Impact of origin-destination information in epidemic spreading

TL;DR

This paper investigates how origin-destination flow patterns in air transportation networks influence epidemic spreading, revealing their importance for microscopic predictions but limited impact on macroscopic epidemic incidence.

Contribution

It introduces a data-driven approach to incorporate origin-destination flows into epidemic models, highlighting their role in detailed infection predictions.

Findings

Origin-destination data improves microscopic infection predictions.

Flow information has limited effect on overall epidemic prevalence.

Data-driven flow estimation enhances epidemic forecasting accuracy.

Abstract

The networked structure of contacts shapes the spreading of epidemic processes. Recent advances on network theory have improved our understanding of the epidemic processes at large scale. The relevance of several considerations still needs to be evaluated in the study of epidemic spreading. One of them is that of accounting for the influence of origin and destination patterns in the flow of the carriers of an epidemic. Here we compute origin-destination patterns compatible with empirical data of coarse grained flows in the air transportation network. We study the incidence of epidemic processes in a metapopulation approach considering different alternatives to the flows prior knowledge. The data-driven scenario where the estimation of origin and destination flows is considered turns out to be relevant to assess the impact of the epidemics at a microscopic level (in our scenario, which populations are infected). However, this information is irrelevant to assess its macroscopic incidence (fraction of infected populations). These results are of interest to implement even better computational platforms to forecast epidemic incidence.