Award Abstract #3P20GM103446-20S2

Predictive Modeling & Optimal Control Framework for Model-Based Epidemic Response in Delaware

Search for this grant on NIH site
Program Manager:


Active Dates:

Awarded Amount:



Steven J. Stanhope

Awardee Organization:

University of Delaware

Funding ICs:

National Institute of General Medical Sciences (NIGMS)


AND ABSTRACT. In this project, our team seeks to develop and evaluate a unique predictive modeling approach that can be applied to the spread of SARS-CoV-2 and subsequently adapted to address other emergent infectious diseases. Robust and accurate predictive models are needed to allow healthcare and public health experts to devise and evaluate interventions for controlling viral spread and mitigating the effects of disease. When new disease-caus- ing viruses arise (such as the recent novel coronavirus SARS-CoV-2), deploying useful predictive models is challenging. First, the transmission characteristics within often diverse populations are not immediately under- stood; and second, many existing model frameworks are based on a necessarily simplified set of serially-com- partmentalized transmissions between susceptible, exposed, infected, and recovered (SEIR) groups that may not accurately represent the realities of the new virus. In the current proposal, we develop and validate a novel modeling approach based on principles of chemical reaction kinetics (CRK). The CRK approach allows us to model the infection/transmission of any virus with the same formalism employed to describe the chemical reac- tion of one molecule (an infected individual) with another (an uninfected). Our approach also employs “Residence Time Distribution” theory, which is typically applied to understand large-scale chemical reactors where reagents move, mix, and interact in a complex manner, to capture elegantly and effectively the uncertainties involved in complex disease processes, especially those resulting in recovery or mortality. In the long-term, our CRK-based system will provide a readily-adaptable and facile framework that can be linked to relevant data streams available through INBRE partner institutions in the state of Delaware, which has a population basis that is broadly repre- sentative of the nation, to allow rapid use. In addition, the model itself will be accessible to researchers, clinicians, and public health experts through a convenient online interface. In the long-term, the model and interface will be vetted and deployed following a detailed Resource Sharing Plan designed to assure usability and impact. In the initial 12-month study proposed here, we begin development of this system by utilizing existing datasets for SARS-CoV-2 to deploy a flexible model framework that predicts fundamental aspects of SARS-CoV-2 spread. In short, a set of ordinary differential equations is developed based on CRK principles where the “reaction rate constants” directly connect to physiological and epidemiological parameters and where recovery and death are characterized by directly determinable “Mean-Time-To-Recovery” and “Mean-Time-To-Death” parameters. We will proceed by addressing two aims: 1) Develop and validate a “Chemical Reaction Kinetics”-based model of COVID-19 infection for the State of Delaware; 2) Develop Optimal CRK Model-Based Mitigation Strategies and Implement a Model and Mitigation Strategy in a User-Friendly Software Appropriate for Policymakers. Execution of these aims will provide a new CRK-based model that will provide a foundation for more advanced modeling tools and will serve as a powerful adjunct to the more traditional models based on SEIR-derived frameworks.

Back to Top