Media, September 2007

MMR – the real villains

Leaks, lies and statistics, and language of industrial strength in the British Medical Journal. A series of articles in July on autism and the MMR vaccine made for fascinating reading for anyone interested in journalism, and what psychologists do with their data.

It all started with a front-page splash in The Observer. The story, an astonishingly poor piece of journalism by Denis Campbell, made three key points: that new research has found an increase in the prevalence of autism to 1 in 58; that the lead academic on this study (Professor Simon Baron-Cohen, University of Cambridge) was so concerned he suggested raising the finding with public heath officials; and that two 'leading researchers' on the team believed that the rise was due to MMR. The story was then recycled in a multitude of national and local newspapers, reaching millions.

In fact, the figure was just one (the highest) of several in a report specifically designed to look at different rates produced by different methods; Baron-Cohen described his unpublished data as 'as accurate as jottings in a notebook', and reports of his concern as 'inaccurate and scaremongering'; and one of the 'leading researchers', Dr Fiona Scott, had to resort to posting online the rather key fact that she 'was never contacted by and had no communication whatsoever with the reporter who wrote the infamous Observer article'.

The Observer largely stuck to their guns – not once, but twice (tinyurl.com/2p3swx and tinyurl.com/ysebs8). (They did, however, remove the piece from their online archive: it should still be available from www.badscience.net/?p=468.) Bad Science's Ben Goldacre, from The Observer's sister paper The Guardian, was belatedly given the opportunity to trash the original report (see tinyurl.com/23fxdp, and in the British Medical Journal at tinyurl.com/3atu5d), adding his voice to coverage in The Times (tinyurl.com/yuzljy), and the website Spiked (tinyurl.com/3x5jkb).  The presumably small number of readers who saw this less 'populist' coverage stroked their beards and tucked into their muesli – another victory for science. But for me, ethical issues remain that are particularly important for our members.

Firstly, should psychologists ever talk to journalists about unpublished data? Ben Goldacre describes it as 'the antithesis of what science is about: transparency, where anyone can appraise the methods, and the results, and draw their own conclusions'. But isn't it a fairly common occurrence for psychologists to be nabbed by journalists after a 'work in progress' talk at a conference? With the snail's pace of most journal publishing, is it any wonder that impatience – from journalists and psychologists alike – pushes data into the public realm before it is 'ready'?

Baron-Cohen agreed with me 'that if a student or anyone else reports results in a public forum, like a conference, then they should expect that the press can repeat these. But a draft report in the process of being written for publication, or an end-of-grant report for a funding body, is not really an announcement in the public domain.'

Secondly, when psychologists communicate with journalists, should they make some attempt at insisting that potential conflicts of interest are communicated to readers? One author on the unpublished paper, Society member Dr Carol Stott, has moved on from Baron-Cohen's team and now works in Dr Andrew Wakefield's private autism clinic in America. In the past she received £100,000 as an adviser to the legal team which failed in seeking compensation for parents who believed that MMR caused their child's autism. Although she says her objectivity was not affected by the sum, it seems incredible that The Observer failed to mention the connections.

In particular, as Baron-Cohen says: 'The key question is why The Observer were in such a rush to plaster a figure like 1 in 58 that had not been finalised or agreed by any of the scientists involved, all over the front page? Was it anything to do with the fact that they had the exclusive interview with Andrew Wakefield, on the eve of his appearance before the GMC?

'If The Observer had bothered to ask us about the status of the "1 in 58" result, they would have discovered that this was just one of several statistics in a draft from 2005. The data has gone and is still going through sophisticated statistical analysis by the country's leading epidemiological statisticians… it's unlikely to end up being the final figure or one that any of the team would stand by.'

I contacted Dr Stott. She asked, 'Why, three years after the completion of data collection, more definitive estimates are not available in the public domain? I have raised this concern frequently with my ex-colleagues and have been assured on several occasions that a final write-up is imminent. I am concerned that had preliminary estimates of prevalence been less startling, the process of verification may not have been undertaken and would almost certainly not have taken so long… this introduces a potential source of reporting bias whereby more extreme figures are subject to greater levels of scrutiny than less extreme figures.'

There are clearly issues here for psychologists to consider. But all in all, as Goldacre noted, 'the real villains of the MMR scandal are the media'. In filling his word quota, Denis Campbell had choices, which would have real consequences for the anxiety of millions of parents, and potential subsequent health of many children. He chose not to include a particularly crucial and well-known study from Japan, which showed that when the triple jab was withdrawn rates of autism continued to rise (see tinyurl.com/yrrhb7). But space was found to mention a new book 'which will fuel the controversy': by a London GP who provides single vaccines privately to babies of parents concerned about MMR.
    Jon Sutton