Andrew Burd: Ethics and clinical trial registration

Andrew Burd These days I do not so often have a gauntlet dispatched in my direction, but when one lands at my feet I have an irresistible urge to pick it up.  By way of explanation, one of my passing pleasures is to engage in debate and discussion in the WAME list serve, a discussion forum for members of the World Association of Medical Editors (WAME).

Questions are posed and discussion ensues, sometimes ending in a consensus position, sometimes not.  It is, for me, an opportunity to ponder topics outwith my daily routine, and occasionally I find myself challenging the unchallenged assumptions of a somewhat privileged elite. 

And so a question was posed regarding clinical trial registration. In essence an editor had received a report of a randomised controlled trial; the study was approved by the local ethical committee; it had been well conducted and the paper was of good quality; but there was no clinical trial registration number.  When the authors were asked about this they responded with a registration number with a date indicating this had been assigned retrospectively.  The editor of the journal was wondering what to do.

It is not a matter of virtue but a fact of the celestial universe that, living in the Far East we start our day before those in the West.  As such I am invariably the first to comment on new questions for debate posted in the WAME forum.  The matter under discussion appeared to me quite straightforward and my pragmatic response was – accept. 

My subtext was “What is the big deal?”  The answer to that question came from some subsequent postings, where the implications of clinical trial registration began to emerge.  It became apparent to me that we were, as is so often the case, entering the arena of confrontation between common sense and vested interests.  I suggested that not all clinical trials are equal and questioned the ethical and indeed moral basis for mandatory registering of all clinical trials, unaware that such a position is like wafting a red flag in front of an elderly bull. 

The “bull” in this case was not Richard Smith or indeed Richard Horton, although both waxed lyrical on this topic as long ago as 1999.  (BMJ 1999;319:865-6).  In a joint editorial they began with the statement “The case for registering all clinical trials – first advanced a decade ago – is now unanswerable.”  I am not writing to offend or upset such luminaries in the field of medical publication.  What I want to do is to look at a complex situation and ask for more precision in definition, debate and understanding. 

It has been put to me, and I quote, “Inviting patients to participate in clinical trials implies a duty to ensure that the results of their participation should be made public.”  My perspective is that there are those who take a very blinkered and absolutist attitude towards semantics; and, to crudely paraphrase, Gertrude Stein would say, “A clinical trial, is a clinical trial, is a clinical trial.” 

I disagree.  Indeed I see that there are clinical trials that are well designed, randomised, controlled, relevant but very, very different in context and consequence.  Unless and until this is appreciated, editorials such as that written by the ‘Richards’ and crusades relentlessly pursued by the blinkered elite will continue.  And they will fail to generate the response that they deserve because the authors, the crusaders, have failed to just step back and consider what clinical trials really are.

 The spectrum is vast but I see three very distinct types of trial.  I see a trial created by an individual doctor who is trying to determine which of two options is better for his patients.  The patients are aware of the dilemma; they participate and on the basis of the collaborative efforts of an informed group of patients and their doctor, an ‘evidence based’ consensus can be achieved.  Another form of clinical trial does not result in an outcome that can be qualified as good or bad.  These are the trials that look for a mechanistic effect that may lead to better patient selection, or prognosis.  Again, this is a trial that is predicated on the ethical doctor trying to do the best for his or her patients in terms of favouring management choices or information analysis.

I think the critical element in these two scenarios is the ethical duty of care of the investigator to his/her study population/patients.  Of course they need to know the results but what is the logical imperative that mandates that the investigator has to register the trial and report the results.  The supposition is that doctors who conduct clinical trials out of their ethical duty of care to their patients (we call this beneficence, by the way) are of dubious integrity and may suppress results they do not like.  What nonsense, and isn’t such a view rather insulting to the majority of honest and ethical doctors?!

No.  The clinical trials that Horton and Smith were referring to are a different matter altogether.  These are trials where money is a stake.  They may even be trials where there is no intrinsic doctor patient relationship.

That is to say that they are usually industry sponsored where the doctor acts as an agent, usually for a fee, to recruit and treat patients.  I absolutely agree that these trials should all be pre-registered and the results should be made public.  Question?  Does that mean NOT publishing them in restricted access journals?  How will that affect impact factors?  But back on point, can arbitrary guidelines be drawn up to define which clinical trials mandate pre-registration and public reporting?

Might I suggest

1) Any industry or institutionally funded trial
2) Any trial with more than 50 subjects

That should do for a start.  I hope my position is clearer.

Let us not call for blanket registration or even reporting of all clinical trials.  Let us specify which trials must be registered and subsequently reported.  Is that so difficult?

Competing interest

I have just received ethical committee approval of a trial where I shall be looking at the micro RNA response to two low dose formulations of standard medications.  It is a randomised prospective, clinical trial and I have been asked to register the study.  This involves filling in many details on a multi-page form, details that would be appropriate if these were new drugs, but in the context of this trial details which are completely irrelevant. 

To add insult to injury I see that a recent paper in JAMA is now acknowledging that even if trials are registered much of the information suggests they are of dubious quality (JAMA 2012;307:1838-1847) and the whole clinical trial registration “ideal” is being questioned.