Blog | Thursday, January 19, 2012

QD: News Every Day--Google tracks flu peaks faster than the CDC


Google can calculate flu incidence peaks a week faster than other diseases surveillance systems, including that of the Centers for Disease Control and Prevention. With hospitals already stretched to capacity and using just-in-time logistics to serve patients, the seven extra days of warning could help them during a pandemic, researchers concluded.

Google Flu Trends uses search engine query data to estimate influenza activity. By counting how often Google sees 45 flu-related search queries, it estimates how much flu is circulating in different countries and regions around the world.

The data are available in near real time, unlike previously developed surveillance systems such as call volume to telephone triage services, over-the-counter drug sales and volumes of emergency department patients with influenza-like illness.

Researchers studied correlation of Google data to influenza and crowding from an inner-city emergency department at an urban academic hospital in Baltimore from January 2009 through October 2010. Data were collected weekly for the emergency departments, the CDC, laboratory-confirmed influenza data and emergency department crowding (patient volume, number of patients who left without being seen, waiting room time, and length of stay for admitted and discharged patients). Pediatric and adult data were analyzed separately when compared to Google Flu Trends.

Results appeared online Jan. 8 at Clinical Infectious Diseases.

Google Flu Trends correlated with number of positive influenza test results (adult ED, r=0.876; pediatric ED, r=0.718) and number of emergency department patients presenting with influenza-like illness (adult ED, r=0.885; pediatric ED, r=0.652). Pediatric but not adult crowding measures, such as total emergency department volume (r=0.649) and leaving without being seen (r=0.641), also had good correlation with Google Flu Trends. Adult crowding measures for low-acuity patients, such as waiting room time (r=0.421) and length of stay for discharged patients (r=0.548), had moderate correlation.

Researchers concluded that Google Flu Trends provides near-real-time surveillance data seven to 10 days before the CDC's U.S. Influenza Sentinel Provider Surveillance Network, correlating well "but not perfectly" with flu activity.

The imperfection was that Google Flu Trends was prone to spikes caused by news coverage rather than sick people seeking information.

One flu peak was not detected by Google Flu Trends, "possibly because of the previous month's flurry of Internet activity surrounding the news coverage of the H1N1 outbreak," the authors wrote. Another flu peak recorded by Google was not mirrored in the number of patients with flu-like symptoms or positive influenza tests, and instead was probably caused by news coverage of the CDC declaring H1N1 as a national public health emergency.