Policy-makers across the country are facing the difficult decision of how to reopen their cities while preventing the further spread of COVID-19. We present a research methodology that combines granular information at the city-level on physical contacts across inviduals along with granular healthcare data and information on the economic and demographic composition of a city to help policy-makers decide which places to open and which ones to keep closed. Together, we help policy-makers discover which policies work to reduce the further spread of COVID-19 while minimizing economic hardships to its citizens.
Note: This work is still in progress and key results may change as we are updating our data and analysis using live data coming in weekly. Please contact us if you plan to use our results for practical applications.
At the moment, we test our framework in three cities: Sacramento, Chicago and New York. A key component of our framework are contact matrices, which capture the average number of contacts between two types of people. For illustration, the figure below plots a contact matrix across a single dimension (age). Already, we can see differences across cities: on average, Chicago residents have more contacts than Sacramento residents, especially for those between 18-49 years old.
We use our framework to test an array of policies that each city could adopt as they move into the next phase of responding to the pandemic.
By combining data on contact matrices, demographics, health outcomes, and ability to work from home we can estimate the effect of a given policy on hospitalizations and employment. The results for each policy in Sacramento and Chicago are plotted below. A given policy's effectiveness can vary dramatically by city.
|Mohammad Akbarpour, Stanford University|
|Cody Cook, Stanford University|
|Aude Marzuoli, Replica|
|Simon Mongey, University of Chicago|
|Abhishek Nagaraj, University of California, Berkeley|
|Matteo Saccarola, University of Chicago|
|Pietro Tebaldi, University of Chicago|
|Shoshana Vasserman, Stanford University|
|Hanbin Yang, Harvard University|