4 Revolutionary Ways to Predict Pandemics

Popularity is Everything

Sociologist Nicholas Christakis has developed a system designed to better predict the spread of infectious diseases based upon a system of ‘social networks’ that we are all a part of. Here social networks are referring to actual networks in reality and not FaceBook or Twitter!

Traditional methods for predicting pandemics use random samples to predict the prevalence of an infectious disease. This method takes time and often will only provide analysis after the fact. A more modern addition to the tools of prediction is Google Flu trends which allows for real time stats about the spread of a pandemic.

What Nicholas has developed is a method by which we can see days, if not weeks in advance, the spread of a pandemic.

Christakis’ system works by identifying individuals at the centre of large social networks. We all have friends, but some have more friends than others. By identifying and following individuals central to social networks through the principle that ‘the friends of randomly selected people have more friends than the randomly selected people themselves’* Christakis was constantly able to predict flu trends amongst groups of university students well ahead of standard methods of prediction.

The system works on the basis that the individuals at the centre will have more contact with other individuals in that network. By asking the university students to nominate a friend, and using the nominated friend as a marker for a pandemic, Nicholas was able to get much faster results.

The usefulness of this theory extends beyond mere prediction. If a health authority can only vaccinate a small amount of its population against an incoming epidemic, the authority should ask a randomly selected group to nominate a friend, and then immunise the friend.

*Consider that, at a party, the guests are more likely to know the host than each other. If you then ask guests to nominate a friend at that party they are likely to nominate the host.

< Previous Page | Next Page >

Leave a Reply

Your email address will not be published. Required fields are marked *