How Your Search Data Can Expose STD Outbreaks

Disclaimer: Results are not guaranteed*** and may vary from person to person***.

STD OutbreaksCatching disease outbreaks early is one of the best ways to stop them and a team from the University of Illinois wants to use Google searches to do this. For some time now, the search engine giant has been providing several institutes, such as the CDC, access to its data stores to aid in the development of analytical models that can track infections in as close to real time as possible.

The models are developed by combing through Google’s data to find search terms that spiked during prior outbreak periods. When considering sexually transmitted diseases, for instance, a jump in people looking up symptoms such as painful urination may be usable to estimate emerging cases.

Google’s data is highly sophisticated and can narrow searches based on city, state, significance, and be combined with other user information to create predictions of how a disease is (or is not) spreading.

Prior to Google’s August invitation for researchers to apply for unlimited data access, scientists had to make do with the more limited information available through Google Trends. This is a tracking feature of popular searches but has limits on how many phrases can be tracked and the tool does not display search results that fail to meet certain volume standards, which can hide values the scientists would want.

This is not the first time Google has involved itself in attempts to track the spread of disease, either. In 2008, the company launched its Flu Trends tool that aimed to provide near real-time tracking of the seasonal illness. Although a step forward in a number of areas, the tool became notorious for over-predicting cases and not being adjusted to take the later introductions of other Google features (like suggested searches) into account. Flu Trends has since been discontinued.

Several other flu tracking tools exist that incorporate Google results and findings from social media posts, among other data sources. Where the University of Illinois’s model would be most effective is in trying to track emerging outbreaks of STDs and other illnesses that have symptoms people are prone to being less vocal about. Someone may post a tweet about getting a fever but only tell their search bar that they are facing bloody diarrhea or genital bumps.

There are, of course, privacy concerns surrounding access to such large volumes of data. The researchers emphasize that even their access does not permit learning details like a searcher’s sex, ethnicity, or geographic location more precise than a city.

Source for Today’s Article:
Jaklevic, M., “Disease Sleuths Analyze Google Searches To Stop Infections,” NPR web site, December 10, 2015; http://www.npr.org/sections/health-shots/2015/12/10/458953265/disease-sleuths-analyze-google-searches-to-stop-infections.