Abstract:We investigate methods to vaccinate contact networks -- i.e. removing nodes in such a way that disease spreading is hindered as much as possible -- with respect to their cost-efficiency. Any real implementation of such protocols would come with costs related both to the vaccination itself, and gathering of information about the network. Disregarding this, we argue, would lead to erroneous evaluation of vaccination protocols. We use the susceptible-infected-recovered model -- the generic model for diseases making patients immune upon recovery -- as our disease-spreading scenario, and analyze outbreaks on both empirical and model networks. For different relative costs, different protocols dominate. For high vaccination costs and low costs of gathering information, the so-called acquaintance vaccination is the most cost efficient. For other parameter values, protocols designed for query-efficient identification of the network's largest degrees are most efficient.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is how to implement vaccination strategies in the most cost - effective way in network epidemiology. Specifically, the author studied the method of maximizing the hindrance of disease transmission by removing nodes (i.e., vaccinating individuals) under a given network structure, and considered the cost of vaccination as well as the cost of collecting network information. The author used the Susceptible - Infected - Recovered (SIR) model as a general model for disease transmission and analyzed the performance of different vaccination protocols under different relative costs.
### Core Problems of the Paper
1. **Cost - Benefit Evaluation**: The paper proposed a framework to evaluate the cost - benefit of different vaccination strategies. Cost - benefit is measured by the net savings per person \(\chi\), and the formula is as follows:
\[
N\chi(f) = \Omega - \Omega'(f) - N f c_{\text{vacc}} - n(f) c_{\text{info}}
\]
where:
- \(\Omega\) and \(\Omega'(f)\) represent the expected epidemic scales without vaccination and with partial vaccination respectively.
- \(N\) is the total number of individuals.
- \(f\) is the proportion of vaccination.
- \(c_{\text{vacc}}\) is the cost of vaccination.
- \(n(f)\) is the number of queries required to obtain network information.
- \(c_{\text{info}}\) is the cost per query.
2. **Comparison of Different Vaccination Strategies**: The paper compared eight different vaccination strategies, including Random, Acquaintance, Random Walk, Two - step Heuristic, Degree, Coreness, Collective Influence, etc. These strategies show different efficiencies under different cost parameters.
3. **Influence of Network Structure**: The paper also explored the influence of different network structures (such as actual human contact networks and networks generated by configuration models) on the effectiveness of vaccination strategies. Through simulation experiments, the author found that different network structures would lead to different optimal strategies.
### Main Conclusions
- **Cost Influence**: For the case of high vaccination cost and low information collection cost, Acquaintance is the most effective strategy. For other cost parameters, Random Walk and Two - step Heuristic usually perform better.
- **Network Characteristics**: Data sets with different network characteristics (such as sexual contact networks, conference networks, hospital networks, etc.) will affect the selection of the optimal strategy. For example, for large - scale sparse networks (such as sex worker networks), Collective Influence is still the most effective although it requires global network information.
- **Cost - Benefit Optimization**: Through numerical simulation and theoretical analysis, the paper provides guidance for selecting the optimal vaccination strategy under different cost parameters.
In conclusion, this paper aims to provide a systematic method to evaluate and select the optimal vaccination strategy by comprehensively considering the cost of vaccination and the cost of information collection, so as to minimize the cost of disease transmission.