The idea of RIAT was born in a specific time period: the early part of this decade. This was when we started realising that clinical trials published in traditional biomedical journals were in the best case scenario extreme summaries of clinical study reports and other documents which contained extensive descriptions of the design, conduct and results. The underlying richness of detail and data, which was astonishing for those who had been used to journal articles, allowed two important aspects of scrutiny. First, we could do many more analyses and look at traditionally neglected aspects, such as harms reporting. This was not possible with journal articles. Second, we and other groups started noticing what we called discrepancies between the content of clinical study reports and their published counterparts. The discrepancies were misreporting in most cases.

When these discrepancies were pointed out to authors, sponsors and editors, no one seemed to take them seriously. We called this abandonment. Then there were invisible trials, those we knew for certain had never been published. We knew because regulators had started releasing previously secret clinical study reports.

We consider the accuracy of the published word so important in science that we thought we had to do something. RIAT was born.

As our website shows, we have a list of important or infamous trials that have been restored and are being restored thanks to RIAT and latterly the generosity of our funders, the Arnold Foundation.

The list comprises full restorations, like that of GSK study 329 on the anti depressant paroxetine.

But full restorations are very time consuming as they entail the initial call for action procedure and then reformatting and re-analysis of the content of a clinical study report into a journal article openly linked to the original report which could be thousands of pages long.

This could act as a disincentive to researchers who are not familiar with such material and do not have access to ghost writers. In addition, we came to realise that the misreporting may only affect a small but crucial part of the trial. We had no reason to spend a lot time restoring other parts which did not need restoration.

This is why we decided to experiment of what we call “focussed RIATs”, looking at the apparently small detail of misreporting that could change the interpretation of a trial or of a series of similar trials in what is known as an evidence development programme.

It may come as no surprise to learn that we are grappling with the misrepresentation of a control as a placebo or inert substance, when in fact the “placebo” turned out to have active properties and possibly confound the assessment of harms.

In addition, thanks to a landmark decision by the Federal Court in Canada we have access to the set of case report forms for the same biologic substance. We are looking at how causality of deaths across the registration trials was ascertained and the quality of record keeping.

Focused yes, easy to publish – so far – no. The first focused RIAT was rejected by a series of journals, including the one that published the original distorted trials. But then, we are used to this and it is the reason why RIAT and similar initiatives need your support if we are to save the credibility of biomedical journals and science. We need your support to spread the word, take on some restorations (full or focused), identify likely distortions or invisible trials and put pressure on the scientific community to act instead of producing articles which point fingers, but provide no solutions.

Tom Jefferson
Rome, 10 December 2019