RAPID: Adaptive Sampling Strategies for COVID-19 Mass Testing

点击翻译按钮获取中文摘要

As states begin to relax limitations on physical distancing in order to restart economic activity, mass testing for COVID-19 will be crucial in identifying and containing infection "hot spots" and to avoid more severe outbreaks. This Rapid Response Research (RAPID) grant will support the collection of time-sensitive data on serological and viral test outcomes. These data, along with census block-level demographic information and mobility patterns, will be used to develop a data-driven strategic framework for mass testing. As testing resources will be limited at the onset, this framework can help guide a testing strategy to use these resources most effectively. The project involves a collaboration with LifeSouth Blood bank and the State of Florida Department of Health to utilize data from north central Florida and is expected to be scalable to other counties, regions, and states across the Nation. This research approach is expected to help mitigate the negative impacts of COVID-19 on public health, society, and the economy.The project will collect community testing data in order to develop a data-driven adaptive-sampling strategy to optimize mass testing within census block groups based on (i) aggregated community population, (ii) daily testing capacity and outcomes, (iii) block group demographics related to contagion and morbidity vulnerability (e.g., race, age structure, employment type, housing crowding), (iv) symptom prevalence, and (v) mobility patterns of people in the block group. The sampling algorithms will “learn” from recent test outcomes (both positives and negatives) to optimize the sampling approach over time, balance exploration and exploitation to avoid overlooking critical areas while ensuring suspected areas are frequently sampled. The project involves four tasks: (i) data collection though collaboration with LifeSouth Blood bank and Florida Department of Health for COVID-19 antibody testing and state testing data; (ii) geographic analysis of community medical and social vulnerability data; (iii) development of an adaptive sampling algorithm to determine the most informative test allocation to regions in the community; (iv) cost-effective analysis to determine the daily budget to balance between testing power and costs.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.