BMA — Access denied: the availability of clinical data

"Positive clinical data sells drugs. But with only half of all trials making it to publication, the full picture is blurred. The BMA demands pharmaceutical companies stop withholding data If you flipped a coin 100 times but only revealed half the results you could persuade people heads had come up every time. When they found out you concealed the negative results they may not find it particularly fair. Yet this is exactly the way in which the results of many clinical trials are presented. Current legislation allows pharmaceutical companies to withhold unflattering outcomes from doctors and patients, resulting in misleading evidence about the benefits and risks of treatments.  It is estimated around half of all trials are never published in academic journals and those with positive results are twice as likely to see the light of day, according to the NHS National Institute for Health Research Health Technology Assessment Programme review 2010.   The BMA strongly opposes the withholding of trial data. BMA medical ethics committee chair Tony Calland (pictured right) says: ‘Access to all clinical trial results is essential to build the reliable knowledge base doctors need to help them prescribe safely and effectively for their patients.  ‘Where results go unreported, this betrays the altruism of research participants, risks repeating studies unnecessarily and threatens public confidence in vital medical research.’  Earlier this year the association signed the AllTrials petition, a joint initiative by the BMJ, Sense About Science, the James Lind Initiative, the Centre for Evidence-Based Medicine and Ben Goldacre’s Bad Science campaign, calling for all clinical trials to be registered and results reported.  The BMJ has asked pharmaceutical company Roche to release full clinical study reports for 10 treatment trials of antiviral drug Tamiflu, which was stockpiled by governments worldwide during the 2009 H1N1 flu pandemic. There is debate over its effectiveness as trial data was not released for independent analysis ..."