OBJECTIVES:  Although much of the clinical trial data generated by pharmaceutical companies in oncology are eventually released, there are concerns regarding the speed at which this information is disseminated. Thus, we conducted a study examining the delays in publication of clinical trial results and the availability of clinically actionable data in company press releases.

METHODS:  We identified peer-reviewed publications and meeting presentations for all clinical trials mentioned in press releases issued by the top seven companies in oncology sales and our pilot company, Amgen, between January 2011 and June 2016. Time to first publication from the availability of trial results was calculated. Availability of results was the earliest date among: initial press release, meeting presentation (minus either 120 or 90 days for regular or late-breaking abstract submission, respectively) or publication (minus 120 days). We conducted survival analyses using the log-rank test and Cox proportional hazards models.

RESULTS:  Across our sample of 101 clinical trials, the median time from the availability of trial results until the first journal publication was 382 days. The vast majority (70%) of releases reported positive results. For those which reported negative results, there was a longer delay to publication (median of 600 vs. 357 days, log-rank p<0.001) and the result remained significant in a model controlling for company.

CONCLUSIONS:  Our study reveals that there is a tremendous amount of information emanating from human subjects research on cancer drugs that is not finding its way into the public domain in a timely fashion. These delays negatively affect both patient outcomes and scientific innovation. We propose two solutions to ensure rapid dissemination of data, including more consistent use of independent scientific preprinting and rigorous enforcement of regulations requiring that sponsors post trial results on public domains such as ClinicalTrials.gov.