Reproduction numbers for epidemic models with households and other social structures II: comparisons and implications for vaccination

TL;DR

This paper compares various reproduction numbers in structured epidemic models, clarifies their relationships, and discusses implications for vaccination strategies, revealing that social structure complicates epidemic control.

Contribution

It provides a comprehensive overview and new inequalities among reproduction numbers in structured models, and analyzes vaccination thresholds considering social structures.

Findings

$R_r$ generally bounds $R_0$ depending on epidemic growth.

Vaccination of $1-1/R_0$ may be insufficient in structured populations.

Sharper bounds for critical vaccination coverage are derived.

Abstract

In this paper we consider epidemic models of directly transmissible SIR (susceptible $\to$ infective $\to$ recovered) and SEIR (with an additional latent class) infections in fully-susceptible populations with a social structure, consisting either of households or of households and workplaces. We review most reproduction numbers defined in the literature for these models, including the basic reproduction number $R_0$ introduced in the companion paper of this, for which we provide a simpler, more elegant derivation. Extending previous work, we provide a complete overview of the inequalities among these reproduction numbers and resolve some open questions. Special focus is put on the exponential-growth-associated reproduction number $R_r$, which is loosely defined as the estimate of $R_0$ based on the observed exponential growth of an emerging epidemic obtained when the social structure is ignored. We show that for the vast majority of the models considered in the literature $R_r \geq R_0$ when $R_0 \ge 1$ and $R_r \leq R_0$ when $R_0 \le 1$. We show that, in contrast to models without social structure, vaccination of a fraction $1-1/R_0$ of the population, chosen uniformly at random, with a perfect vaccine is usually insufficient to prevent large epidemics. In addition, we provide significantly sharper bounds than the existing ones for bracketing the critical vaccination coverage between two analytically tractable quantities, which we illustrate by means of extensive numerical examples.