A Method for Reducing the Severity of Epidemics by Allocating Vaccines According to Centrality

TL;DR

This paper introduces a graph-based stochastic epidemic model and evaluates vaccine allocation strategies based on centrality measures to effectively reduce epidemic severity across Texas counties.

Contribution

It presents a novel epidemic modeling approach and demonstrates that centrality-based vaccine allocation can outperform traditional methods.

Findings

In-degree and inverse betweenness centrality methods are most effective.

Centrality-based strategies outperform random allocation.

The model can be integrated with conventional approaches for better epidemic control.

Abstract

One long-standing question in epidemiological research is how best to allocate limited amounts of vaccine or similar preventative measures in order to minimize the severity of an epidemic. Much of the literature on the problem of vaccine allocation has focused on influenza epidemics and used mathematical models of epidemic spread to determine the effectiveness of proposed methods. Our work applies computational models of epidemics to the problem of geographically allocating a limited number of vaccines within several Texas counties. We developed a graph-based, stochastic model for epidemics that is based on the SEIR model, and tested vaccine allocation methods based on multiple centrality measures. This approach provides an alternative method for addressing the vaccine allocation problem, which can be combined with more conventional approaches to yield more effective epidemic suppression strategies. We found that allocation methods based on in-degree and inverse betweenness centralities tended to be the most effective at containing epidemics.