Early detection of local COVID19 outbreaks by normalized SPC charts

Yuval Foox is an M.Sc student in the Department of Industrial Engineering in Tel Aviv University

Thursday, December 9, 2021, 12:00 PM at Room 206  and Via Zoom 

Abstract:

As the COVID-19 pandemic and its variants continue to spread, it is important to identify infected areas as soon as possible.

In order to allocate resources properly (vaccines, manpower, etc.) it becomes important to test and analyze the COVID19 positive rate daily in each area (e.g. neighborhoods, districts, small towns) as each area behaves differently over time.

We use Statistical Process Control (SPC) based decision-making to support rapid responses to changes and early detection of local outbreaks, as specifically requested by a large Health Maintenance Organization (HMO), with whom we are collaborating.

Our method applies a normalized p-chart which addresses different sample sizes, that reflect the different number of people that are tested daily in each location, as well as other effects to train and learn the z-score limits for alerts.

While maintaining high and balanced levels of precision and recall, the proposed model outperformed other monitoring methods, as well as the current HMO's heuristic that has been used. Examples based on real data will be presented

Bio:

Yuval Foox is an M.Sc student in the Department of Industrial Engineering in Tel Aviv University, under the supervision of Prof. Irad Ben-Gal. He holds a B.Sc in Industrial Engineering from Tel Aviv University. He works as a data science team lead at AppsFlyer