I looked at a recent news story about a paper on asthma and paracetamol. According to the Daily Mail Taking paracetamol regularly ‘triples risk of asthma’ and the headline on the Daily Telegraph site is “Paracetamol use found to increase risk of asthma three-fold”.
Given the statement in the first headline (I know it’s in single quotation marks, but really – do they honestly think that single quotes will stop someone concluding that it is definitely the case?) and the Daily Telegraph’s assertion that regular users of paracetamol are nearly three times more likely to suffer from asthma, you might be forgiven for assuming that this is a done deal. After a quick glance at the stories, I took a look at the NHS “Behind the Headlines” take on the coverage of this. The Daily Telegraph’s and the Daily Mail’s assertions are of particular interest given that Behind The Headlines says this:
The study behind this report is a case-control study and by virtue of its design, it cannot prove causation (that paracetamol increases asthma). It also cannot rule out reverse causation (that asthma causes people to take more painkillers). However, when considered alongside the findings of other studies, these results suggest that there may be an association between use of paracetamol and asthma that needs further exploration. This is not a new concern, and researchers have been investigating this for some time.
So while the media reports are implying that causation has been shown, it appears that the Behind The Headlines team (BTH) tell us it would be impossible to do so in a study of this type. I don’t know why the media seem to feel unable to use the phrase “x is associated with y” rather than the inaccurate “x causes y”. Also of interest is the last sentence of the quoted section above – this isn’t new. It’s something that has been looked into for some time now and still needs further investigation, but you wouldn’t know it from the media coverage. I tink the Daily Mail were aware it wasn’t new though.
Weirdly, when I looked through google’s news search the Mail’s report stated that the study was published today in the Lancet. The site had the story down as being updated at 8:21 AM on 19th September 2008, so I assumed the Mail story related to a paper published today in the Lancet. The only thing is, I couldn’t find a Lancet paper on Pubmed relating to asthma and paracetamol from as recently as 2008 – I got a single hit from 2000. BTH has the paper down as being from Eur Respir J 2008, doi:10.1183/09031936.00039208. I think that perhaps what happened was that the Mail updated an old story because they had a new story about paracetamol and asthma. The Mail report that showed up on google news is here. The Mail report on the current study is actually here. Confused? I was. Maybe they should have an ‘originally published’ tag on stories they update, if nothing else it would at least save morons like me from ending up in a state of confusion over what is today’s asthma news and what was 2000’s asthma news. Attempting to understand science stories in the press can be a real pain in the arse sometimes.
Also of interest: The news reports stated that “Paracetamol tripled the risk of asthma”. Given that I have no idea what the chances were of someone having asthma if they haven’t taken paracetamol, I have no idea what it might mean for their chances of having asthma to triple if they do take paracetamol. This is something that Ben Goldacre has raised in his book Bad Science [it’s in chapter 13, if you’re reaching for your copy as you read this]. Newspapers tend to give the scariest/most impressive figures (for example, the relative risk instead of the absolute risk – see here for more on relative v absolute risk). In the case of the media reporting of a Mischief PR press release, there was a statement that there had been a 39% increase in household reports of wasps following introduction of fortnightly bin collections. Perhaps the press should have found the time/space to inform us of the “natural frequency” of asthma attacks, which might mean more to me than being told there is a “tripling of the risk”. Bandolier briefly explains natural frequencies here. Their explanation concludes with this: “Natural frequencies help people to make sound conclusions, whereas conditional probabilities tend to cloud minds”.
Links: Click here for the abstract of the paper.