Tuesday, May 12, 2009

New and Hot: How to Predict Pandemics

May 8, 2009
arxiv.org/abs/0905.0728

Researchers from the British Colombia Centre for Disease Control released a model that enables us to estimate the basic reproduction number R0 far earlier in an outbreak than has been possible before. The basic reproduction number describes the expected number of new infections caused by a typical individual during a certain period of time. It is an important factor for estimating the final size of an outbreak and can therefore by very useful for global health officials trying to figure out how to deal with an outbreak in its early stages—for example, swine flu.

The model requires three quantities as input; these quantities are apparently relatively easy to measure early on in an epidemic. These quantities are:

1)infection rate probability- for viral diseases, this is related to the viral load of infectious individuals and can be calculated from their blood serum.
2)removal rate probability: how quickly individuals are removed from the scenario through factors such as death, quarantine and other means to reduce social contact.
3)generation interval distribution - rate at which new cases appear.

The typical disease spreads in three phases: in the early stages, the pattern of disease spread is stochastic or random and therefore hard to predict. Later, after more people have the disease, the other phases are more straightforward: exponential growth and then decline of the number of infected individuals. Knowing the basic reproduction number in conjunction with our knowledge of social networks (how often people meet and where) can help predict the spread of disease even while it's in its early, stochastic phase.

-Elaine C.

1 comment:

  1. How does one go about determining the accuracy of these models? Has this new model held up empirically?

    ReplyDelete