Our short-term goal is to rapidly build an integrated surveillance and forecasting system to support public health decision making for the Commonwealth of Pennsylvania (PA). We aim to integrate an array of federal and state level cases data, meteorological and environmental data, pharmaceutical sales, electronic health records data, and human judgment into an ensemble of mechanistic and statistical forecasting models. Our overall objective is to build an integrated surveillance system to better inform public health officials for the state of PA. We hypothesize an ensemble forecast trained on high-resolution real-time data from a variety of sources will support public health decision making and reduce the impact of COVID-19 on the state of PA. To meet our overall objective and test the above hypothesis, we plan to pursue the following aims:
a. Build a real-time syndromic surveillance system for PA to monitor the COVID-19 outbreak.
b. Train a stacked ensemble of statistical models to forecast the COVID-19 trajectory.
c. Develop a long-term surveillance system that integrates real-time data and models to detect novel infections and monitor current circulating diseases.
Results from the above expect outcomes will have a positive impact by reducing the COVID-19 burden on Pennsylvania and a broader impact on syndromic surveillance research, human judgment, and statistical models to inform public health decisions. Output from this work will support scientific output in applied statistics and rapidly address, and aim to reduce, the impact of the novel coronavirus.
- Centers for Disease Control (CDC) Fluview
- Johns Hopkins University's Center For Systems Science and Engineering COVID-19 Data repository
- The Delphi group at Carnegie Mellon University's COVIDcast
This project is supported by students participating in Lehigh University's Mountaintop Summer Experience, a 10 week intensive program with a focus on transforming ideas into practical tools.
tom mcandrew: mcandrew@lehigh.edu