People’s Google search histories can reflect much about their lives. For example, a parent whose child has come down with the flu may turn to Google to seek information on pediatric flu treatment.
In the past decade, researchers have tried to harness real-time Google search data to capture instant snapshots of geographical disease patterns. These efforts have met with mixed success, but new research published in PLOS Computational Biology shows that Google searches could help alert officials to dengue outbreaks in underdeveloped countries.
Dengue is a mosquito-borne tropical virus that causes flu-like symptoms and sometimes death. It infects an estimated 390 million people every year, but it can be difficult to monitor with traditional hospital reports in countries lacking efficient disease surveillance systems.
“The wide availability of internet throughout the globe provides the potential for an alternative way to reliably track infectious diseases, such as dengue, faster than traditional clinical-based systems,” says Mauricio Santillana of Harvard Medical School, who led the new study.
Santillana’s team investigated whether dengue-related Google searches could have helped health officials estimate dengue prevalence in real time during past outbreaks in Mexico, Brazil, Thailand, Singapore, and Taiwan—a diverse group of countries with varying levels of public health data availability.
Using historical data from Google’s “Trends” tool, the researchers tracked the top ten dengue-related search queries made by people in each country over several years. They also gathered government health data that would have been available at the time of estimation.
The researchers fed both datasets into a mathematical modeling tool called AutoRegression with GOogle search queries (ARGO), which they had initially developed in 2015 to track influenza in the United States. For each country, ARGO calculated near real-time estimates of dengue prevalence.
To gauge the quality of ARGO’s estimates, the researchers compared them with estimates made using five other methods that incorporate Google search data. For Mexico, Brazil, Thailand, and Singapore, ARGO’s estimates were the most accurate.
However, ARGO’s dengue prevalence estimates for Taiwan were not so accurate. This may be because dengue infection patterns were less consistent from year to year in Taiwan than in the other countries.
Nonetheless, ARGO could help revive hopes in the power of Google search data in light of disappointing results from past disease-tracking efforts, such as Google Flu Trends. And while country-wide data may not have much practical use, ARGO could potentially be improved to track dengue at a more local scale.
“This alternative way of tracking disease could be used to alert governments and hospitals when elevated dengue incidence is anticipated, and provide safety information for travelers,” Santillana says.
Reference: Yang S, Kou SC, Lu F, Brownstein JS, Brooke N, Santillana M (2017) Advances in using Internet searches to track dengue. PLoS Comput Biol 13(7): e1005607. https://doi.org/10.1371/journal.pcbi.1005607
Image Credit: Anderson Mancini