What we work on
The November 8 issue of the British Medical Journal (BMJ) features an article titled “60 seconds on TrialsTracker”. TrialsTracker is a website that reports the publication status of clinical trials registered within the last ten years on the ClinicalTrials** website. It provides information about medical entities regarding the number of trials they registered and the number of trials they published.
The BMJ article focuses on the role of drug companies in the sad story of unpublished trials. It points out that three of the five entities with the lowest numbers of published trials are the drug companies Sanofi, Novartis and GlaxoSmithKline. The other two worst performers are a government institute and a public hospital cooperation. This goes well with the public perception that drug companies are publishing only what fits into their marketing strategy, but hide data that might help patients without benefit for the companies. On the other hand, the top performer that has published all registered studies is Shire, also a pharmaceutical company. Though the BMJ author acknowledges this fact, he does not mention that the next 17 ranks for most complete reporting are also held by drug companies.
The inquisitive user will find that TrialsTracker is very helpful to dig deeper into the data. It is possible to change the reporting order of the entities according to the percentage of studies published versus those registered. Now the first 22 ranks of the top performers are pharmaceutical companies. The first non-commercial entity appears at rank 23. Although the worst performer again is a drug company, Ranbaxy Laboratories, with no trials published at all, the next 15 lowest performers in terms of percentage of published trials are universities, hospitals, and government institutes. Thus, when looking at the percentage of studies published versus those registered instead of absolute numbers, the reputation of the drug companies improves.
So is the pharmaceutical industry just a poor victim of biased reporting? Possibly not: the data available at TrialsTracker is exclusively based on the studies registered at the ClinicalTrials website. And somehow, the 100% reporting rate of Shire seems to contradict common experience with clinical trials, with some of them ending up with no reportable results. A quick look into the EudraCT database# shows that Shire, which has 96 trials registered at clinicaltrials.gov, has 153 trials registered in EudraCT. Here, several trials show the status “removed from public view” in the results section. Therefore, Shire might perform worse than it seems on TrialsTracker. Hopefully, the European Medicines Agency will soon make the EudraCT data available to everyone, or at least to TrialsTracker to promote transparency.
Taken together, the data on TrialsTracker seems to indicate that unpublished trials are not solely a problem of drug companies but of many different organisations that initiate studies. Transparency is a powerful tool to improve the efficiency of healthcare systems, and TrialsTracker can play an important role in it. However, if the data is used to assign blame publicly, it might fall short of its aim. Science is driven by trial and, maybe even more importantly, by error. This includes failed attempts to complete a study. The new tool shows a widespread failure to publish registered studies, and should trigger efforts to ameliorate the overall situation. If the media pick out certain groups to blame, it may create an atmosphere of fear that will reduce the willingness of specific medical entities to initiate studies.
* Trials Tracker
** ClinicalTrials
# Eu Clinical Trials Register