The recent Public Accounts Committee report generated significant social media commentary, and I was pleased that many commentators referred to the ABPI commissioned study assessing rates of clinical trial disclosure by industry.
This study, published in the peer-review journal Current Medical Research and Opinion (CMRO), is one of the most current studies to address this question and is significantly different to a number of previous studies that have also addressed this topic. Yet these differences are not often captured in articles about rates of disclosure of clinical trial information by industry and this can lead to conflicting and potentially confusing messages on how ‘transparent’ the industry is. The most common citation being that ‘half of all trials are not published’. It’s interesting to note that if a figure is repeated often enough, it is more likely to be remembered regardless of its accuracy.
Therefore, in this blog I want to talk about the ABPI commissioned research in more detail and articulate why the figures below, taken from the CMRO study, are the most representative benchmark of clinical trial transparency by the industry to date.
Over three quarters (76%) of all industry-sponsored clinical trials for new medicines recently approved by the European Medicines Agency (EMA) had some results disclosed within a year of completion or of regulatory approval.
Rates of disclosure have continued to rise and almost nine in ten (89%) of these trials had disclosed results by 31 January 2013.
Before I look at this study in more detail, it is important to examine the differences between this and studies that have come before it. Many of the historic studies that look at clinical trial transparency covered trials conducted during specific periods between 2000 and 2008, so the figure is somewhat out of date and are not directly applicable to any individual medicines which have been approved for use in patients (for example Ross et al 2009, Bourgeois et al 2010, and Ross et al 2012). They also usually have investigated only one of the following:
So comparing results from study to study is often very challenging, subject to many limitations and a little like comparing ‘apples to pears’. Hence figures used to support claims such as ‘only half of all trials are published’ are misleading and may substantially underestimate the overall proportion of trials whose results are actually available in the public domain. For example, the study by Bourgeois et al 2010 (one of those used to support the claim that half of trials are not published) notes that although 66.5% of industry-sponsored trials were published, results for many of those not published were available electronically – combined this gives a disclosure figure of 88%.
So what is the ABPI commissioned research showing us that others aren’t? In the section below section, I highlight some key points about the study which people interested in the issue of clinical trial transparency need to understand.
In the ABPI study we investigated all industry sponsored clinical trials conducted in patients on all new medicines approved for use in Europe (by the European Medicines Agency) during the three year period from January 2009 to December 2011 inclusive. In doing so, we sought out all trials related to each of these 53 new active substances, not just those found in a single registry, (and whether or not they had been submitted to regulatory authorities). We then looked for disclosure of the results, either through publication in the literature or through posting of summary results on clinical trial registries. Due to the nature of the lengthy clinical development process, although we were examining recently approved medicines, the supporting trials had been conducted over a period extending back to the late 1990s and sometimes even earlier. Therefore, it was a good sign of progress that the majority of trials we assessed were posted on one or another of the registries we examined (even though public registries have only been available since around 2000, when the US NIH first released the ClinicalTrials.gov website).
Our study showed that, for medicines recently approved for use, the results for only 11% of trials remained undisclosed at the end of January 2013 and, particularly for those drugs approved for use in 2010 & 2011, the figure is now below 10%. The key studies on which regulatory authorities base their evaluation of the benefits and risks associated with a new drug are the large controlled trials carried out in Phase III of the drug’s development. Our study showed that only 8% of these trials related to recently approved drugs remained undisclosed at 31 January 2013 and, for the drugs approved in 2011, this has fallen to only 1%.
The ABPI commissioned research covered new medicines which had been developed during a period of transition. Prior to the availability of public registries on which summary results could be posted, important results of clinical trials were made public through conference presentations and publications in scientific literature. Historically, small early phase studies (which serve primarily to inform decisions about whether to progress drug development to larger more significant Phase III studies) were often not acceptable for publication as single research papers, while the majority of Phase III trials would have been routinely published if they supported regulatory applications.
Over the last decade or so, it has become standard practice to provide advance details of all new trials in public registries, and such registries are becoming the first repository for summary results, especially for the smaller Phase II trials. Thus disclosure of the results of the older trials in our study would be more likely through publication in the scientific literature, while first disclosure of the results of more recent trials is more likely to be on registries. When looking for public disclosure of results it is, therefore, important to search a variety of sources, which the CMRO study undertook.
Finally, it is noteworthy to us that two reputable journals, (both of which have been actively involved in the on-going discussion around clinical trial transparency) rejected our manuscript without even conducting a peer review.
Whilst this study highlights a positive trend of increasing levels of disclosure for industry-sponsored clinical trials, it shows that more remains to be done. The ABPI is committed to continue its work with key stakeholders to ensure an international framework for clinical transparency across the breadth of clinical research, both industry and academic, is implemented. For more information about our activities in 2014 see my next blog on this subject.
Dr Bina Rawal
ABPI Research, Medical and Innovation Director