November 17, 2005

Fears of a bird flu epidemic are overblown

I do not entirely understand the proliferation of ominous pandemic-bird-flu articles. They greatly exaggerate the magnitude of the danger as well as the imminence of catastrophe. Are there not more mischievous things a New York Times reporter can accomplish with his column-inches?

The “flu-is-coming” story is quickly becoming a journalistic standby. If you recall, it was ubiquitous at this time last year. I was skeptical (and right).

Last year, on October 31, I wrote: “I doubt this year’s vaccine shortage will lead to the highly fatal pandemic the fear-mongering media have been insinuating. There have been three high-magnitude influenza outbreaks this century: the 1918-19 “Spanish” flu, the 1957-58 “Asian” flu, and the 1968-69 “Hong Kong” flu. If you assume the probability is time-independent, the country has a pretty good chance of getting through this year more-or-less unscathed.”

I will recycle that prediction for 2005. All reasonable people—here I am excepting New York Times reporters, lottery-ticket buyers, and Scientologists—agree that influenza outbreaks are independent. It is like flipping a coin. If you flip a coin once, there is a 50 percent probability it will land heads. If you flip a coin 10 times and see 10 heads, on the 11th flip you face the same 50-50 odds.

Pandemics are not “cyclical.” There is no 38-year rule. If 37 years pass pandemic-free, the influenza virus will not vow to try extra-special hard to cause a pandemic.

In 1997, this precise virus—bird flu (H5N1)—had the media in a frenzy. The Center for Disease Control issued stern warnings. Authorities killed 1.5 million chickens.

The human casualties? Six.

In 1976, United States authorities panicked over a “swine-type” influenza An “outbreak” in Fort Dix, New Jersey. The authorities proceeded to attempt to vaccinate the entire country.

The “outbreak?” A single death.

However, the vaccine killed 25 people.

Unapologetic mis-headlining quickly inflames the hysteria. Reuters wrote, “Flu Pandemic Could Kill 150 million, U.N. Warns.” They were referring to a comment by a single official, who guessed a pandemic could kill from “5 million to 150 million.”

But Michael Osterholm, professor at the University of Minnesota School of Public Health, can top that. He estimated between 180 and 360 million casualties, a range he arrived at, as far as I can tell, by multiplying the number of deaths from the Spanish flu by the ratio of today’s population to 1919’s.

Why not multiply the bird flu’s fatality rate, roughly half, by today’s population of six billion? Three billion casualties? A ridiculous extrapolation, but no stupider than Osterholm’s.

The first expert volunteered an essentially useless sentence quite likely to be true, whereas the second volunteered a more specific but less defensible guess.

The tradeoff, when you don’t know anything, is between saying something useful and saying something true.

The Spanish flu casualty rate, the only non-imaginary part of Osterholm’s model, is a wildly inappropriate proxy for a hypothetical future influenza casualty rate. In 1919, pharmacies lacked antibacterial treatments which can be obtained cheaply and easily today. In the case of the Spanish flu—like most flus—the substantial majority of casualties were not caused by the virus directly but by secondary bacterial infections. Today, those are quite treatable.

It is possible that a pandemic will begin tomorrow. But it is not likely. The danger is exaggerated.