RAPID: Projecting Possible Scenarios of COVID-19 in the United States

Models of emerging infections provide invaluable insight into the potential impacts and trajectories of unfolding epidemics. In a rapidly changing situation, like that of the ongoing COVID-19 pandemic, unforeseeable changes in behavior, control measures, pharmaceutical interventions like vaccination, and changes in the virus itself make it challenging to provide accurate disease forecasts more than 3-4 weeks into the future. Many policy decisions require multiple months of advanced planning, however. Scenario projection models provide an essential framework for assessing what might happen during an outbreak in the longer-term, such as 3-6 months, under scenarios with well-defined assumptions about things like control measures, vaccination, and variant spread in the future. Because of real uncertainty in how an epidemic situation may unfold, combining scenario projections from multiple models with different assumptions about the future can provide a more complete picture of the epidemic possibilities. The useful production and interpretation of multiple scenario projections rests critically with the careful definition of questions, the explicit specification of scenarios, and the concise communication of these assumptions and findings, tasks that are not always accomplished in the use of epidemiological modeling for policy. This project will provide training opportunities for a graduate student. Through collaboration with the U.S. CDC and other public health agencies, the COVID-19 Scenario Modeling Hub project is well poised to shape the research and implementation capacity of forthcoming U.S.-wide epidemic and disease prediction initiatives. This project supports the coordination of the COVID-19 Scenario Modeling Hub (SMH), an effort that aims to provide decision makers with a form of "situational awareness" under multiple plausible futures by serving as a central, harmonized source of timely scenario modeling projections. The SMH aims to propel the science of scenario projection modeling, thus leading the way for future, long-term coordinated multi-modeling exercises. These efforts are focused on (1) defining tractable policy questions, (2) designing standardized scenarios for multiple modeling groups to answer those questions, with specific focus on limiting uncertainty in scenario-related components, while allowing for different model assumptions and implementations in other aspects; and (3) communicating and interpreting model results to help guide policy and highlight legitimate sources of scientific disagreement. Additionally, this project also supports ongoing adjustment and improvement of the Johns Hopkins University COVID-19 Scenario Pipeline model, which both contributes to the SMH and provides the SMH coordination team with a useful tool for experimenting with scenario specifications.