His original post kicks off the debate:
"I keep getting asked about the Atlantic Magazine article, Does the Vaccine Matter? by Shannon Brownlee and Jeanne Lenzer, two reporters whose particular bias is that we as a nation are "over treated." As a generalization that's probably true, and finding examples isn't hard. Unfortunately by taking as their main example flu vaccine during a pandemic, they have not only picked the wrong example but created more confusion at a time when there's already too much.
There is nothing judicious about Jefferson, whose problem was described by one of my colleagues as "methodolatry," the profane worship of the randomized clinical trial as the only valid method of investigation. In this case his evidence base isn't even relevant, because we aren't dealing with seasonal flu but pandemic flu. For the record, what he is saying about the uncertainties about flu vaccine efficacy in the over 65 age group isn't new. In fact we've discussed it here, several times (here, here), going back a couple of years. But it is also clear that the vaccine offers protection in the age groups that matter for this pandemic, the people under age 50.
Nor is Jefferson, as claimed in the article, someone who "knows the flu-vaccine literature better than anyone else on the planet." That's an absurd claim. The literature is vast and he knows only a tiny part of it. But insofar as there is an acknowledged expert on vaccine efficacy, it would be biostatistician Elizabeth Halloran, who reviewed the clinical trial and experimental challenge literature recently in the American Journal of Epidemiology. We wrote a longish post on the subject here. There is general agreement, even among so-called skeptics, that the vaccine works in the under 60 age group, precisely the group at issue with the swine flu vaccine.
Jefferson's points about vaccines in the elderly are not new, as I've said. But even though he is extreme in his views, looking at Jefferson's own recent papers in the literature show conclusions much weaker than the picture painted in this article..."
It continues from there.
Then, the authors come back with a long post
"Blogger Revere describes our article, published in The Atlantic, “Does the Vaccine Matter?” as an instance of “methodolatry.” Revere and his followers, are so enamored with their new-found term that they fail to examine, using facts rather than insinuation and emotion, whether this article was worthy of that term – a term that is useful and could be used to describe a very real problem.
A layperson wrote the following response to Dr Gorski’s praise for Revere, which we think is relevant here. She wrote:
Can someone…address the points of the article rather than just saying
for flu vaccine is unethical? I understand that point very well.
But, can someone discuss the challenges the authors made to the 50% mortality reduction statistic? How about the 2004 low production year where mortality didn’t fall? And, I’m most interested in their assertion that the ethical moratorium on for aggressive breast cancer treatment increased mortality risk. What about situations like that?
I’m not an anti-vaccine person (I got my flu shot last week), I’m just a curious lay person.
Does it take a layperson to get scientists to act like scientists?"
Revere Responds:
Brownless/Lenzer: It is a sprited defense of a straw man. You interjected the issue of efficacy of flu vaccine among the elderly, where the RCT data are sparse, into the question of pandemic preparedness, where the vaccine is primarily aimed at the young age group, for which RCTs for influenza vaccine do exist (see the link to one of our posts discussing the Halloran review). I didn't address the points you raised for two reasons. I have addressed them in posts here (links provided in the post) at least twice in the last several years; but more importantly, they are not relevant to the issue at hand, use of vaccines for the swine flu pandemic. I have also posted here, several times, on the method used for estimating excess mortality in flu (here's the link). Both P&I and All Cause is used, and is a choice. Both Simonsen and Thompson have done so, as I recall, and it is Simonsen's work that called attention to the lack of mortality change despite increase in vaccination among the elderly over an extended time period. Note that such a comparison is much more fraught than comparing the observational and RCT studies, because it is an ecological design comparison: you are trying to infer the interior of the tables (the joint probabilities) from only a knowledge of the marginal probability estimates. We know that both confounding of all kinds (especially temporal and geographic) and effect modification can drastically affect that comparison (even reversing its direction) and clearly over this time period both are operating.
http://scienceblogs.com/effectmeasure/2009/10/journalists_sink_in_the_atlant.php
Printer-friendly format
Email this thread to a friend
Bookmark this thread
(1000+ posts)
Send PM |
Profile |
Ignore
