Sim Epidemic

Dotnaught writes "Scientific American has an intriguing story about EpiSims, an outbreak simulator. Designed by Los Alamos National Laboratory (LANL), it deals with a social networking of a different sort: 'To understand what a social network really is and how it can be used for epidemiology, imagine the daily activities and contacts of a single hypothetical adult, Ann. She has short brushes with family members during breakfast and then with other commuters or carpoolers on her way to work. Depending on her job, she might meet dozens of people at work, with each encounter having a different duration, proximity and purpose.'"

Comments

Problems with models

[CtrlPhreak]

A friend of mine works with the WHO which has solicited many different people to do work like this for them. In light of all the terrorism talk and threats of bio-terrorism, we've had talks on this. Different universities etc. The problem with it is that no model is able to conform to historical records of various outbreaks well enough across the board to develop policy on. One model is highly based on an aids breakout of the 1980s or an asian flu epidemic, and the model fits well to it. But when the model is applied to different epidemics they don't work out. There are just so many factors differing by area, culture etc. Think close knit community vs big city. Also the way things are transmitted. You would have to have a different model for each scenario which is very difficult/costly. Until we can predict everything going into a situation these models aren't very useful.

Models working with more people is definitely a step forward, but just an improvement.

Re: Problems with models

[Alwin Henseler]

You would have to have a different model for each scenario which is very difficult/costly. Until we can predict everything going into a situation these models aren't very useful.

I'd argue against the 'not very useful'. These models may not provide accurate predictions for real outbreaks, but may improve the insights into the variables involved, and how things interact.

In the case of a real outbreak, authorities may take measures like release public warnings, quarantine certain areas/building, ask people to avoid certain activities for a while, etc.

Will that slow down the outbreak? Will it confine casualties to a certain area? Can a measure help to lower mortality rates?

With a real outbreak, that sort of data is crucial. If you know how to slow down the spread of a disease, you buy time for healthcare workers. Time to track down infected people, time to treat people before incubation period expires, etc. Keeping the number of casualties down, may avoid a mass panic. So having some knowledge about the variables involved can make a big difference, and save countless lives.

So these models may be pretty helpless for predicting exact numbers, but still useful anyway.

But, this is SLASHDOT...

[mcrbids]

To understand what a social network really is and how it can be used for epidemiology, imagine the daily activities and contacts of a single hypothetical adult, Ann....

Of course, this is slashdot. If Ann was a slashdotter, her epidemiology would consist only of contact between Ann and her parents, at the dinner table, during the approximately 45 minutes per day that Ann leaves the cellar.

Perhaps it should read something like:

To understand what a social network really is and how it can be used for epidemiology, one must not be a slashdotter. Imagine the daily activities and contacts of a single hypothetical adult, Ann....