THE presence of the University of Dundee on a list of offending institutions which sponsor medical trials but don't publish the results was a surprise.

But there the respected University is, listed on the new TrialsTracker, promoted by campaigners for transparency in medical research.

The tool has been developed by the AllTrials.com campaign, which says failure to publish research findings - even if they are negative or inconclusive - holds back science and is unethical.

The TrialsTracker compares the official database of registered trials with results listed US-based services PubMed or ClinicalTrials.gov.

Institutions with high levels of studies which have not reported results two years after the trial concluded are named and shamed on a new website.

Listed alongside the more traditional pantomime villains of Big Pharma, such as Novartis, which currently has 37.6 per cent of the trials it has sponsored unpublished, GlaxoSmithKline (22.6 per cent) and Bayer (39.7 per cent), sits the University of Dundee - which has sponsored 35 eligible trials since 2006, of which 21, or 60% have apparently not published results.

This puts it just outside the top 100 offenders in the world, with a record marginally worse than that of the International Centre for Diarrhoeal Disease Research in Bangladesh.

The University disputes the methodology which campaigners have used.

But the system was developed by academics at the University of Oxford's Centre for Evidence-Based Medicine. One, Bad Pharma author Dr Ben Goldacre, says it is an attempt to tackle the scandal of unpublished trial results.

Clinical trials involving patients and healthy volunteers are the best way of testing whether a medicine is safe and effective. But trials with negative results are twice as likely to be unpublished as those with positive results.

At its worst, this can mean we are all denied knowledge about the potential benefits and side-effects of a medicine. "Withholding the results of clinical trials is unethical, because it prevents doctors and patients making informed choices about which treatment works best," Dr Goldacre said.

Sense About Science which runs the AllTrials campaign says we should also be outraged at the betrayal of 8.7 million patients who have taken part in unpublished studies over the last 10 years.

A spokesman for the University of Dundee told me it did not accept the findings, claiming that 28 out of the 39 clinical trials listed have in fact been published, while five have not yet published results, and some others were ongoing. "Rather than 60 percent of studies the more accurate figure would be 13 per cent," he said.

However Dr Goldacre is unpersuaded. All trials have had to use unique identifying numbers since 2006, he points out. "Unless there has been a spectacular error in the software, these discrepancies will have arisen simply because Dundee have failed to enter the correct information into the public trials register, and their academic publications.

"There are no ongoing trials in our dataset. So the data for Dundee University trials in the public trials registers must be incorrect."

Dundee shouldn't be complacent if any failure to publish has been inadvertent or a mistake he adds: "Errors in the public registers about trials at Dundee represent a serious shortcoming by the University and its researchers."

If Dundee has been included unjustly, is it a price worth paying to bring the issue to the fore? Ensuring the missing trials are correctly published will update the tracker, and remove their unwanted listing, Dr Goldacre points out.