PaperlayerView exact revision on arXiv

Source-linked AI summary

Vaccination strategies against COVID-19 and the diffusion of anti-vaccination views

Rafael Prieto Curiel, Humberto González Ramírez

arXiv:2009.13674v1physics.soc-phcs.SI

TL;DR

Limited vaccine supply and possible refusal make vaccination strategy and uptake important to COVID-19 control. The paper simulates disease and anti-vaccine-view diffusion on networks, finding that central individuals are more likely to adopt anti-vaccine views and that those views can reduce the benefits of targeted vaccination.

  • Problem

    Limited vaccine supply and possible refusal make it important to understand vaccination strategies and anti-vaccine-view diffusion on networks.

    Source: §1 · p. 1, §5.1 · p. 10

  • Method

    The paper simulates COVID-19 spread and vaccination strategies on networks alongside an SIR-like process in which exposed individuals may adopt anti-vaccine views.

    Source: §1 · p. 2, §3 · p. 5, §3.3 · p. 5

  • Results

    Central individuals are more likely to adopt anti-vaccine views, which can impose costs on optimal vaccination strategies.

    Source: §1 · p. 2

  • Takeaways & Limitations

    Targeting highly exposed people, including healthcare staff and other highly connected professions, offers the highest social gain in the paper’s analysis.

    Source: §5 · p. 10

    The study assumes the network for sharing anti-vaccination views is the same as the network for virus transmission.

    Source: §5.1 · p. 10

Abstract

from arXiv · show

Miss-information is usually adjusted to fit distinct narratives and can propagate rapidly through communities of interest, which work as echo chambers, cause reinforcement and foster confirmation bias. False beliefs, once adopted, are rarely corrected. Amidst the COVID-19 crisis, pandemic-deniers and people who oppose wearing face masks or quarantines have already been a substantial aspect of the development of the pandemic. With a potential vaccine for COVID-19, different anti-vaccine narratives will be created and, likely, adopted by large population groups, with critical consequences. Here, we analyse epidemic spreading and optimal vaccination strategies, measured with the average years of life lost, in two network topologies (scale-free and small-world) assuming full adherence to vaccine administration. We consider the spread of anti-vaccine views in the network, using a similar diffusion model as the one used in epidemics, which are adopted based on a persuasiveness parameter of anti-vaccine views. Results show that even if an anti-vaccine narrative has a small persuasiveness, a large part of the population will be rapidly exposed to them. Assuming that all individuals are equally likely to adopt anti-vaccine views after being exposed, more central nodes in the network are more exposed and therefore are more likely to adopt them. Comparing years of life lost, anti-vaccine views could have a significant cost not only on those who share them, since the core social benefits of a limited vaccination strategy (reduction of susceptible hosts, network disruptions and slowing the spread of the disease) are substantially shortened.

1 Introduction

Limited vaccine supply and refusal make targeted vaccination important, while anti-vaccine narratives may further hinder vaccination strategies. The paper models epidemic spread and anti-vaccine-view diffusion on networks to assess vaccination outcomes and the views’ costs.

1 Introduction

  • Targeted vaccination may be needed because mass immunisation may not happen soon and vaccine production and distribution pose major challenges.Unvaccinated individuals can still benefit when vaccination reduces susceptible hosts and limits disease spread.Source: §1 · p. 2
  • Opposition to vaccination may amplify outbreaks and friction vaccination strategies, while adherence can influence coverage and the time needed to eliminate transmission.The introduction describes diverse anti-vaccine narratives, including safety concerns, conspiracy theories, and alternative medicine.Source: §1 · p. 2
  • The paper considers vaccination strategies when logistical restrictions or refusal mean that only a portion of the population may be immunised.Its analysis compares five strategies with different vaccination rates.Source: §1 · p. 2
  • Anti-vaccine views can make central individuals more likely to refuse vaccination, imposing costs on optimal vaccination strategies.The model assumes individuals are equally likely to adopt these views after exposure.Source: §1 · p. 2

2 Modelling vaccination and anti-vaccination views on a network

The section motivates vaccination strategies that account for limited uptake and network structure, then frames anti-vaccination views as a network-diffusion process. It identifies uncertainty about vaccine properties and the scale of future refusal.

2.1 Dealing with non-universal vaccination

  • Limited vaccination strategies must balance reducing and delaying infection peaks with reducing the final infected population, since targeting vulnerable people may save lives without reducing viral circulation.Targeting highly exposed people may slow overall exposure, while age-based and random strategies reflect other priorities.Source: §2.1 · pp. 2–3

2.2 Modelling immunisation in a population

  • Network-based vaccination research must account for heterogeneous contacts, while the effects of non-adherence in network settings remain unexplored.The section notes that refusal by central nodes could make targeting the most connected people ineffective.Source: §2.2 · p. 3

2.3 Opinion dynamics and anti-vaccination views

  • Opinion-dynamics models describe how opinions spread and change through interactions, while network communities can reinforce similar views and confirmation bias.The section links this setting to the potential spread of opposing views toward immunisation if social and disease networks are entangled.Source: §2.3 · p. 4

2.4 Anti-vaccination views

  • The section reports substantial existing vaccine opposition but notes that COVID-19 vaccine immunity duration, reinfection, and the eventual scale of refusal remain unknown.These uncertainties bear on whether persistent herd immunity can be attained and whether outbreaks could recur.Source: §2.4 · p. 4

3 Methods

The methods model disease and anti-vaccine-view spread on networks, comparing vaccination strategies by years of life lost and epidemic duration. They assume permanent vaccine immunity and a permanent adoption decision after first exposure to anti-vaccine views.

3 Methods

  • The SVIR simulation spreads infection across network contacts until no infected nodes remain, then estimates deaths and years of life lost using age-dependent mortality.The population’s age and age at death are sampled from scaled Beta distributions, and the model treats age and network structure as independent for simplicity.Source: §3 · pp. 4–5

3.1 Vaccination strategies

  • Five strategies vaccinate by degree, betweenness centrality, peripheral position, age, or random selection.The first three use network attributes, Age uses age, and Random uses no node information.Source: §3.1 · p. 5

3.2 Comparing vaccination strategies

  • Strategies are compared using average years of life lost and time until infection stops, with simulations repeated 500 times across vaccination rates.The model assumes a perfect vaccine that grants permanent immunity, although the authors note immunity could wane.Source: §3.2 · p. 5

3.3 Diffusion of anti-vaccine views

  • Anti-vaccine views spread through contacts in an SIR-like process, with first-time exposure prompting a permanent adoption decision based on persuasiveness θ.Only those who adopt the views are assumed to reject vaccination; exposed non-adopters and unexposed individuals may be immunised.Source: §3.3 · pp. 5–6

4 Results

Targeting connected or central nodes reduces years of life lost more efficiently than random vaccination, though it can prolong transmission. Anti-vaccine views can spread quickly, expose much of the population, and erode the benefits of targeted vaccination.

4 Results

  • Targeting central nodes leaves most people susceptible at a 20% vaccination rate, while random vaccination leaves many recovered; central targeting also slows spread and can nearly double transmission time.The time to finish the pandemic can increase even as fewer people become infected.Source: §4 · pp. 6–7, Figure 1 · p. 6
  • 60% random vaccination has the same impact in years of life saved as 30% Degree or Centrality vaccination.Without vaccination, the population loses up to 4.7 years of life; with universal vaccination, the loss is negligible.Source: §4 · p. 6

4.1 The topology of the network

  • On scale-free networks, Peripheral vaccination accelerates the pandemic’s end, whereas on small-world networks it can delay it; vaccinating younger nodes saves fewer years and prevents fewer casualties than other strategies.The Peripheral strategy targets people likely to be infected last; younger nodes have low lethality in the model.Source: §4 · p. 7, §4.1 · p. 7
  • Across both network types, Degree and Centrality strategies reduce years of life lost, but their effectiveness is slightly lower on small-world networks.Removing central nodes can nearly double the time for infection to propagate among those who still become infected.Source: §4.1 · p. 7

4.2 Anti-vaccine views

  • Anti-vaccine diffusion depends nonlinearly on persuasiveness: medium values expose many people, while high values allow the views to percolate through the network.Diffusion stops quickly at small or large persuasiveness, and takes around 10% of the virus-dynamics steps at medium values.Source: §4.2 · pp. 7, 10
  • For small positive persuasiveness, the most central nodes are two to three times more likely to share anti-vaccine views than the least central nodes, assuming equal adoption likelihood after exposure.Central nodes are more exposed, particularly in networks with hubs.Source: §4.2 · p. 8
  • Even at small persuasiveness, most people can encounter anti-vaccine views without adopting or passing them on; at θ = 0.3, about 70% have contact on a scale-free network.The views may spread through multiple narratives with similar dynamics.Source: §4.2 · pp. 8–9, Figure 3 · p. 9
  • With 30% Degree vaccination, years lost rise from 0.3 to nearly two when 20% of people support anti-vaccination views; this analysis assumes the opinion and virus networks are the same.The authors note that opinions can also spread online or through other channels that do not require physical contact.Source: §4.2 · p. 10

5 Discussion

The discussion favors targeted vaccination of highly connected people for its social benefit and cautions that anti-vaccination views may undermine that benefit. Its conclusions are qualitative, and the opinion-diffusion analysis assumes the same network for opinions and virus transmission.

5 Discussion

  • The authors' simulations provide a qualitative account of pandemic evolution, not estimates of COVID-19 years of life lost at different vaccination rates.Such estimates would require considering comorbidities, access to health services, and other factors.Source: §5 · p. 10
  • Targeting vaccination to highly exposed people with many contacts offers the highest social gain because vaccinated people also stop spreading the virus to others.The discussion prioritizes healthcare staff and other high-contact, highly central professions early in vaccination.Source: §5 · p. 10
  • Measuring burden in years of life lost shows that prioritizing elderly people may reduce casualties without reducing years of life lost compared with random vaccination.The authors use this measure to put the impact of elderly-first targeting in perspective.Source: §5 · p. 10

5.1 Anti-vaccine views will be highly relevant

  • Anti-vaccination narratives may attract undecided people with genuine vaccine concerns, while vaccine refusal could substantially increase the time to reach herd immunity.The discussion calls for clear, concise, evidence-based public communication if a vaccine becomes available.Source: §5.1 · p. 10
  • Under the assumed shared opinion-and-virus network, higher-centrality individuals are more likely to share anti-vaccination views, which is detrimental to optimal vaccination strategies.Central nodes play a relevant role in both anti-vaccination-view diffusion and pandemic evolution.Source: §5.1 · pp. 10–11
  • The authors propose targeting central nodes with evidence-based communication as a potential immunisation strategy against misinformation.Influential social-media figures could promote evidence-based views or act as misinformation hubs.Source: §5.1 · p. 11
Loading 2009.13674v1…