News and Events

May 18, 2018

Pandemic Simulation used in Gates Shattuck Lecture

On April 27, 2018, Bill Gates was the featured speaker for the Shattuck Lecture, held during the 2018 Annual Education Program for the annual meeting of the Massachusetts Medical Society and New England Journal of Medicine. Mr. Gates lecture, titled “Epidemics Going Viral: Innovation vs Nature,” explored the issues and challenges faced with modern epidemics.  Specifically, the lecture focused on the lack of global preparation for a pandemic.

Health systems and their dependent infrastructure can—and frequently do—break down in the face of epidemics, making the potential destruction from a pandemic a frightening plausibility. The risks posed by emerging pathogens, or even bioterrorism, are large and increasing, so the probability of a wide-scale and lethal pandemic occurring in the next 50 or so years is significant.

Recent outbreaks and epidemics have demonstrated both the lack of preparedness in health systems as well as the potential lethality of the problems. Swine flu (H1N1) in 2009 and Ebola in 2014 exposed our inability to control the spread of disease and failure to mount appropriate public health measures locally and globally.  Clearly, there is an obvious need for early detection and global response systems.

To demonstrate the destructive power of an unchecked epidemic, Mr. Gates leveraged the modeling expertise of the Institute for Disease Modeling (IDM).  IDM specializes in both developing software for the mathematical modeling of infectious disease as well as researching the interventions and methodology necessary for interrupting transmission and eliminating pathogens. For this lecture, Dr. Hao Hu, the Sr. Research Manager of IDM’s Epidemiology section, and Bryan Ressler, a Sr. Software Engineer at IDM, worked together to create simulations modeling pandemic scenarios for the spread of an influenza-like pathogen.

Specifically, Dr. Hu utilized IDM’s Epidemiological MODeling software (EMOD), an open-source agent-based model, to determine the likely outcome for the spread of an airborne virus. To capture the risk of a global pandemic, the model was configured to simulate what would happen if a newly emerged virus—i.e., one in which health systems are not prepared for and likely do not have a vaccine to prevent—appeared in South East Asia. The model utilized typical SIR (Susceptible-Infectious-Recovered) dynamics: all individuals were assumed to be susceptible, and immunity to the virus did not wane. The virus was given an R0 (the basic reproductive rate, or the number of secondary infections caused by a typical case) of 2.1. For comparison, seasonal flu has an R0 of 0.9 – 2.1, with a mean of 1.3; the R0 of the 1918-1919 pandemic-causing Spanish flu is estimated to have ranged from 1.4 – 2.8, with a mean of 2. Individuals in the model could travel both locally and by airplane, thereby moving the virus within and across countries. With an initial seeding of about 800 infections, the simulation demonstrated how an unchecked epidemic can very quickly evolve into a pandemic within a wholly susceptible population.

To visualize the simulation results, Dr. Hu made use of IDM’s Vis-tools, visualization software developed by Bryan Ressler to explore spatio-temporal data. Both worked closely with the team from the Bill and Melinda Gates foundation to produce a presentation-ready visualization for the lecture. Several new features were developed for Vis-tools, including new map layouts, color gradients, and a “presentation mode” to maximize the viewable screen. To create the pandemic video used in the lecture, the visualization from Vis-tools was subjected to minor post-processing video editing to add text and pauses.

Overall, IDM is proud to have been asked to contribute modeling and software expertise for such an important lecture on such an important topic.

If you would like to embed the pandemic video in an article or website, it is available here. The following are the instructions and citation information for using the video:

For the transcript of the Shattuck Lecture, please click here.