Controlling Epidemics

Controlling Epidemics
Department of GeographyUniversity of MarylandCollege Park, MD 20742	Dr. Catherine Dibble	cdibble@geog.umd.edu Phone: 301-405-0637 FAX: 301-314-9299

Epidemics of infectious diseases such as SARS, smallpox, and periodic killer influenzas such as the 1918 Flu can cause severe harm: from mortality rates as high as thirty percent, from lost productivity and complications even for those who recover, and especially from social and economic disruption due to panics and economic inactivity. No matter how well we research new vaccines or train our public health workers and National Guard Troops to help with contact tracing and quarantine enforcement, effective geographic deployments and timely interventions can be crucial for controlling epidemics. Yet predicting how a particular epidemic will unfold and where it will strike first, or next, via mathematical models has been impossible due to the complex interactions of millions of individuals traveling via richly connected high speed transportation networks. Similarly, if we try to make predictions based upon empirical models of the behavior of past epidemics, we are limited to those few events for which we have data, each of which was shaped by the particular historical accidents or travel whims of infectious individuals. We have developed agent-based computational laboratory models of infectious disease epidemics that spread among mobile individuals who travel between cities on richly structured transportation networks. The advantage of such simulation models is that we can rewind and rerun variants of the historical tapes of our epidemic simulations as many times as necessary in order to develop a complete understanding of the dynamics of particular epidemics among particular populations and particular transportation systems. In addition, we have developed a complementary evolutionary search and optimization tool (GA) to systematically evaluate many thousands of scenarios in order to determine the most effective interventions and deployments for a very broad range of potential scenarios and outcomes. We can answer the following questions for any infectious disease affecting any population on any transportation network:
1. Which cities are at greatest risk from an epidemic? Where should we stockpile vaccines? Where should we base public health reinforcements a priori? Which cities should emphasize epidemiological public health awareness and training programs for first responders, Civilian Reserves, and other citizens? 2. Which cities could serve as the most effective epidemiological “firebreaks” to control the spread of an epidemic to other parts of the nation? Which cities should be targeted for emergency vaccinations, if they are available? Which cities should be targeted for public health reinforcements to enforce contact tracing and/or quarantines (if costs are fixed for each city? if costs vary by population?)? 3. In order to minimize disruption of transportation services and economic activity, which transportation links are crucial for controlling the spread of the epidemic? Which airline flights, train routes, or highways should be blocked or carefully monitored (e.g. by deploying infrared fever sensors as for SARS)?
Each risk or intervention strategy can be identified both for potential epidemics (of unknown or of hypothesized geographic origin(s)) and for actual epidemics (where the initial geographic origin(s) and epidemic extent are known). ONR has funded us to develop the capability for providing these answers, which is now complete. We have not yet been funded to develop a calibrated model for the United States, but we could do so quickly if the need arises and if funding becomes available.

Figure 1:	A GeoGraph model of SARS, where individual agents travel between communities and bar charts for each community show the epidemiological status of its population. Green agents are healthy, pink are infected, red are sick, gray are dead, and white are recovered and immune. The model includes super-spreader events and seasonally adjusted infectivity. Gray links are base links, such as highways or trains, yellow links are high-speed shortcuts in the landscape, such as airline routes.
Funding
PI	Office of Naval Research, Grant N000140310062: A GeoGraph Simulation Platform for Modeling Mobile Agents on Richly Structured Network Landscapes, 2002-2003, $74,951.
PI	Office of Naval Research, Grant N000140310062: GeoGraph Network Models of Epidemics and Civil Violence, 2003-2004, $80,000. (renewal)