37 Disclosing the results of clinical trials: how is the pharmaceutical industry doing?

Abstract

ObjectivesTo evaluate disclosure of clinical trials registered by pharmaceutical companies using an independent, semi-automated tool (TrialsTracker; https://trialstracker.ebmdatalab.net/#/).MethodFor the top 50 pharmaceutical companies (2014 global sales; EvaluatePharma, London, UK), registered interventional phase 2–4 clinical trials completed in 2006–2015 were identified in TrialsTracker, which calculates annual disclosure rates for sponsors of over 30 studies registered on ClinicalTrials.gov. The proportion of trials with results disclosed by April 2017 was analysed by company membership of the European Federation of Pharmaceutical Industries and Associations (EFPIA) and Pharmaceutical Research and Manufacturers of America (PhRMA).ResultsIn total, 323 clinical trial sponsors were listed in TrialsTracker, of which 69 were pharmaceutical industry sponsors and 31 were ranked in the top 50 pharmaceutical companies. Of these, 25/31 were EFPIA/PhRMA members and 6/31 were non-members. The disclosure rate for each year from 2006 to 2015 was 42.9%, 54.4%, 81.0%, 86.1%, 84.6%, 87.2%, 89.3%, 82.1%, 84.1% and 73.4%; reporting of clinical trial results became mandatory in 2008. The disclosure rate (disclosed trials/eligible trials) between 2006 and 2015 was greater for all pharmaceutical industry sponsors (7037/9511 [74.0%]) than for non-industry sponsors (9074/19866 [45.7%]) (p<0.01). For the top 50 companies, results were disclosed for 4761/6235 trials (76.4%) between 2006 and 2015, with similar disclosure rates for EFPIA/PhRMA members (4336/5697 [76.1%]) and non-members (425/538 [79.0%]).ConclusionsAccording to TrialsTracker, the pharmaceutical industry has disclosed the results of three quarters of trials completed between 2006 and 2015. The disclosure rate for pharmaceutical industry sponsors is greater than for non-industry sponsors. Because TrialsTracker excludes sources other than ClinicalTrials.gov (e.g. company websites), this figure may be an underestimate.