When people get sick they usually sleep more, use their phone less, and stay in bed all day watching DVD box sets of their favorite shows. Such patterns can be detected and that’s exactly what students at Massachusetts Institute of Technology have successfully achieved. Anmol Madan, the guy behind this idea, gave 70 undergraduate students living in a dormitory a mobile phone. Preinstalled was software that tracked their movements, phone call and text message usage. A daily survey on both their physical and metal health was also taken. Data was collected for a period of 10 weeks and it was found that the students who came down with the flu moved around far less than usual and stopped making late night and early morning phone calls. Madan tweaked his “disease finding” algorithm to search for similar patterns, and with 90% accuracy he could tell when someone was infected.
This work has inspired Leon Danon at the University of Warwick in the UK to try and do similar research in an unspecified Northern European country using the anonymous cellphone records of 10,000 people and try to match their health data with their device. He’s waiting for approval. Think of the potential if your mobile phone could not only tell you that you’re getting sick, but that there’s a really bad case of the flu going around your part of town and that you should start taking your vitamins? Similar profiles could be built for other diseases. Privacy advocates are probably going mad at the very thought of this, but it’s always give and take with issues like this. How much information would you be willing to give up if you, as a parent, got a text from your child’s school saying that there’s been a lot of flu going around and so you should make sure your little boy or girl has all her shots?