Restoring invisible and abandoned trials: a call for people to publish the findings

Sean Lynch, in a rapid response, cautioned that there are a variety of benign or legitimate reasons that a clinical trial may not have been published. It actually appears to me that Doshi et al have anticipated this possibility in their paper: "3. Issue a “call to action” by publicly registering your possession of data sufficient for publication: ... This declaration offers original sponsors and trialists an opportunity to publish or formally correct their studies within the next 365 days..."


But counter-examples don't get around the fact that large numbers of clinical trials have been kept "invisible" for strategic reasons. Nor do they address the equally large number of "abandoned" publications of clinical trials with deceptive analyses and unsupported conclusions. The thing that defines our academic journals and studies as "academic" in the first place is peer review, both before and after publication. Academic Freedom means the freedom to debate, to criticize, to offer alternative analyses, to expand the vision of individual scholars to that of "the academy" of scholars as a whole. Withholding trials from publication for whatever reason is like withholding the raw data of the trials. It insures that there can never be peer review by the academy. In either case, they cannot be called "academic."

Doshi and colleagues (1) raise important challenges to publication bias and the standards of trial reporting. I agree with many of their suggestions and these have already been highlighted in other journals such as the Lancet (2) and by initiatives such as the Unpublished Trials Amnesty Repository of the Cochrane Skin Group (3).

I suggest that there are other more subtle finacial factors and human factors and possible design errors (only becoming obvious in retrospect) that are explanations for "invisible" or "abandoned" trials which need consideration.

1. There is a change in the business of the sponsor so that a drug goes "off patent" or further R&D with this product is no longer a commercial priority and the investigator cannot renegotiate funding or other funding streams
2. The research team are affected by human resource issues such as sick leave, change in personnel and timelines for completion of the trial are no longer realistic.
3. There are changes in working conditions of investigators or in health service delivery between the successful application for funding and intended start date, that affect recruitment into the trial.
4. There are unanticipated barriers to recruitment in the study. This might happen even if the trial was thought well-designed and subject to rigorous peer review before launch. I am aware of publicly-funded trials with such difficulty.
5. Advances in science can paradoxically make a design either obsolete, unlikely to give useful conclusions or unethical.
6. A final consideration might be that trials are not abandoned as such but postponed - and this has not been clearly stated in the trial register (I put my hand up for one of these).

As well as publishing the trial, or at least the protocol, with which I agree; it is important that the difficulties facing the investigator are highlighted.

An issue I sometimes have with the conclusions from secondary research is that this dimension affecting feasibility of research is not captured. It is all very well identifying the need for a trial and likely methods from analysis of previous trials, but I feel that authors should also report delays or recruitment difficulties. If not, there is a danger of making a bigger and better mistake!

1.Doshi P et al
Restoring invisible and abandoned trials: a call for people to publish the findings BMJ 2013;346:f2865
2.Tfelt-Hansen P
Unpublished Clinical trials with Sumatriptan
The Lancet, Volume 374, Issue 9700, Pages 1501 - 1502, 31 October 2009
3.Unpublished Trials Amnesty Cochrane Skin Group
csg@nottingham.ac.uk