Small-number effects and cooperativity enhance epidemic containment by regional measures*

Epidemic containment with minimal restrictions for citizens is a central goal for policy makers. We showed earlier that population subdivision can resurrect certain stochastic effects that reduce the impact of an epidemic [Bittihn & Golestanian, Chaos, 2020]. Here, we ask whether similar effects also enhance the efficacy of regional containment measures triggered locally when infection numbers reach a critical threshold.
We build a stochastic meta-population model using data on COVID-19 spread in Germany, Italy, England, New York State and Florida, including their respective regional structures and lockdown efficiencies. As a performance measure, we determine the restriction time, that is, the total time individuals will experience restrictions over the next 5 years.
We show that the leakiness (the rate of cross-regional infections) and the regional structure itself are decisive: Below a critical leakiness, cooperative, small-number effects substantially lower the restriction time required by regional measures compared to national ones. Moreover, a more fine-grained county structure makes control significantly easier. Our analysis shows the importance of monitoring cross-regional infections and tight regional control [medRxiv 2020.07.24.20161364].