Subscribe to DSC Newsletter

Swine Flu Virus Detected Earlier Via Data Mining Techniques

Kirkland firm noted outbreak in Mexico early

Weeks before the Centers for Disease Control and the World Health Organization alerted the public to a growing number of swine flu cases, a startup based in Kirkland already had a hunch something was up. Veratect Inc., a 2-year-old company with fewer than 50 employees, combines computer algorithms with human analysts to monitor online and offline sources for hints of disease outbreaks and civil unrest worldwide

By Jessica Mintz

The Associated Press

Weeks before the Centers for Disease Control and the World Health Organization alerted the public to a growing number of swine flu cases, a startup based in Kirkland already had a hunch something was up.

Veratect Inc., a 2-year-old company with fewer than 50 employees, combines computer algorithms with human analysts to monitor online and offline sources for hints of disease outbreaks and civil unrest worldwide. It tracks thousands of "events" each month — an odd case of respiratory illness, or a run on over-the-counter medicines, for example — then ranks them for severity and posts them on a subscription-only Web portal for clients who want early warnings.

The idea fueling Veratect and similar companies is that blogs, online chat rooms, Twitter feeds and news media and government Web sites are full of data that public health agencies could use to respond faster to problems like outbreaks of swine flu.

Veratect attracted attention in recent days by publicly posting a timeline of the outbreak and publishing short reports on Twitter, where more than 4,000 people signed up to receive updates.

But skeptics question whether these companies can reliably detect meaningful signals from all the noise online or whether they are mainly good at spotting patterns in hindsight. Complicating the picture, the companies are reluctant to disclose their sources and methods.

Veratect's chief executive, Robert Hart, says the company alerted clients to a potentially severe outbreak before the general public learned of swine flu. Veratect's computer systems, which troll the Web for reports that seem out of the ordinary, unearthed clues, and a team of about 30 analysts, many of them multilingual holders of public health degrees, chased down the ones that seemed most alarming.

Veratect says it posted a report to clients on April 6 describing an unusual number of respiratory illnesses in the Mexican state of Veracruz, then sent an e-mail on April 16 to the Centers for Disease Control pointing to an outbreak of atypical pneumonia in Oaxaca state, after officials there issued an alert.

A key clue came in Mexican media reports on April 6 indicating a Veracruz community called La Gloria — now considered a swine flu hot spot — was starting to point fingers. Local residents blamed waste from a nearby pig breeding farm for the respiratory illness, while health officials pinned it on a fly.

"Playing the blame game is one of those indicators" that something unusual is going on, said Dr. James Wilson, Veratect's chief scientist. When the company posted the La Gloria information, it treated the incident as a matter of "moderate severity."

To be sure, not everything Veratect turned up was related to the outbreak. Veratect told its clients of a Canadian lawyer hospitalized in late March after a trip to Mexico, but on Tuesday the company said he has since tested negative for swine flu.

Even with the flaws, clients like World Vision, the large Christian humanitarian organization based in Federal Way, pay Veratect for its intelligence.

Recently, World Vision shifted resources — water purification tablets and education staffers — to areas Veratect thinks might see cholera outbreaks, said Brian Carlson, the head of technology for World Vision's global relief efforts.

Full article: Click Here

Views: 283

Comment

You need to be a member of AnalyticBridge to add comments!

Join AnalyticBridge

On Data Science Central

© 2020   AnalyticBridge.com is a subsidiary and dedicated channel of Data Science Central LLC   Powered by

Badges  |  Report an Issue  |  Privacy Policy  |  Terms of Service