While reading a blog post written by a chiropractor bloke who is definitely not a chiropractor*, I noticed a claim that had been made regarding the vaccination status of children with whooping cough.
This claim was that vaccination could not be guaranteed to be 100% effective. This is true. However, the claim that “85.9% of children with whooping cough (pertussis) were fully immunised” was then made without giving readers any clue as to how this should be interpreted. What does it mean to say that the majority of children with whooping cough had been vaccinated against the disease? Well, it’s important to bear in mind that if the majority of children are vaccinated against whooping cough then it could easily be the case that the majority of those diagnosed have been vaccinated. This would not in itself say anything particularly meaningful about whooping cough or the pertussis vaccination. Are children who have been immunised more or less likely to be diagnosed with pertussis than those who have not? Does either group suffer from more or less severe complications? These are the sort of questions I’d be interested in finding out the answers to. I would not be particularly interested in learning simply how many of the children with whooping cough have been vaccinated, as it would not really mean anything to me on its own. This number alone would not inform me.
Let’s get hypothetical and pluck some numbers out of the air. If you have 1,000,000 children and 95% are vaccinated, then 950,000 children have some protection against pertussis, while 50,000 are left unprotected. Say 100 children catch pertussis. Of these children with pertussis, 86% are vaccinated and only 14% are unvaccinated. These figures might, at first glance, make it seem as if vaccination is pretty much useless – just look at how many vaccinated children got sick. However… of the 50,000 children who are unvaccinated, 0.028% have got whooping cough. Of the 950,000 children who were vaccinated 0.009 have got whooping cough. Using these hypothetical figures, we can see that it is easily possible for the rate of infection with whooping cough in the unvaccinated to be three times that in the vaccinated despite the potentially misleading figure that shows 86% of children with whooping cough had been vaccinated and may make us doubt the effectiveness of the vaccine.
In the comments section on the chirolive blog, someone asks the following question of the author:
Could you explain how the 85.9% figure relates to your argument and why you feel the percentage of children with whooping cough who are vaccinated is the best way to represent the data.
I think this is an excellent comment and the author doesn’t yet seem to have given a direct answer to the twin question that has been posed here. When people use numbers without giving any kind of context or explaining how the number supports their argument then you have to ask yourself how meaningful that statistic is. After all, I could claim that the majority of children who test negative for whooping cough have been vaccinated against pertussis – it’s a whopping 97.2% in the BMJ study** that is linked to on the chirolive blog – but on its own, taken out of context, would that figure really be meaningful? Would it tell us anything useful?
*I have amended this post to make clear that the author of the Chiropractic Live blog is not actually a chiropractor. He does, though, still practise chiropractic. According to his blog: “I still practise chiropractic and am a passionate advocate of chiropractic.” (June 11th 2009). So he practises chiropractic, but is not a chiropractor. I’m pleased to have been able to clarify that for you.
**The PDF is available via this link. The abstract is here. While I did make clear that the BMJ study had been linked to on the chirolive blog, I failed to provide a direct link myself. I apologise for this uncharacteristic omission. Backup copy: whooping cough.
These edits were prompted by criticism from the chiropractor bloke who is definitely not a chiropractor, but who practises chiropractic in this blog post. I don’t feel that he has adequately addressed my criticism of his use of a single statistic, but you can judge this for yourselves should you wish to do so. If you do read his blog post for yourself it will at least mean that you do not have to rely on my biased interpretation of his writings. After all, I may have “a bee in my bonnet”.
Forthcoming blog posts
I’m writing something on the weapons of influence that may be useful to advocates of alternative medicine (I may be being a trifle ambitious in attempting to marry Cialdini’s work on influence and persuasion with Holford, McKeith and homeopaths, but we’ll soon see).
I’m also part way through a rant about psychics/mediums giving advice to people on how to live their lives.