When the University of Hong Kong speaks, I listen and take notes. These folks have dealt with everything from H5N1 through SARS, MERS, and H7N9. Via the South China Morning Post: Hong Kong's first 'flu forecasts' to predict timing and scale of outbreaks just like the weather, HKU scientists say.
Hong Kong will soon be able to predict the timing and intensity of a flu outbreak as accurately as the weather, thanks to a groundbreaking study by scientists at the University of Hong Kong and Columbia University in the United States.
The city's first flu forecasts will be able to pinpoint the peak of an outbreak three weeks in advance with accuracy as high as 93 per cent, said Dr Eric Lau Ho-yin, assistant professor at HKU's school of public health.
"These forecasts provide information on lead times that can be valuable for the public and health officials," Lau told the South China Morning Post.
"The public can choose to get a flu vaccine to protect themselves against infection before the peak season comes, while the government can anticipate how to plan health policies."
Lau said he expected the forecasts could be put into use this autumn, before the city is hit by the winter flu peak.
Hong Kong has seen a decrease in the number of flu cases this week after reaching the summer peak in late June to early July.
Between June 12 and 29, there were 173 cases of influenza requiring admission to intensive care, of which 127 patients died, according to the city's Centre for Health Protection.
This summer's flu season has been milder than that of the winter this year, which killed 496 people between January and April, compared to 133 deaths the previous winter.
Lau said it was more difficult to predict flu patterns in Hong Kong than in the US due to the city's subtropical climate. In the US, seasonal flu outbreaks were restricted to the winter, he said.
The flu forecasts employ a computer model to generate multiple simulations that mimic the behaviour of an outbreak. They have been used by Columbia since 2013 to publish weekly regional flu forecasts for more than 100 US cities.
Predictions for Hong Kong will be based on flu surveillance data published weekly by the government and a modified formula to suit the city's situation. The formula has to account for the dynamics of a subtropical climate and will incorporate mathematical techniques that prompt the forecast system to recalibrate itself and let go of certain assumptions.
Researchers ran a test case based on data from 1998 to 2013 to retrospectively generate weekly flu forecasts, and they found the system was able to predict both the timing and magnitude of the peaks of 44 flu epidemics over 16 years caused by specific flu strains, Lau said.
"Prediction accuracy varied depending on the strength of the outbreak and how far in advance the prediction was made."
Results from the study were published yesterday in the journal PLOS Computational Biology.