As the World Health Organization (WHO) upped the pandemic alert level to five last night, suggesting a pandemic is imminent, a group of scientists are piecing together forecasts to help authorities understand how the novel virus could spread.

 

Based on a “worst-case scenario”, which assumes that no measures are in place to stop the virus from spreading further, their calculations suggest that the illness-case count could rise to one thousand in the USA by 17 May. During the same period, and under these conditions, cases would be expected to continue to pop up sporadically in Europe, according the GLEaM model.

 

“The model doesn’t show a dramatic increase in cases over the next few weeks, even in the ‘worst case scenario’,” comments Alessandro Vespignani, who leads the team, for ETHF News. “[But] it does identify geographic hotspots of virus activity that we need to keep an eye on.”

 

Over the next three weeks some unaffected countries will be at an increased risk of seeing signs of the virus too, according to this model. Regions at risk include transport hubs in Africa such as South Africa, Morocco, Egypt and Nigeria, as well as the populous Asian countries of China and India.

 

The US forecast is “stunningly similar” to results produced by another group modelling how the swine flu outbreak could spread across the USA, according to Vespignani. This group, based at Northwestern University in the state of Illinois, are using a completely different model. The close match between the forecasts is quite amazing and encouraging, he says.

 

Vespignani and his team plan to update the forecast daily, publishing the results on a freely accessible website. The group hope to incorporate information about mitigation strategies into the model.

 

By raising the pandemic alert level to five the WHO is urging all countries to activate their pandemic response plans. These will include distributing antiviral drugs to infected people and their contacts, and closing schools to help stem the spread of the virus. Data reflecting these activities will be added to the model from today, as and when they become available. So the forecasts will eventually give a more representative picture of how the virus is actually spreading, helping to determine what effect response measures are having on the unfolding epidemic. “We expect this will produce scenarios that are more optimistic,” says Vespignani.

 

A supercomputer at Indiana University runs the model, which is updated continually with the number of people confirmed to have been infected with the swine flu virus. To estimate the future course of the epidemic, Vespignani and his team divided up the global population into little boxes that measure 15 miles by 15 miles on the world’s surface. The calculations take into account transport systems, people’s commuting patterns, the dynamics of influenza transmission, and demographic information about the population in each area.

 

“The idea is that we can see how we can use this forecasting tool along the way [during] a real time event,” says Vespignani.

 

Some experts say models that forecast the spread of infectious diseases can be inaccurate as they don’t take into account how different groups of people interact with each other. But Vespignani and his team stress the predictions can be used by policy makers alongside other tools.

 

Various health authorities dealing with the outbreak, including a group of scientists at the European Centres for Disease Control and Prevention, are receiving GLEaM’s forecasts, adds Vespignani.

 

Over the past few years the amount of research into the spread of infectious diseases using mathematical models has surged, mainly as a result of the 2003 SARS outbreak and the continuing risk of a bird flu pandemic. Previous modelling work has looked at how the spread of a new flu virus would be affected by school closures, vaccination policies, and frequent fliers. “This is the first time state-of-the-art modelling has been used in real time,” says Vespignani.