Autism, MMR and bad reporting, redux

Speaking of Bad Science and bad reporting and how the two seem to go together so frequently, Ben Goldacre goes after The Observer big time in yesterday's Bad Science column. The Observer, of course, continued to do its bit on behalf of scientific illiteracy with scare story on 9 July over yet another purported link between the MMR vaccine and the apparent rise in autism in the UK. What got me was the headline: New health fears over big surge in autism. And, as I recall, this was the lead story in The Observer that day.

Where to begin? Well, maybe the place to begin is to go on Guardian Unlimited and type in "MMR" in the search box. Astonishing how much the Observer story got wrong, isn't it? As Mrs Wufnik innocently asks, why do we still buy this newspaper? Plus we now know some new things we didn't know before. Like the fact that the "study" forming the basis of the Observer story isn't complete yet. And that the "experts" raising new concerns aren't really "experts," and one of them hasn't even raised concerns. And that Andrew Wakefield, the MD who started the controversey and remains unreprentant about it, paid children at a birthday party £5 each for blood samples. The list goes on and on. Goldacre's piece should be required reading for every journalist in the UK who even thinks of writing on this topic. Even the Reader's Editor at The Guardian grudgingly admitted that the story conflated two issues--the alleged rise in autism and the MMR controversey--that should not have been, although he remains strangely defensive on other problems with the story, and one wonders how the head of news referred to in this column could be such a dunderhead and still remain an editor.

Two other things bother me about the story, though, and, in fact, the way the entire autism "debate" seems to be conducted in the UK press. The first is the very strange lack of any discussion of a possible genetic issue here. It's just not discussed seriously in the UK press. Now, it's not that frequent a talking point in US coverage either, but it does show up from time to time. My own convictions here changed radically following reading an article in Wired several years ago that laid out a pretty compelling reason for at least considering the proposition seriously--the emergence of "autism clusters" in the US, kind of like "cancer clusters". Called The Geek Syndrome, it's worth digging out at the library, since the version available on-line is considerably shorter than the original article, and barely mentions the geographic aspect at all. The original article simply makes an interesting observation--there appears to be a higher (and growing) incidence of autism and Asperger's Syndrome in geographic regions where there is concentration of people involved in work relating to mathematics, computers and the sciences than in regions where these people don't aggregate. What to make of that? Assuming that autism doesn't come from some sort of communicable virus, this at least suggests some genetic, even hereditary, component.

The other point relates to defining and categorizing someone as autistic. There are real issues here, and categorizations aren't always clear cut. I used to work with men who had aphasia, who were generally older, who generally had aphasia as a result of strokes where the underlying lesion was clearly identifiable in terms of brain locality, and where the clinical categories were reasonably well defined. And still, diagnosis and categorization could often prove problematic. In the case of autism, clinicians and diagnosticians are usually dealing with children of different ages and levels of development, where there is no obvious organic cause, and where there may be a great deal of other stuff going on. Now, we haven't followed this literature very much, but the issue of whether definitional criteria have expanded does seem a reasonable one to ask. We posted on this a while back in relation to the "surge" of children being diagnosed with "behavioural disorders" in the US. And as we noted at the time, there have been a number of suggestions that one possible reason for this may be the increasingly broad metric being used to call something a "behavioural disorder."

Now, I don't necessarily expect reporters to be up on this stuff. But I would expect them to at least know that these are issues worth considering before they attempt to link issues that are only being inked, and irresponsibly at that, in an attempt to generate a good headline. And if this is too much trouble for reporters or editors to go to, at the very least someone at The Guardian or The Observer should have enough sense to run it past Ben Goldacre first. They do work for the same organization, after all. And at least they won't look stupid. Or worse.