The University of Pittsburgh team was assembled to improve the nation’s ability to detect, control, or prevent naturally emerging infectious disease threats or intentionally released microbial pathogens. To achieve this we have established a multidisciplinary Research Group with all the expertise and skills necessary to develop, test, evaluate, and deploy sophisticated agent-based (individual-based) computational models of infectious disease threats. Expertise domains represented on the University of Pittsburgh team include infectious diseases, epidemiology, ecology, biostatistics, time series analysis, non-linear dynamics, network theory, and decision theory. We plan to do the following to achieve our objectives:
- Identify and obtain epidemiological data sets that have high value in the creation of computational models of infectious disease threats, including historical and archived data sets, “real-time” epidemiological data, “proxy” epidemiological data on other infectious diseases that are not immediate threats but which could nonetheless reveal important transmission patterns, and simulated data sets, and work with the Informatics group to develop a library of such high value epidemiological data sets.
- Develop and evaluate (1) new software tools and methods for analysis of infectious disease epidemiological data sets, especially algorithms for analysis of non-stationary time series data, and use these tools to analyze high value epidemiological time/space data sets to detect epidemic disease transmission patterns (2) new software tools and methods for simulation of infectious disease epidemics, especially agent-based models and (3) new software tools and methods for “docking ” or calibrating simulated epidemics generated by the computational models to real world epidemiological data. Use these tools to create a suite of archetype agent-based (individual-based) models for important infectious diseases that are transmitted by different mechanisms, including infectious diseases transmitted primarily by the respiratory route (smallpox and SARS) and vector-borne diseases (dengue and West Nile). Use these tools to study epidemic patterns of “proxy” infectious diseases (influenza, measles) in order to develop generalizable predictive models.
- Utilize agent-based models as “experimental epidemiology” test beds to evaluate epidemic intervention strategies such as preventive vaccination, post-exposure vaccination, isolation, quarantine, or other interventions, alone or in combination. Also use agent-based and micro-simulation models to generate epidemics of non-existent but nonetheless hypothetically plausible infectious disease threats (such as genetically modified smallpox, or other currently unknown but conceivable microbial agents) and then evaluate conventional and unconventional strategies to contain such hypothetical threats.
- Develop new methods to measure and express the complexity versus utility of any particular infectious disease model, and develop new methods for evaluating and integrating results from more than one computational model when their outcomes differ. The intent is to develop and deploy user-friendly desk-top and internet-accessible versions of these computational models that are visually and conceptually accessible to policy-makers, as well as to the scientific community at large.
The University of Pittsburgh team intends to work closely with the NIGMS and other components of the MIDAS network, and in the event of an infectious disease emergency to rapidly refocus its efforts onto tasks of national importance.
