decision trees and swine flu
Monday, August 24th, 2009I have a masters in public policy degree, and people sometimes ask me what exactly do you learn in a public policy program. The answer is a little bit of lots of things — some economics, some statistics, some political science, how to write a 2 page memo, a fair amount on whatever topic you specialize in. But probably the core skill they try to teach is about decision making with imperfect information. And when I was in grad school, the main example they used to illustrate this issue is the 1976 swine flu vaccination decision.
The short version of it is that in February 1976 two soldiers at Fort Dix died from a swine flu, with a significant number of soldiers infected, and no exposure to pigs involved. This was known to be related to the flu that had caused the 1918 pandemic, most people didn't have immunity to it, there was a theory of cyclical pandemics that suggested that we were overdue for one, and this strain had proven itself to have the ability to kill otherwise healthy young men. The government asked the vaccine manufacturers to produce a vaccine, and in the fall, they swung into action with a full-fledged immunization drive.
As it turns out, swine flu did not reemerge in the fall, and the vaccination campaign was suspended (after about 40 million people had been vaccinated) when some of the vaccinated individuals came down with Guillain-Barré syndrome.
With hindsight, it's clear that the vaccination drive was unnecessary. But with the information available, was it the wrong thing to do? The people who made the decision still think it was the right call, writing: "When lives are at stake, it is better to err on the side of overreaction
than underreaction. Because of the unpredictability of influenza, responsible
public health leaders must be willing to take risks on behalf of the public." My public policy classmates were less convinced, generally concluding that it made sense to produce the vaccine, but that they should have held off on immunizing people until there was some evidence that the swine flu had returned in the fall. We felt the president had been pushed by politics to overreact to a low-probability but high risk event, fearing the headlines if tens of thousands died and the government had done nothing.
Which brings us to today, and the headlines that half of Americans could be infected by the swine flu this year. So, I'm glad that they're making vaccines. But I'm not sure whether I want to line up — or line my kids up — to be immunized. If you've got some time to spend thinking about what the vaccine strategy should be, the CDC is holding two web-based public forums to collect input and is looking for participants.
For us, a complicating factor is that D has a history of mild asthma. And viral infections are his primary trigger. He's been doing so well that we've had him off of the inhaled steroid, but I emailed his pediatrician to ask if she thought we should start him again with the start of the school year. She responded that the risks of the inhaled steroid were very low, while the flu was an unknown, so she'd advise going ahead with the steroid. So we're going to do that, and otherwise we're waiting and seeing.
What are you doing? Will you get vaccinated if offered it? Do you think your local school system is overreacting, underreacting, or getting it right? Do you have a plan for what you'll do if schools are closed?