(mcfarlandmo via wikimedia commons)

(mcfarlandmo via Wikimedia commons)

A flu forecasting system that uses some of the same techniques as modern weather forecasting was not only able to  predict the timing of the 2012-2013 influenza season, but was able to do so nine weeks before the flu season peaked.

Columbia University researchers say there were able to correctly predict that flu activity in the southeastern United States would peak in December 2012, and that the rest of nation wouldn’t see its flu season crest until the first weeks of 2013.

A study published in Nature Communications revealed the results of the researcher’s first large-scale demonstration, which tested the forecasting system in 108 U.S. cities.  The World Health Organization (WHO) says between 250,000 to 500,000 people worldwide die from influenza.  In the United States, between 3,000 to 49,000 people die from the flu every year, according to the Centers for Disease Control (CDC). The CDC says about 45 percent of Americans receive annual flu inoculations.

The demonstration of Columbia University’s flu-predicting system follows the team’s study last year that predicted, in retrospect, the peak of the New York City flu season for each year between 2003 and 2008.

Predictions made in last year’s study were the initial test of the system and were limited to one city.  The researchers said the new and expanded demonstration made its predictions in real-time for a number of cities throughout the country.

A microscopic image of the H1N1 ('swine flu') influenza virus - In 2009, the World Health Organization declared this new strain as a pandemic.

A microscopic image of the H1N1 (‘swine flu’) influenza virus . (wikimedia commons)

The timing of flu season can vary from year to year and from region to region.  Influenza usually arrives any time between December and April, but when it does hit, a city can go from having no, or very few cases, to thousands within a short time period.

“Having greater advance warning of the timing and intensity of influenza outbreaks could prevent a portion of these influenza infections by providing actionable information to officials and the general public,” said  study author Jeffrey Shaman,  assistant professor  at Columbia University’s Mailman School of Public Health.

This advance warning could provide people with information that might encourage them to get their flu vaccination, become more aware of their own personal health,  and perhaps exercise extra caution when they’re around people who sneeze and cough.

The prediction system could also help health authorities prepare for an upcoming influenza outbreak.  It could help inform their decisions on how many vaccines and antiviral drugs to stockpile, and whether to take other measures such as closing local schools in the event of a virulent outbreak.

To test their prediction system on a large scale, the researchers began to perform weekly estimates for 108 US cities in November 2012.  The results of their weekly estimates were posted online and shared with the CDC.

They found that by the end of 2012, or four weeks into the flu season, their forecasting system had made accurate predictions in 63 percent of the cities they measured. That accuracy increased as the flu season moved forward. By the fourth week, the forecast system had accurately predicted the peak of the flu season in 70 percent of the country.

The flu forecasts were able to provide precise lead-times of up to nine weeks in advance of the peak, instead of the usual two-to-four weeks offered by other prediction methods that use historical data instead of real-time data.

A flu shot may sting a little bit but the US CDC recommends a yearly flu vaccine as the first and most important step in protecting ourselves against flu viruses. (Photo: US Navy)

A flu shot may sting a little bit but the US CDC recommends a yearly flu vaccine as the first and most important step in protecting ourselves against flu viruses. (Photo: US Navy)

In reviewing the results of their measurements, the researchers noticed regional differences regarding the accuracy of the system.

“We were able make better predictions in smaller cities. Population density may also be important,” said Shaman. “It suggests that in a city like New York, we may need to predict at a finer granularity, perhaps at the borough level. In a big sprawling city like Los Angeles, we may need to predict influenza at the level of individual neighborhoods.”

Shaman and his colleagues are gearing up for the 2013-2014 flu season and will be ready to put their forecasting system back into action as soon as it begins.

“Right now there are few cases of the flu, but as soon as the needle starts to move, we will start making predictions,” said Shaman.

Flu forecasts this season will be more available to the public because they will be posted on a soon-to-be launched website hosted by Columbia’s Mailman School of Public Health.