of an audit method that others can use to assess their own performance in trials publication, and see how and where it can be improved. It’s also confirmation that internal records are – with a bit of digging – a good data source for information on trial registration (now a requirement of ethical approval for trials) and publication.
As one of my co-authors, Senior Clinical Research Fellow Dr Ben Goldacre said: “It is testament to the integrity of the research units audited here that they were keen to collaborate on an audit of their publication performance. I hope that others around the world will join Oxford and use the audit tools we have produced, and shared, with this paper. Nobody can achieve perfection, anywhere in medicine, but audit is one of the most basic and powerful tools we have, to try.”
Another interesting assessment made in the paper was whether trials sponsored by industry were any less likely to be published than non-industry trials, and they weren’t. As we say in the paper “No clinical trial characteristics were found to be significantly associated with non-publication,” which is reassuring, given ongoing debate about whether industry is less keen than other sponsors to make trial results known.
To our knowledge, this is the first published audit of publication performance for major research institutions, but that is certainly not “job done.” In a section of the paper called “How do publication performance of trials supported by the NIHR Oxford BRC and BRU compare with other institutions?” we report that (roughly speaking) we may do better than average, but will have to strive to meet the World Health Organisation position statement that trial results should be publicly reported within 12 months of their end. We also need to decide how best to ensure that our unpublished trials are published; as we say in the paper “Based on the audit findings, the audit team and BRC/BRU Directors produced a set of recommendations to improve the transparency regarding which clinical trials receive BRC/BRU support and to encourage their timely publication. The BRC/BRU Directors have committed to improving performance in these areas, and we are confident that with some simple changes and renewed prioritisation, this can be achieved.” One such recommendation states “Contact investigators of trials overdue publication to investigate reasons for non-publication and together agree an action plan.”
In August 2014, I wrote “watch this blog for more news on the publications review, as well as the other commitment we are making today: that information about trials patients may be eligible to join is written in language they can understand.”
With this blog, I give you the results from the audit; call me to task if at some stage not too long from now we can’t tell you more about the progress of tracking the missing publications. And rest assured: our endeavours to ensure patients who might want to join trials can find out about them in ways they can understand are pretty ceaseless: it should be easy, but like so many such things, it is far from that. More on this soon.
When our host Trust signed AllTrials, we committed the BRC and U to action above and beyond simply endorsing the campaign as an ideal. I am delighted that the Patient Involvement Working Group funded this work, and that we have completed the audit and identified where, and how, we can do better to ensure the results of all current and future studies are published. Patients join trials for countless reasons and we owe it to them to ensure that results are openly shared with all who could benefit. More than half a million people take part in NHS trials every year; their contributions to medical knowledge must never be wasted by non-publication of results.