Facing COVID-19 in Mexico City: from network epidemiology to modular economic reactivation.

The COVID-19 pandemic has had a terrible toll in cities across the world, including Mexico City. The main transmission route of SARS-CoV-2 requires close physical contacts between people; reducing these contacts has been the basis of mitigation strategies worldwide. The mexican government instituted a federally mandated soft (voluntary) lockdown (Jornada Nacional de Sana Distancia), reducing general mobility among the population to about 25% of its original levels for the months of April and May. With economic activities gradually and slowly reopening, the question of how many people can return to the public space without an increase of the epidemic activity remains open.
In this work, we used network epidemiology models to identify the effects of economic reactivation on the epidemic dynamics [https://arxiv.org/abs/2008.12688]. We used an contact network empirically reconstructed from mobile device localization data for Mexico City [https://arxiv.org/abs/2007.14596]. We modelled the lockdown as a removal of a 75% of links in this network. We then modelled economic reactivation as the reconnection of a fraction of nodes, representing people returning to the public space. We compared the epidemic dynamics for different fractions of randomly reactivated nodes, ranging from 0% (ie. continuing the lockdown) up to 100% (a full economic reactivation). We observe that the downward behaviour of epidemic curves is lost if more than 5% of the population returns to the public space; furthermore, a reactivation of 50% of the population leads to an epidemic dynamic indistinguishable from a full economic reactivation.
As an alternative, we analyzed the epidemic behaviour if the economic reactivation is done by reactivating nodes belonging to the same module or community within the network. We observe that by reactivating whole modules instead of randomly dispersed people, a larger fraction of the population can return to the public space without risking a new spike in the epidemic dynamics. Furthermore, by spreading the reactivated population throughout small, non-adjacent modules, a large fraction (~50%) of the population can return to economic activities without negative effects in the epidemic curve. We consider that such reactivation can be achieved in practice through a combination of public policy measures to manage the economic and health damages caused by the pandemic.