Reflecting on clinical trials and transparency, it is pertinent to ask whether, amongst the wide variety of clinical trials (of different designs, sizes, durations, purposes, sponsors etc), from the viewpoint of transparency, all trials are equally important.
Let’s not forget that the safety of patients is the paramount concern, and the drug development process within the modern pharmaceutical industry has been constantly refined over the last 70 years or so (ever since the trial use of penicillin in single patients) with this in mind. The sequential testing from single dose tolerability in small numbers of volunteers, through multiple dose safety and pharmacokinetics to the first assessment of potential efficacy in 20 or so patients is designed so that the process can be stopped if there is any sign of an unexpected effect. At each stage, the objective is to support the decision to move forward to the next stage or not. While disclosure of the results of these small trials is desirable, historically, on their own, these early phase studies were never going to command space in hard copy journals, and neither do they provide substantive evidence for how the medicine should be used in practice.
It is great news for everybody that the coming together of digital platforms with the commitment to improved transparency has provided the trial registries (particularly ClinicalTrials.gov) that are the ideal publicly-available repository for the results of these types of trial. But don’t forget that these registries have been available for less than ten years; and our recent results indicate that overall disclosure of industry-sponsored trials – including these small early phase studies – is now over 90% (ref).
So, in the discussion on transparency, it is critical to make a distinction between small trials which, designed to meet a decision milestone, were never expected to be published, and larger, late stage trials which, having failed to show the expected effect, were not published. Only the latter could impact a systematic review of a medicine’s value, and, looking forward, with the usage of registries mandated to track trials from inception to results, at least summary results of all such trials are now being made publicly available.
Early studies (from the 1700s onwards) in life-threatening diseases including infections and scurvy showed clear-cut benefits in small numbers of patients, both for treatments and in the use of inoculation and vaccination for prevention. With more complex diseases, the clinical endpoints are often less clear-cut and the use of surrogate endpoints (often laboratory tests indicative of an expected improvement), has led to an increasing reliance on randomised controlled trials (RCTs) in ever larger numbers of patients, with the aim of demonstrating statistically significant average improvements in outcomes.
However, the concept of RCTs retains an inherent flaw – the basic assumption that all human subjects are similar; however well-matched the samples, there is bound to be variation, and the most likely reason for failure of a phase 3 trial to demonstrate significant efficacy is that the patient groups included some who would benefit significantly and others who might not.
Such studies don’t help solve the daily challenge faced by the practicing physician – that of selecting the right medicine for the individual patient for whom the medicine’s benefit will outweigh the risk. We can hope that this may be improved through the development of stratified and personalised medicines, providing the costs can be covered somehow. But that’s another story.
Consultant, Livewire Communications