1speed
Incredibly profound yet fantastically flawed
manipulating results with double talk..
http://www.reuters.com/article/2015/10/06/us-health-pneumonia-flu-vaccine-idUSKCN0S028O20151006
ok, more people with pneumonia as a result of flu had skipped the flu shot...by the transitive property, the flu shot prevents pneumonia....
Yeah, I'm actually in favor of a flu vaccine (because ... drugs!) but that article is terrible. I imagine the JAMA article on which it is based is a lot more specific in its conclusions. I mean, if its study design is as simple as this article makes it sound, there'd be no way it would have made it through a review. The don't define what the "comparison" group is exactly, and that's kinda important, right? They'd have to match their test group on things like age, comorbidities, medical history, etc. Any of those factors could be just as strongly related to incidence of pneumonia as "not getting a vaccine." And worse, the article implies that their basic conclusion is tied to a frequency value. Their rate of non-vaccination in the study group was about 83%, and that's enough to make them say that a vaccine prevents pneumonia. But what if 150 of the 162 were between the ages of 65 and 80? That's 92%. Should they have then concluded that being young prevents flu? I get that the article has to dumb down the message a bit to reach a broad audience, but they could do a better job than this. Just cite the author's own conclusion and don't try to talk about the data -- just reference that the study was complete and passed a vigorous JAMA review. Adding the "highlights" the way they did just makes it look like bad science.