April 2015

Using “Big Data” to Optimize Public Health OutreachAnswering the Call to Action

Author Affiliations
  • 1Medical student at Harvard Medical School, Boston, Massachusetts
  • 2Department of Dermatology, Massachusetts General Hospital, Boston

Copyright 2015 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.

JAMA Dermatol. 2015;151(4):367-368. doi:10.1001/jamadermatol.2014.3176

Studies show that more than 90 million Americans use the Internet as a resource for health information, often by entering queries into popular search engines. These queries are logged into mass archives of “big data” that are freely accessible to the public. New studies have shown that big data can be used to track disease incidence and to gauge America’s interest in public health concerns. Dermatology has yet to grasp hold of this wealth of information.

The best-studied application of big data has been in influenza. As influenza spread, Americans assessed their symptoms with online searches. Researchers used “Google Flu Trends” to predict weekly influenza activity across the United States in real time. This was an exciting breakthrough, considering that traditional Centers for Disease Control and Prevention surveillance data are reported with a 2-week lag, often after the peak of an epidemic.1 Early Google Flu Trends models overestimated the absolute incidence of influenza. As methodology was refined to correct inaccuracies, investigators realized that online search volume reflected not only disease activity but also media publicity.2 In fact, search engine query data have proven to be an invaluable asset in gauging public interest. Google Trends has been used to gauge interest in breast cancer screening,3 bariatric surgery,4 and uterine fibroids.5

First Page Preview View Large
First page PDF preview
First page PDF preview