Do you Google your health symptoms instead of visiting your doctor? According to a recent article in The Atlantic, roughly 72% of Americans searched the internet for health information during the past year and one-third self-diagnosed a health problem using information obtained online.
According to the article, public health officials use this online activity to monitor and map diseases. A recent study published in PLOS Computational Biology found that trends in Wikipedia page visits can be used to predict flu outbreaks up to one month in advance.
Researchers from Los Alamos National Laboratory created a predictive computer model by comparing internet traffic to 10 flu-related Wikipedia pages to Centers for Disease Control and Prevention (CDC) flu reports. The researchers also applied the predictive model to dengue, tuberculosis, and HIV, but achieved varying levels of success. For example, the model predicted outbreaks up to one month in advance for dengue and one week for tuberculosis. It was unable to predict HIV outbreaks, most likely due to the virus’s (1) lack of “seasonality” (e.g., following seasonal patterns) and (2) ability to live inside a person for an extended period of time without causing symptoms.
While the article notes that using internet activity to predict disease outbreaks is not new, it suggests that using Wikipedia may offer scientists a way of overcoming existing obstacles in disease forecasting. For example, Twitter data is costly to obtain and some companies, such as Google, keep such data private, due to fears that computer hackers may manipulate it to create the appearance of a disease outbreak. Conversely, Wikipedia provides public access to hourly traffic data for its pages.
Additionally, using Wikipedia allows researchers to bypass the bureaucracy of large-scale disease tracking, as the CDC’s flu-surveillance report has a two-week lag time.
