Identifying Risk Factors of Medical Conditions through the Analysis of Internet Search Engine Queries
presentation by Dr. Elad Yom-Tov, a Senior Researcher at Microsoft Research in Israel.
Epidemiological studies for discovery of disease risk factors are usually performed by analyzing the physical activities and demographic data of different populations. As people's behavior online has become reflective of their activities in the physical world as well as the virtual one, we propose to use such behavior to discover risk factors of medical conditions, drawing on anonymized queries submitted to a general-purpose Internet search engine.
There are several challenges in this process: First, a cohort of people who are likely to be suffering from the condition of interest needs to be discovered based solely on their queries. Second, queries need to be generalized to describe facets of user behavior. In my talk I will describe our solution to these challenges and how the Self-Controlled Case Series method can be applied to the processed data in order to identify new risk factors. I will show some of the risk factors discovered by application of the proposed method, and discuss the advantages and drawbacks of our analysis.
