"... there must be an outbreak of serious illness; immunity levels must be low; theThat seems a reasonable bar for a compulsory medical procedure. But does flu vaccination reach this bar?
vaccine must be effective, safe and available; and vaccine uptake must be low."
First, how serious is flu? If we put aside the "estimated" death rate and simply look at death certificates, there are less than 2,000 flu deaths/year in the US. Some flu deaths are due to secondary conditions that occur as a result of flu, so flu may not be listed on the death certificate. Sometimes mathematical models are used instead to infer flu's impact.
Flegel says there are 4-8,000 flu deaths/year in Canada, a country of 34 million people. His own journal, the CMAJ, said in 2003 that Canadian flu deaths had been estimated at 500-1500/year, but applying a new CDC model increased the estimate to 700-2500/year.
Flegel's numbers come from a 2007 paper by Schanzer et al. Based only on mathematical modeling and new assumptions, the authors concluded that the real rate of flu deaths was 4,000-8,000/year. However, this model is fatally flawed: it predicted that over half the deaths during periods of influenza activity were due to influenza, when Simonsen et al. have reliably shown that less than 10% of winter deaths in any season were influenza-related.
In fact, Schanzer's Figure 2 attributes as many deaths to flu as to cancer in Canada! Think of how many people you knew who died of cancer, then think of how many you knew who died of flu. Schanzer needs to go back to the drawing board.
Second, it is hard to gauge the degree of immunity in a population, since we don't know which antibodies are protective, so we don't know which specific ones to measure. Elderly people did not get much swine flu in 2009 because they had pre-existing immunity--but there was no way to know that till after the fact. And there is no way to know how immune our population is to any one flu strain, because immunity from prior years' strains provides some degree of immunity to other strains. But we only learn how much retrospectively.
Third, how safe and effective are flu vaccines? With flu vaccines being made from new strain combinations almost every year, one never knows either how effective or how safe this year's version is going to be. Flegel says efficacy is 86%, but others suggest efficacy is 60% or less, and related to the degree of match between vaccine strains and those circulating.
There was no mention of the interesting Canadian finding, supported by about 6 studies, showing that people who received the 2008 flu vaccine were twice as likely to become clinically ill from the 2009 swine flu virus as those who had not received flu vaccine in 2008. See this article, this one and this letter. The finding raises a new safety issue for flu vaccines, one whose magnitude is uncertain.
Flegel apparently forgot to check his citations. They failed to support several of his assertions, as pointed out in Letters to the editor by several well known flu experts, including an author of the Cochrane influenza vaccine review he cited.
IMHO, the editorial and letters are a welcome science-based addition to the discussion of mandatory, yearly flu inoculations for healthcare workers and others.