This post was published by Dr Carl Heneghan on 20th May 2013
If you hadn’t noticed, today is international clinical trials day. Celebrating the day, in 1747, when James Lind started his famous trial of treatments for scurvy amongst seafaring men. What was unique about Lind’s approach was the use of experimental controls to compare different interventions, six in total: all of the seamen involved in the trial received the same diet plus the addition of either cider, elixir vitriol, nutmeg, seawater, vinegar and the eventual cure, deemed to be a luxury at the time, oranges and lemons. It took only six days for the effective intervention to become apparent.
Six years later Lind published ‘A treatise of scurvy’, which like many of today’s trials was invariably ignored. It was some 20 years later that lemon juice became common-place amongst seamen.
What would have happened if Lind had decided not to publish the results? Keeping the results instead to himself and his own seafaring crew. Would the lack of publication impacted on the use of a highly effective treatment – or led to delay -for what was a devastating condition at the time?
Yet it currently seems we have not taken to heart the lesson of publishing all results, irrespective of their findings. It was with this issue in mind – and many more examples that include direct harm to patients – that the AllTrials campaign was launched earlier this year. It is becoming increasingly clear that all clinical trial results should be published and reported in full.
Indeed, it seems a sizable number of clinical trials have never been published. I am not even sure, if anyone really knows, exactly how many clinical trials there are.
Here is my best estimate, and feel free to criticize, add to or take away from the final number:
1) From a search of PubMed (using the excellent PuBMed Trend database); there are 450,555 RCTs published between 1965 and 2012;
3) EMBASE has been shown to retrieve about 16% more clinical trials than MedLine;
4) About a further 15% of studies are only found through hand searching and other databases;
5) Introduce a fudge factor of +/- 10% to allow for gross over estimates or currently unknown hideouts for unpublished trials .
My best estimate is therefore there are 1,202,080 RCTs (+/-10%) trials of which less than half are currently available for scrutiny.
The AllTrials campaign calls for the “information on what was done and what was found in these trials could be lost forever to doctors and researchers, leading to bad treatment decisions, missed opportunities for good medicine, and trials being repeated.”
All trials past and present should be registered, well at least those in the last 20 years otherwise it will take forever, and the full methods and the results reported.
More than 50,000 people, and 250 organizations, have already signed the petition:
I’m pretty sure James Lind would have signed it.