Skip to main content

Tracking an epidemic requires computer models—but what if those models are wrong?


Laura Castanon’s article on epidemic tracking speaks on the use of computer models and how those can become inaccurate, due to the nature of social networks. More specifically, researchers are saying that “the number is never constant”, referring to the average number of new cases caused by each infected person. Castanon explains that the basic simplified models show an epidemic’s beginning as an exponential curve reliant on this “average” number, with the curve leveling off when people’s networks become saturated with sickness. However, new models are being created with the idea that “not all networks are created equal”, and that in reality, we cannot rely on just this average reproductive number.

The primary takeaway that the article presents is that networks can vary a significant amount based upon age, location, culture, and several other factors. With this knowledge comes tedious work in accurately assessing all the factors in epidemic tracking. Researchers suggest that instead of using some approximate reproductive number to predict epidemic spread, using quantitative data from the past to attempt to anticipate the spread. A simpler model relies on assumptions that could highly skew predictions.

It is interesting to think about how network contagion can follow the principles we know to be true, and yet be so affected by external factors that can be extremely difficult to accurately measure. Assessing whether or not people fall into high traffic locations often or the number of connections they may have drastically changes our forecasts, and as such we need to evaluate the situations as the complex issues they are.


Leave a Reply

Blogging Calendar

November 2018