June 30, 2003

Typhoid Annie

Here is an idle question for the Peanut Gallery and assorted cleverologists out there - is there much written about the application of game theory to epidemiology? (that is, that a non professional-social scientist might digest? and preferably articles available online).

I ask because of an Economist article from last week (no point linking, it’s password protected) which says that despite the fantastic scientific advances that helped identify and explain much about SARS within weeks of its being discovered, it seems that old-fashioned public health measures such as quarantine were what really contained the disease. Now, I’m not going to get started on my theory about SARS being a Sky-news disease of the week sort of crisis. But this did remind me of Danny Boyle’s recent film 28 Days Later which is about the UK population being infected and practically wiped out by a nasty, zombie-creating disease. It’s pretty gory stuff, but very interesting in its own way.

Without spoiling it, I can only say the idea of quarantine is pretty essential to the plot (which has lots of holes, but anyway). One of the characters graphically describes a scene at Paddington Station where thousands of people are trying to escape plague-ridden London by getting to Heathrow and abroad. Of course, other countries won’t want possibly infected people to enter, so in the film, the panicked hoards are prevented from escaping and subsequently become infected and die.

In the presence of a particularly virulent and infectious disease, individuals will desperately want to escape from an area of infection (either to escape infection or to receive better treatment). For the larger group (or simply other individuals in a lower risk area), containing potentially infected individuals is essential, even if that increases those individuals’ likelihood of being exposed to a disease. Is this simply a question of co-ordination, or do the raised stakes make quarantine in this type of scenario a zero sum game? Or is it a game at all, given the existence of the Leviathan state stopping people from becoming far flung vectors?

Posted by Maria at June 30, 2003 10:49 AM | TrackBack

I’m not sure it’s a game of the sort you have in mind, but I’m certainly no game theorist. I’m not even a cleverologist, though I wish I were: it’s a great job title. Though the outbreak problem may not be a game, it’s certainly amenable to formal modeling. I think Malcolm Gladwell has a nice sketch of this area (probably with some useful footnotes) in “The Tipping Point.”

Posted by: Kieran Healy at June 30, 2003 11:52 PM

Multi-person prisoner’s dilemma with negative payoffs. Everyone is better off with quarantine than without, but best of all off if there is general quarantine but they themselves are allowed to run away.

Posted by: dsquared at July 1, 2003 02:36 AM

I’m not an epidemiologist (but I have played one on the internet). I do know a little about the models they use, however.

Typically, a compartmental model is employed. In the usual analysis, the population is divided into 4 compartments comprising susceptible, exposed, infectious and recovered individuals. From the empirical data, it’s then possible to guestimate rate constants between compartments, and derive differential equations which model disease growth.

As far as I know, game theory isn’t used, but a recent Wired article did mention a newer approach using network science, which allows one to take differences in individual susceptibility and behaviour into account, and derive spatial models. Unfortunately, the article is a little short on detail.

Posted by: antirealist at July 1, 2003 08:07 AM

Very interesting post

Posted by: David at October 21, 2003 06:54 AM
Post a comment

Remember personal info?