Back

Failure of Concentric Regulatory Zones to Halt the Spread of COVID-19 in South Brooklyn, New York: October-November 2020

Harris, J.

2021-11-21 epidemiology
10.1101/2021.11.18.21266493 medRxiv
Show abstract

We relied on reports of confirmed case incidence and test positivity, along with data on the movements of devices with location-tracking software, to evaluate a novel scheme of three concentric regulatory zones introduced by then New York Governor Cuomo to address an outbreak of COVID-19 in South Brooklyn in the fall of 2020. The regulatory scheme imposed differential controls on access to eating places, schools, houses of worship, large gatherings and other businesses within the three zones, but without restrictions on mobility. Within the central red zone, COVID-19 incidence temporarily declined from 131.2 per 100,000 population during the week ending October 3 to 62.5 per 100,000 by the week ending October 31, but then rebounded to 153.6 per 100,000 by the week ending November 28. Within the intermediate orange and peripheral yellow zones combined, incidence steadily rose from 28.8 per 100,000 during the week ending October 3 to 109.9 per 100,000 by the week ending November 28. Data on device visits to pairs of eating establishments straddling the red-orange boundary confirmed compliance with access controls. More general analysis of device movements showed stable patterns of mobility between and beyond zones unaffected by the Governors orders. A geospatial regression model of COVID-19 incidence in relation to device movements across zip code tabulation areas identified a cluster of five high-mobility ZCTAs with estimated reproduction number 1.91 (95% confidence interval, 1.27-2.55). In the highly populous area of South Brooklyn, controls on access alone, without restrictions on mobility, were inadequate to halt an advancing COVID-19 outbreak.

Matching journals

The top 5 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.