Influence of network dynamics on the spread of sexually transmitted diseases

TL;DR

This paper presents a flexible model of social network dynamics for sexually transmitted diseases, revealing that static network assumptions underestimate epidemic thresholds and that disease duration influences epidemic spread.

Contribution

It introduces an analytically tractable dynamic network model fitted to survey data, improving understanding of epidemic thresholds in social networks.

Findings

Static network data underestimates epidemic thresholds.

Dynamic network thresholds increase with infectious period.

Static thresholds serve as lower bounds for real epidemic thresholds.

Abstract

Network epidemiology often assumes that the relationships defining the social network of a population are static. The dynamics of relationships is only taken indirectly into account, by assuming that the relevant information to study epidemic spread is encoded in the network obtained by considering numbers of partners accumulated over periods of time roughly proportional to the infectious period of the disease at hand. On the other hand, models explicitly including social dynamics are often too schematic to provide a reasonable representation of a real population, or so detailed that no general conclusions can be drawn from them. Here we present a model of social dynamics that is general enough that its parameters can be obtained by fitting data from surveys about sexual behaviour, but that can still be studied analytically, using mean field techniques. This allows us to obtain some general results about epidemic spreading. We show that using accumulated network data to estimate the static epidemic threshold leads to a significant underestimation of it. We also show that, for a dynamic network, the relative epidemic threshold is an increasing function of the infectious period of the disease, implying that the static value is a lower bound to the real threshold.