Using Google to Gauge Dengue Disease Activity

Joe Hannan

A new study shows that tracking Google searches may be an effective means of monitoring mosquito-borne disease activity.

A new study shows that tracking Google searches may be an effective means of monitoring mosquito-borne disease activity.

The study, published in the July issue of Computational Biology, suggests that a similar framework used to track influenza globally can also apply to diseases such as dengue. In other words, Google searches pertaining to dengue — when put through a statistical model – correlate to dengue activity, researchers say.

Among vector-borne diseases, the Centers for Disease Control and Prevention (CDC) estimates that as much as one third of the world’s population is at risk for dengue infection. The virus is one of the chief causes of death and illness in the planet’s tropic and subtropic regions. The CDC puts annual infection estimates at about 400 million people. The study authors called it one of the fastest-growing global diseases.

Related: Verily Looks to Swat Zika with Population Health Initiative

In response to the World Health Organization’s call for better early detection methods of dengue, the study authors investigated whether internet searches could serve as a disease activity barometer. The team turned its attention to data from Mexico, Brazil, Thailand, Singapore, and Taiwan. They then compared aggregated dengue time-series data from government agencies in each country to online search volume and online search term selection in Google.

“We used the multivariate linear regression modeling framework ARGO (AutoRegressive model with Google search queries as exogenous variables), previously used to track flu incidence using flu-related Google searches, to combine information from historical dengue case counts and dengue-related Google search frequencies with the goal of estimating dengue activity one month ahead of the publication of official local health reports,” the researchers wrote.

The comparison showed that the ARGO model outperformed current disease-tracking methodologies in 4 of 5 studied countries. Researchers saw fewer errors during periods of low infection and peaks of activity.

Current models in Taiwan outperformed ARGO, with internet access likely being a contributing factor. Internet penetration in Taiwan, researchers noted, may cause an inflated level of perceived disease activity. They added that ARGO outperformed current models in Brazil and Singapore — where the level of internet access is moderate – compared to Mexico and Thailand, where levels are lower.

“Web penetration is nevertheless still an important factor,” researchers said.

Despite the limitations, the researchers concluded that using Google and the ARGO method offers useful insights regarding disease activity.

“Our findings confirm that combining historical dengue incidence information with dengue-related Google search data, in a self-adjusting manner, leads to better near real-time dengue activity estimates than those obtained with previous methodologies that exploit the information separately.”

Keep up with the latest news, views, and insights by signing up for our newsletter.

Related

Microsoft, Machine Learning, and Better Mosquito Defense

Applying Big Data to Disease Outbreaks

Infants Potentially Exposed to Zika Virus Should Undergo Routine Eye Exams