On Top of the Flu
News Nov 12, 2015
Harvard statisticians have devised a new method to track the flu via Internet search data, potentially providing public health officials and consumers alike with advance warning this flu season.
The system adopts an approach — tracking searches for key words and phrases such as “flu,” “flu symptoms,” “treating the flu” — that has been tried before with varying levels of success, but combines it with additional data to improve accuracy.
The result, according to Samuel Kou, a professor of statistics, is the most precise method yet.
“If you see a spike in search volume, it probably indicates that something is going on,” Kou said.
The history of tracking the flu using search data goes back several years. The most notable effort, Kou said, was that of Google Flu Trends, released in 2009 and discontinued last summer. Though Google Flu Trends used sophisticated algorithms and was revised repeatedly, its predictions often missed the retrospective analysis of the Centers for Disease Control and Prevention, which tracks actual reports of people with flu-like symptoms seeking medical attention across the country.
CDC data are considered something of a gold standard in disease surveillance, Kou said, but it takes one to three weeks for reports to be compiled, making it hard for public health officials to stay ahead of the flu.
Though most sufferers recover relatively quickly, the disease can be deadly, killing 500,000 annually around the world. In the United States, it kills an estimated 3,000 to 5,000 people each year.
Kou worked with Ph.D. student Shihao Yang and Mauricio Santillana, a lecturer in applied mathematics, to deliver results in real time.
The approach, called ARGO, for AutoRegression with Google search data, combines Google data with historical records from the CDC and information on seasonality of the flu. It also accounts for changes in the inner workings of Google’s search engine and shifts in search behavior. People learn as they search for information, Kou said, changing their queries and becoming better searchers.
“If I want to search for something, I do it better now than I did two years ago. Besides, Google’s search engine evolves, and so does the interaction between people and the engine,” Kou said.
The project began about a year ago after a casual conversation on the topic between Kou and Santillana, who also has an appointment at Harvard-affiliated Boston Children’s Hospital. The team is currently working to make ARGO open source and widely available.
Getting to Know the Microbes that Drive Climate ChangeNews
A new understanding of the microbes and viruses in the thawing permafrost in Sweden may help scientists better predict the pace of climate change.READ MORE
How Many People Die From TB Every Year?News
Discrepancies between the estimates for global tuberculosis deaths is due to different methodologies and data sources used by each institution. The results highlight the need to improve the modeling approaches in these countries in order to understand the true burden of the disease and design adequate health policies.READ MORE
ExPecto Patronum! Magical Machine Learning Tool Summons DNA Dark Matter DataNews
A new machine learning framework, dubbed ExPecto, can predict the effects of mutations in the so-called “dark matter” regions of the human genome. ExPecto pinpoints how mutations can disrupt the way genes turn on and off throughout your body.