Using modelling to estimate the number of flu-related deaths makes a lot of sense when many of those who die are in nursing homes, not hospitalized, and not tested for flu. "Many are close to death, with flu being only one factor leading to their demise," notes today's Washington Post. I would expect that several times the number of "proven" flu deaths from seasonal flu occur each year.
Flu deaths in children, particularly this year, are another matter entirely. CDC has asked medical providers to perform influenza tests on all patients hospitalized for presumed flu, and in all those who die of possible flu. Therefore, mathematical models are unnecessary for determining the number of flu deaths in hospitalized patients, since they are all being counted. This should include all children with serious flu-related illness. Thus CDC's surveillance of hospital deaths should already include every pediatric mortality case.
According to the Washington Post's David Brown, "The new estimate includes deaths that occurred outside hospitals, patients who tested negative for H1N1 but almost certainly had it, and other overlooked cases." Brown goes on to quote CDC's Dr. Schuchat regarding these new numbers: "We don't think anything has changed," Schuchat said. "We think our 540 number (for US child deaths) is a better estimate for the big picture."
Tested negative but almost certainly had it??? Overlooked cases??? By whose criteria? And children who die outside a hospital will almost certainly be autopsied, if the diagnosis is in question.
The front page of today's Bangor Daily News notes Maine's third swine flu death. The article, however, points out that all three people who died had "very serious underlying medical conditions." These are precisely the patients who die from flu each year. Though sad, such deaths aren't really front page news any other year.
Multiplying the number of known pediatric deaths, using a mathematical formula appropriate for deaths in nursing home patients, reflects a lack of scientific integrity and instead suggests fearmongering.
Today's new mortality estimates are consistent with CDC's known PR tactic of "predict[ing] dire outcomes" in order to increase vaccine uptake. See: Doshi, Peter. “Viral Marketing: The Selling of the Flu Vaccine.” Harpers Magazine. March. 2006.