Real-Time Forecasting Global Epidemics

Vittoria Colizza presents GLEAM, a simulation tool used to predict and combat the spread of disease.

Vittoria Colizza is dedicated to the fight against new and emerging diseases across the world. New diseases pop up in some geographic regions while diseases long thought to be wiped out re-emerge in other countries. She cites the H1N1 virus in 2009 as an epidemic that needed very little help to become a worldwide pandemic. In her We Solve for X presentation of GLEAM Colizza shares her solution to the question  – “How do we radically improve our communication from modelers to public health authorities and the general public in order to fight pandemics?”

The ebola outbreak in West Africa is an area of great concern to Colizza. First she studies the spread of the disease, both the speed of the spread and the growth of the affected area. Next is risk assessment to predict the possibility that ebola could spread worldwide. If the virus does spread worldwide, the question becomes how it can be controlled and then battled.

Containing and combatting a pandemic is a problem that needs to be continually reevaluated. The last influenza epidemic was in the 1960s, Colizza says, and human movement and travel patterns are completely different now from the previous containment methods.

After a lengthy problem statement Vittoria presents her proposal for preventing the spread of an epidemic. First she studied the factors that contribute to spread of disease. People, personal interactions, movements, and disease progression all factor into a disease’s rate of infection. Several data sources were used to gain information, from population density maps to travel patterns and movement predictions. Her overall solution is called GLEAM – The Global Epidemic And Mobility model.

GLEAM first looks at people and where they live geographically, in relation to the major hubs of transportation. Mobility data is next, how people move in their region and between regions. The epidemic is modeled next, looking at different infection scenarios. The three systems are combined to create a worldwide model for spread of the specific disease.

GLEAM is a highly functional simulation that Vittoria frames using engineering principles. Geographical regions are broken down into pieces like the mesh in a finite element analysis. Several scenarios are run by randomizing several factors together and studying the outcome, like a Monte Carlo simulation. The GLEAM website itself explains a little of the stochastic algorithms and binomial and multinomial processes used to calculate the spread of disease, but it looks like a control diagram to me. This is a great talk with a rich problem dissection and a highly detailed look at the factors that work together to be a part of the solution.