Computer models help forecast spread of Zika virus

With about 60 countries and territories worldwide reporting active Zika virus transmission, predicting the global spread of the mosquito-borne illness has been challenging for public health officials.

However, researchers are leveraging large-scale computational models that integrate socio-demographic and travel data as well as simulations of infection transmission—requiring the computing power of 30,000 processors simultaneously—to project the path of the disease.

The Global Epidemic and Mobility (GLEaM) model has been used in the past to simulate the spread of Ebola, H1N1 flu, and other outbreaks on a worldwide scale. However, in forecasting Zika, researchers have relied more on the historical patterns of mosquito-borne diseases such as chikungunya and dengue.

While the Zika virus can also be transmitted sexually, their computer model does not take that mode of transmission into consideration. They describe the Zika virus epidemic as “characterized by slow growth and high spatial and seasonal heterogeneity, attributable to the dynamics of the mosquito vector and to the characteristics and mobility of the human populations.”

In fact, mosquitoes bring an added level of difficulty to the equation, given the uncertainty of their travel behaviors, abundance and lifecycle depending on temperature, as well as the relationship between Zika and its host mosquitoes.

According to Alessandro Vespignani, professor of physics and director of the Network Science Institute at Northeastern University, what makes Zika such a challenge to track and predict is that as many as 80 percent of people infected with the virus are asymptomatic, and it is primarily transmitted by mosquitoes and spread internationally through travel.

But by combining real-world data on populations, human mobility and climate with elaborate stochastic models of disease transmission, a team of 14 researchers—half drawn from Northeastern—has devised projections for the number of Zika cases in the Americas through January 2017.

“Whatever the disease surveillance systems tell us, it is just the tip of the iceberg,” contends Vespignani. “What they are able to monitor and detect is much less than what is the reality of the gestation in many places.”

Also See: Registries helping CDC track pregnant women infected with Zika virus

The good news is that modeling algorithms do not predict very large Zika outbreaks in the continental United States, according to Vespignani. While the state of Florida has had the highest risk for Zika transmission, he notes that researchers “project at most a total of a few hundred cases in Florida in the next three months” and “in other places of the continental U.S., there are very minimal probabilities of a few cases.”

This downward pressure on the spread of Zika is being aided by the end of mosquito season which is reliant on weather conditions. Mosquitoes multiply fastest in tropical and sub-tropical habitats, and dropping temperatures this fall and into winter are unfavorable for mosquitoes to thrive.

“At this point, Florida and the Gulf Coast are approaching the winter season. And although it’s not as cold as in the northeastern U.S., it’s not as good for mosquitoes,” says Vespignani. “The models indicate that by January and February the number of new Zika cases should be declining drastically.”

Still, there is continued bad news for the U.S. territory of Puerto Rico, which has been ravaged by the virus, and, if current trends continue, as many as 25 percent of the population of almost 3.5 million will become infected with Zika by the end of the year.

Like chikungunya and dengue, Vespignani says Zika isn’t a threat that is going away anytime soon, and he predicts that it will come in waves as a seasonal phenomenon. “The main question now is what will happen in Asia and Africa?” he concludes, where more than 2 billion people could be at risk from contracting the virus.

In addition to Northeastern University, the research team modeling Zika included the University of Florida in Gainesville, Fla., Bruno Kessler Foundation in Trento, Italy, Bocconi University in Milan, Italy, Institute for Scientific Interchange Foundation in Torino, Italy, Fred Hutchinson Cancer Research Center in Seattle and the University of Washington.