Isn\’t this a known effect?

Researchers in Canada have found that for every million doses of the H1N1 pandemic vaccine that were administered in 2009/10 there were two extra cases of Guillain-Barré syndrome which can be fatal in extreme cases.

Somewhere lurking at the back of my mind is the idea that this is a known possible side effect of many vaccines?

8 thoughts on “Isn\’t this a known effect?”

  1. It’s a known hypothesis that Guillain-Barré syndrome may be associated with immunization, but there’s little evidence to support the hypothesis except in the case of swine flu vaccination.

    What this paper did is look at the entire immunized population of Quebec and find a slight blip in GBS onset in the weeks after immunization. It’s possible this is because the vaccine is a cause of GBS, it’s possible that it brings forward GBS that would have happened anyway, or it’s possible that it just makes people more likely to consult a doctor if there’s something a bit wrong with them.

  2. All estimations of such effects are based on assumptions that are hard to check. A result of 2 in 10^6 could easily be explained by the assumptions being a little wrong – as PaulB has explained.

    That’s why modern epidemiology has such a poor success rate – many effects, if they exist at all, are so small that they are hard to pick out. Smoking cigarettes is an exception because the effect is huge.

  3. Sounds like an acceptable trade off given there were >>2 lives not lost to flu per million doses administered.

  4. Actually I’d opine that epidemiology is fine and a valuable discipline, what is not fine is the politically motivated attempts to use it as a tool for actions that can’t be sustained by the findings.

    Basically, even in the best designed studies, the potential confounding effects mean that a result has to be very marked to be worthy of direct attention without very careful attention or further study. John Briggs of Numberwatch fame used to claim that an excess of over 100% was the practical minimum that should normally be considered.

    What you tend to see all too often is the announcement of a small apparent correlation with claims that this is definitive. I recall one such “study” regarding Dioxin contamination in an area, and from the epidemiological data they managed to find a couple of syndromes or diseases with excesses of 20-30% out of a grab bag of possible issue. What they dishonestly didn’t mention is that there were just as many syndromes with a 20-30% under representation. So they wanted to claim for damages for the overs, but not gain a credit for the unders.

    Pity, as this undermines a generally useful tool.

  5. “How on Earth do they factor things out tightly enough to correlate two cases per million?”

    If the normal rate of Guillain-Barré syndrome was 15 per million and they looked at a population of ten million vaccinated people then they would expect to see about 150 cases. If they saw 170 cases that would suggest an additional factor.

    I appreciate that’s still only a 13% increase, but it’s easier to spot than is inferred from the original numbers.

  6. Well, there are known knowns, and unknown knowns…. I won’t continue the quote, but just say that Rumsfeld knew whereof he spoke. And Mr Worstall might pay attention to it next time he writes about “global warming”.

Leave a Reply

Your email address will not be published. Required fields are marked *