A groundbreaking study published in the Proceedings of the National Academy of Science demonstrated that cell phone data can be used to predict seasonal disease patterns.
Researchers from Princeton University and Harvard University tracked the spread of Rubella in Kenya by using anonymous mobile phone records of more than 15 million people. This type of work was difficult to do in the past, because of a lack of technology usage by individuals in low-income and underdeveloped countries. However, as the use of cellphones becomes more prevalent, studies such as this become more feasible.
The researchers had access to records of mobile phone usage and movement in Kenya between June 2008 and June 2009 (with the exception of February 2009 which was missing from the dataset). They were able to link more than 12 billion phone communications to a province. This enabled the researchers to determine a daily location for each user and the number of trips the person took between provinces each day. They were able to then compare the data to a highly detailed dataset on rubella incidence in the country.
The data matched, and the patterns of cell phone movement lined up with the figures on the incidence of rubella. The disease spiked three times a year—primarily in September and February along with May in a few locations. Thus, the cell phone data predicted the spread of this infectious disease.
The researchers expected this pattern, because rubella is more likely to spread when children interact with each other after holiday breaks and at the start of school. In most of Kenya, the risk decreased during the rest of the school-term months. The exception was in Western Kenya where the risk during breaks in school was higher than when school was in session.
The analysis showed that mobile phone data could be used to detect seasonal human movement patterns and help to understand childhood disease. The researchers hope to apply their methodology to other diseases shaped by human movement like measles, malaria, and cholera.