Retrospectively registered trials: the Editors’ dilemma

Prospective clinical trial registration aims to address publication and reporting bias. Unfortunately, not all clinical trials are registered before they start. Here we discuss the dilemma faced by editors when receiving submissions reporting a clinical trial that was not registered prospectively, and a new policy for increasing transparency when a trial was registered after participant recruitment has begun.

1

Registering clinical trials in an appropriate trial registry before the first participant is enrolled into the study creates a publicly-available record of the researcher’s intentions; including key details such as how many patients they need to recruit, what their outcomes will be, how they intend to measure these outcomes, etc.

The existence of this publicly-available record is vital to address the well-documented problems of publication bias and reporting bias. Without prospective trial registration, researchers could decide to change their outcomes once they have analyzed their results, omit results that don’t ‘fit’ with their hypotheses, or not publish them at all if they really don’t like what they find.

Distortion of the medical literature in this way is a very real problem for everyone. It affects the evidence base on which doctors make clinical decisions when they see their patients.

In an ideal world, all clinical trials that are carried out would be registered before the first patient is enrolled.

The real world

Unfortunately, we do not live in an ideal world and, for many reasons, not all researchers publish their findings or register their studies before they start.

This is a problem because people who give up their time to take part in research expose themselves to the inherent risks of clinical research. They do this in the belief that the findings will be published and will benefit others.

The Editors’ Dilemma

Receiving a manuscript reporting a trial that was not registered before the study started presents a difficult dilemma for the editor. What should they do? Reject the manuscript? Consider it for publication? There are ethical arguments both ways.

On the one hand, considering the manuscript could undermine the aims of trial registration – to guard against publication and reporting bias and their effects on the literature/evidence base. It may also remove the incentive for researchers to register prospectively through the knowledge that they will still be able to publish their results.

On the other hand, there is an ethical responsibly to publish research that human participants have taken part in. Researchers may genuinely not know that they need to register their study before patient recruitment begins, or not realize their study fits the WHO definition of a clinical trial, which is not confined to randomized controlled trials.

They might have conducted valuable research which could hold potential benefits for others. There is also a question of where to draw the line, is registering a month into recruitment ok, but after analyzing the results not ok? What if they registered just before they started analyzing their results?

Neither scenario is perfect, but this is a very real dilemma which editors grapple with on a regular basis.

Neither scenario is perfect, but this is a very real dilemma which editors grapple with on a regular basis.

BioMed Central journals consider retrospectively registered studies on the basis that rejecting straight off all studies that were not registered prospectively would lead to vast amounts of research involving human participants going unpublished.

How widespread is retrospective registration?

Along with my colleague Jigisha Patel, Associate Editorial Director (Research Integrity), I carried out some research to find out the extent to which retrospective registration occurs. To do this we looked at all clinical trials published in a one-year period in the BMC series.

To our surprise, retrospective registration is widespread and prevalent in the journals we looked at. 67% of the studies identified were registered retrospectively, ranging from registration within a few days of participant enrolment right through to registration after submission to the journal.

Our new policy

To prevent non-publication of research involving human participants, we will continue to consider retrospectively registered studies, but want to do this with complete transparency. Therefore, in light of our findings, we are updating our editorial policies.

For all manuscripts reporting the results of clinical trials, authors will be asked for the date of registration and the date of enrollment of the first participant. For all studies that were not registered prospectively, authors must include the words ‘retrospectively registered’ after the trial registration number at the end of the abstract.

A long way to go

There’s still a long way to go, and increased transparency alone when a trial has been retrospectively registered is not sufficient to solve the problems. Much more needs to be done to improve the rates of prospective clinical trial registration. We all play a role in this – institutions, funders, ethics committees, journals.

Much more needs to be done to improve the rates of prospectively clinical trial registration.

I look forward to the day when all clinical trials are registered prospectively and editors no longer face the dilemma of whether to consider a retrospectively registered study.

However, until that day arrives, we all need to work together to increase prospective registration, encourage registration of all studies that have been started but not registered, and ensure that where a trial was registered retrospectively that this is transparent and clear to the reader.

View the latest posts on the Research in progress blog homepage

One Comment

Bernhard Uehleke

In case of retrospectively registered trials, it might help to check the study plan registered, which that one, which was filed for the ethics comittee. Official ethic comittees should store the application doscuments and give out information in unclear situations. Studies, which miss prospective study registration and votum and documentation of an ethic comittee bear possibility of bias and misinformation.

Comments are closed.