Modelling the impact of vaccine hesitancy in prolonging the need for Non-Pharmaceutical Interventions to control the COVID-19 pandemic.

Background: Vaccine hesitancy - a delay in acceptance or refusal of vaccines despite availability - has the potential to threaten the successful roll-out of SARS-CoV-2 vaccines globally. In this study, we aim to understand the likely impact of vaccine hesitancy on the control of the COVID-19 pandemic.
Methods: We modelled the potential impact of vaccine hesitancy on the control of the pandemic and the relaxation of non-pharmaceutical interventions (NPIs) by combining an epidemiological model of SARS-CoV-2 transmission with data on vaccine hesitancy from population surveys.
Results: Our simulations suggest that the mortality over a 2-year period could be up to 7.6 times higher in countries with high vaccine hesitancy compared to an ideal vaccination uptake if NPIs are relaxed. Alternatively, high vaccine hesitancy could prolong the need for NPIs to remain in place. Conclusions: While vaccination is an individual choice, vaccine-hesitant individuals have a substantial impact on the pandemic trajectory, which may challenge current efforts to control COVID-19. In order to prevent such outcomes, addressing vaccine hesitancy with behavioural interventions is an important priority in the control of the COVID-19 pandemic.

Introduction

The COVID-19 pandemic has simultaneously resulted in high global mortality and major economic disruptions. As a control measure, non-pharmaceutical interventions (NPIs) such as social distancing and mobility restrictions have been put in place worldwide and have successfully reduced transmission of the virus. However, these interventions are unsustainable in the long-term[1] and current hopes to control the pandemic rely heavily on vaccination.

In December 2020, the first vaccine against SARS-CoV-2 was approved; by May 2021, 14 vaccines had been licensed (https://vac-lshtm.shinyapps.io/ncov_vaccine_landscape/) and more than 1.3 billion vaccination doses administered worldwide (https://ourworldindata.org/covid-vaccinations#). Their reported efficacy against symptomatic disease ranges from 50% to over 95%[2-6]. Given the high basic reproduction number for SARS-CoV-2 (estimates range between 3–4)[1] high levels of vaccine uptake will be required to achieve herd immunity[7], particularly if children are not vaccinated during the first phase of roll-out.

One major concern that threatens to limit the impact of vaccination is vaccine hesitancy[8]. Population surveys have found that between 14%[9] and 27%[10] of adults say that they will not accept a vaccine if available, whilst between 14%[9] and 19%[10] say that they are uncertain. There is a large variation in vaccine hesitancy between countries, with the proportion saying that they would get a SARS-Cov-2 vaccine if it became available, ranging from 40% for France[10] to 89% for China[9]. In many countries, vaccine hesitancy is heterogenous across sub-populations depending on gender, age, ethnicity, religion, or socioeconomic status[9-11]. Surveys have highlighted the key drivers of SARS-CoV-2 vaccine hesitancy are related to concerns about the accelerated pace of vaccine development[11], side-effects[10] and the spread of misinformation about the pandemic[8]. Underlying reasons of vaccine hesitancy are a complex interaction between trust in government and health authorities[9] coupled with new information —and misinformation— on the vaccine safety and disease risk arising everyday[12].

In the present study, we aim to understand the likely impact of vaccine hesitancy on future control of the pandemic, using a mathematical model of SARS-CoV-2 transmission[7] to explore vaccine hesitancy through its impact on population coverage. We capture the effect of reduced coverage using measured levels of vaccine hesitancy from behavioural survey data[10] on self-reported intention to be vaccinated. Survey results are disaggregated by age and translated to vaccination coverage ranges per age group. Pandemic trajectories with low vaccination coverage due to vaccine hesitancy are compared to an ideal counterfactual assuming no vaccine hesitancy, in which we assume that a small proportion (5%) of the population cannot be reached for vaccination. This value is based on maximum vaccination uptake reported for England’s current COVID-19 vaccine rollout (https://www.england.nhs.uk/statistics/statistical-work-areas/covid-19-vaccinations/). We model each scenario with both a high and a moderate vaccine efficacy profile that represents the range of efficacies of currently approved vaccines. Informed by current vaccine roll-out in high-income countries, we assume that vaccination started in January 2021 and is implemented at a rate that results in a total campaign of 10 months to fully vaccinate the population above 15 years old.

Our simulations suggest that mortality could be higher in countries with high vaccine hesitancy compared to an ideal vaccination and this could prolong the need for NPIs to remain in place. We show that to reduce this impact, vaccination campaigns could include less vulnerable groups, like children. Vaccine hesitancy is an important public health priority that needs to be addressed in order to control the current pandemic.

Methods

Vaccine hesitancy data

Attitudes towards COVID-19 vaccination were obtained from the Imperial College London YouGov Covid 19 Behaviour Tracker Data[10]. This data set includes weekly surveys about people’s behaviours in response to COVID-19 (including vaccines) as well as standard demographic questions on age, gender and household structure. Ethics approval and informed consent were not required given that all data was publicly available and de-identified.

We extracted the survey results from 8th to 15th February 2021 for 10 European countries. To assess vaccine hesitancy, we used data from one question pertaining to COVID-19 vaccine acceptance in which participants were asked to what extent they would definitely get a COVID-19 vaccine, if it became available to them next week. Answers were obtained on a numeric scale ranging from “Strongly agree – 1” to “Strongly disagree – 5”. To capture survey uncertainty, answers per age group were used to parameterise a multinomial distribution, from which we drew 100 replicates. To capture further uncertainty associated with the translation of survey response to vaccine uptake, for each replicate, coverage per age group was estimated assuming the probability of vaccination as a beta distribution with means: 0.98, 0.75, 0.50, 0.25 and 0.02 for survey responses 1, 2, 3, 4 and 5, respectively. Coverage distributions per age group, median as well as the 10% and 90% quantiles are shown in Table S5 and Fig. S2.

Mathematical model

We used a previously developed mathematical model for SARS-CoV-2 transmission and vaccination[7] (Fig. S1). The age-structured deterministic SEIR-type compartmental model incorporates an age-specific probability of infection determined by age-based contact matrices. Susceptible individuals become infected at a rate that depends on the level of infection in the community. Following infection, cases proceed to mild infection or a clinical disease pathway, which includes hospitalisation, oxygen support and intensive care. Waning immunity is captured by recovered individuals returning to the susceptible compartment following an erlang distribution.

Vaccination is modelled as an additional dimension disaggregating the population into those who have not received the vaccine (v0), those who have received the vaccine but are not yet protected (this stage represents the two-dose vaccine schedule and the need to wait ~28 days from dose 1 for protection to develop) (v1 and v2), those who have received the vaccine and are protected (v3 and v4) and those who have received the vaccine but are no-longer protected (v5) (if vaccine-derived immunity is not life-long). In this model, only those who are currently infected do not receive the vaccine. Protection due to vaccination is modelled at two stages in the model; (1) reducing the probability of infection upon exposure (efficacy against infection) and (2) reducing the probability of hospitalisation being indicated after developing disease (efficacy against hospitalisation and death).

Parameters

Parameters for SARS-CoV-2 infection, health care capacity, age-distribution and contact patterns are based on previous work[7,13] (Tables S1 and S4). Given these parameters, transmission probability is estimated based on reproductive number (Rt), which is given as an input for each simulation as a function of time. Vaccine-induced immunity was assumed lifelong, while natural immunity was assumed to last for an average of one year[14]. To produce simulations representing the different vaccines approved to date, each scenario was run for two vaccines: one with high efficacy (94% efficacy against infection[2]) and one with moderate efficacy (63% efficacy against infection[3]). For both vaccines we assume an additional 60% efficacy against hospitalisation for breakthrough infections, resulting in an overall vaccine efficacy against hospitalisation and death of 98% for the high efficacy vaccine and 85% for the moderate efficacy vaccine. A summary of key parameters is given in Tables S1–S6. The model code is freely available at https://github.com/mrc-ide/nimue[15].

To mimic current vaccine rollout plans, vaccination is introduced in the population at the beginning of January 2021. We assumed a constant vaccination rate (κ), at which all individuals aged 15 years and above (~78% of the population) will be vaccinated over a 10-month period. This rate is implemented for all scenarios modelled, since we assume vaccination rate is constrained not by vaccine uptake but by the supply and delivery of vaccines. Therefore, lower levels of coverage, result in shorter vaccination campaigns; given that in the model, once coverage targets are met, vaccination is ceased. To illustrate the effect including children vaccination, vaccination rate was maintained constant and vaccination period was extended such that all individuals age 5–15 years could be vaccinated.

Vaccines are targeted by age groups at the constant rate κ, prioritising older age groups: with 80+ years vaccinated first and then sequentially including additional age groups in 5-year age-bands down to 15–19 years for adults only vaccination simulations and down to 5–10 years for simulations including children vaccination.

Reproductive number profiles

To simulate a representative pre-vaccination scenario, we generated a reproductive number profile in which Rt was the same as R0 (R0 = 3[13]) up to April 2020, subsequently decreased to 1 to represent the impact of NPIs against the first wave, and then rose to 1.5 during the latter half of 2020 to represent a second wave. Following the introduction of vaccination in January 2021, we set Rt to increase in 10 fixed steps. Each step representing the lifting of NPIs. The time for each step increase was determined by estimating when vaccination coverage had reached levels such that the herd immunity threshold due to vaccine immunity was reached. At the end of the vaccination period, Rt remained at a value such that the herd immunity threshold was maintained, given final vaccination coverage and vaccine efficacy against infection.

To estimate the coverage needed for each Rt step, the following herd immunity threshold equation was used:When analysing the impact of lifting NPIs, the Rt profile following the introduction of vaccination was generated based on an ideal scenario for vaccination uptake. Conversely, when evaluating the degree to which NPIs would need to remain in place, the Rt profile after the introduction of vaccination was set up based on vaccine coverage due to vaccine hesitancy.

Scenarios

We consider two potential scenarios for vaccine coverage target per age group: an ideal scenario 2020 with 20 cases. A simulation was run for each vaccine coverage scenario for both adult-only vaccination campaign and vaccination campaign including children. As an output for each simulation, we estimated the number of deaths and hospitalisations associated with COVID-19 over the 2-year period from 1 January 2021 to 31 December 2022.

To generate country-specific simulations, we parameterise the model with data on the population size and age distribution of the country (https://population.un.org/wpp/) and representative contact matrices obtained from a systematic review of social contact surveys through the socialmixR package (https://github.com/sbfnk/socialmixr). The model was then fitted to reported daily cases and deaths up to 31 December 2020 by varying three parameters - the start date of the epidemic, the initial R0 and the effect size of changes in mobility on transmission (using mobility data from Google (https://www.google.com/covid19/mobility)). Model fitting was performed using a Metropolis Hastings MCMC based sampling scheme as previously described[16]. The resulting fit generates a fitted R0 as baseline, an Rt trajectory up to the introduction vaccination in January 2021, after which, Rt was set to increase by 10 fixed steps, up to the theoretical herd immunity threshold based on an ideal vaccination schedule (as described above). The pandemic trajectory was evaluated using country-specific data on vaccine hesitancy and demography for the two coverage scenarios described above and assuming vaccination for individuals aged 15 years and above only.