Abstract
Background Results of phase III vaccine clinical trials against COVID-19, although encouraging and well above initial expectations, have only reported on efficacy against disease and its severity. We evaluated the impact of vaccination on COVID-19 outbreak and disease outcomes in Ontario, Canada.
Methods We used an agent-based transmission model and parameterized it with COVID-19 characteristics, demographics of Ontario, and age-specific clinical outcomes derived from outbreak data. We implemented a two-dose vaccination program, prioritizing healthcare workers and high-risk individuals, with 40% vaccine coverage and vaccine efficacy of 95% against disease. Vaccines were distributed at a rate of 30 doses per day per 10,000 population with a 6-day schedule per week. We projected the impact of vaccination on attack rates, hospitalizations, and deaths. For scenario analyses, we varied the vaccine efficacy against infection, under the assumption of 5% pre-existing population immunity.
Results With no protection against infection, a two-dose vaccination campaign with a time interval of 21 days between doses reduced attack rate, hospitalizations, and deaths by 44.6% (95% CrI: 34.5% - 54.3%), 63.4% (95% CrI: 56.1% - 69.9%), and 70.0% (95% CrI: 62.6% - 75.8%), respectively. These reductions were improved with increased vaccine efficacy against infection, with similar estimated ranges in the corresponding scenarios with a 28-day time interval between vaccine doses.
Conclusions Vaccination can substantially mitigate ongoing COVID-19 outbreaks, even when vaccines offer limited protection against infection. This impact is founded upon a relatively strong vaccine efficacy against disease and severe outcomes.
Introduction
Despite unprecedented public health measures, such as stay-at-home orders, school closures, and physical distancing (1–4), the novel coronavirus disease 2019 (COVID-19) continues to have severe global health and economic consequences (5,6). In Canada, these measures have been comparatively effective in flattening initial outbreaks (7,8). However, the devastating outcomes of the second wave and the high level of susceptibility to COVID-19 (8) have underscored the need for a safe and effective vaccine to control ongoing outbreaks.
The target product profile (TPP) by the World Health Organization (WHO) provided a roadmap for potential COVID-19 vaccine candidates (9). The TPP indicated a preference for candidates that demonstrate a population-based efficacy of at least 70%, and a point estimate of 50%, against transmission and/or severe disease outcomes. Results from phase III vaccine clinical trials have been encouraging thus far, with Pfizer-BioNTech and Moderna reporting an efficacy of over 90% against symptomatic disease (10,11), exceeding the TTP target range. As of December 9, 2020, Health Canada has authorized the Pfizer-BioNTech vaccine for distribution (12), with 3 other candidates under review (13). Therefore, there is an urgent need to understand the potential population-level impact of vaccination campaigns with vaccine prioritization (14).
We sought to evaluate the impact of a COVID-19 vaccination campaign, based on a scenario with two doses distributed either 21 days or 28 days apart, on attack rate and adverse clinical outcomes in Ontario. We extended an agent-based model of disease transmission (15) to include vaccination with an age-specific uptake distribution similar to that of past seasonal influenza epidemics and the 2009 H1N1 pandemic (16,17). We evaluated a roll-out strategy that prioritizes higher risk adults (i.e., healthcare workers, elderly, and comorbid individuals), followed by the general population, to minimize transmission and severe outcomes (14). Our results indicate that vaccination, even with a vaccine that offers limited protection against infection, could have a large impact on reducing hospitalizations and deaths in Ontario.
Methods
Model structure
We extended a previously established agent-based COVID-19 transmission model (15) and included vaccination to simulate outbreak scenarios. The natural history of COVID-19 was implemented in the model by considering individual status as susceptible; latently infected (not yet infectious); asymptomatic (infected and infectious but with no symptoms); pre-symptomatic (infected, infectious and in the stage before symptomatic illness); symptomatic with either mild or severe/critical illness; recovered (and not infectious); and dead (Appendix, Figure A1). We binned the model population into five age groups of 0-4, 5-19, 20-49, 50-64, and 65+ years old based on the demographics of Ontario, Canada (18), and parameterized the model with estimates of the proportion of the population with comorbidities associated with severe COVID-19 (Appendix Table A1) (19,20). Interactions between individuals were informed using an empirically determined contact network (21). The daily number of contacts for each individual was sampled from a negative-binomial distribution with age-dependent mean and standard deviation (Appendix, Tables A2).
Disease dynamics
We implemented disease transmission in a probabilistic manner whereby susceptible individuals were exposed to infectious individuals (i.e., asymptomatic, pre-symptomatic, or symptomatic stages of the disease). Infected individuals entered a latent period as part of an average incubation period of 5.2 days (22). For those who went on to develop symptomatic disease, the incubation period included a pre-symptomatic stage prior to the onset of symptoms (23). The duration of the pre-symptomatic stage was sampled from a Gamma distribution with a mean of 2.3 days (23). The infectious period post-symptom onset was sampled from a Gamma distribution with an average of 3.2 days (24). We considered an age-dependent probability of developing mild, severe, or critical illness after symptom onset. Infected individuals who did not develop symptoms remained asymptomatic after the latent period until recovery. Asymptomatic individuals were infectious for an average of 5 days, which was sampled from a Gamma distribution (24,25).
Based on the number of secondary cases generated during each stage of the disease (26), we parameterized the infectivity of asymptomatic, mild symptomatic, and severe symptomatic stages to be 11%, 44%, and 89% relative to the pre-symptomatic stage (27). We assumed that recovered individuals could not be re-infected during the same outbreak scenario.
Infection outcomes
In our model, mild symptomatic cases recovered without the need for hospitalization. Persons with severe illness or illness requiring critical care used hospital beds in this model. We parameterized the model for the use of intensive care unit (ICU) and non-ICU beds based on recent COVID-19 hospitalization data stratified by the presence of comorbidities in Ontario (20). We assumed that all symptomatic cases who were not hospitalized self-isolated during the symptomatic period. For self-isolated cases, daily contacts were sampled from an age-dependent contact matrix derived from a representative sample population during COVID-19 lockdown (28). The time from symptom onset to hospital admission was uniformly sampled in the range of 2 to 5 days (15,29). The lengths of non-ICU and ICU stays were sampled from Gamma distributions with means of 12.4 and 14.4 days, respectively (30,31).
Vaccination
We implemented a two-dose vaccination campaign, achieving 40% vaccine coverage of the population over the course of the campaign. Vaccine distribution mirrored that of the age-specific vaccine coverage estimates for seasonal and 2009 pandemic influenza (16). We reviewed vaccine program distribution rates for influenza to determine the roll-out capacity (17,32–34). With the expected shortage of vaccines in the initial roll-out, we assumed that 30 individuals per 10,000 population are vaccinated per day in Ontario. A 40% vaccine coverage was achieved within 40 weeks with a 6-day immunization schedule per week.
Vaccination was sequential with prioritization of: (i) healthcare workers, individuals with comorbidities, and those aged 65 and older (i.e., protection cohort); followed by (ii) individuals aged 18-64 (i.e., disruption minimization cohort) (14). Children under 18 years of age were not vaccinated. The vaccine coverage in each cohort (Appendix: Table A3) followed the uptake for different age groups reported in Canadian influenza vaccination coverage studies (16,35). Pre-existing immunity as a result of a primary infection with COVID-19 was not a factor in the vaccination of individuals.
Timelines of vaccination and vaccine efficacy
We considered a vaccination campaign with two doses distributed either 21 days or 28 days apart. Based on results of clinical trials (10,11), we considered vaccine efficacies of 52% and 95% against symptomatic disease, Ve(d), after the first and second doses, respectively (36). If individuals developed COVID-19, we also implemented an equivalent reduction in the likelihood of developing severe disease. We implemented a 14-day interval after the first dose to reach half of Ve(d) and a 7-day interval after the second dose to reach Ve(d). In the absence of data for the protection efficacy against infection, Ve(i), we varied this parameter and considered scenarios in which (i) the vaccine did not protect against infection (i.e.; Ve(i)=0); (ii) vaccine efficacy against infection was 50% lower than the protection against disease (i.e.; Ve(i)=0.5Ve(d)); and (iii) vaccine efficacy against infection was equal to the protection against disease (i.e.; Ve(i)=Ve(d)) after each dose of the vaccine.
The vaccine-induced protection was implemented as a reduction factor in disease transmission, likelihood of developing symptoms, and severity of disease when a vaccinated individual encountered an infectious person and if infection occurred. For individuals with comorbidities or those aged 65 and older, we assumed a reduction in vaccine protection efficacy: Vp = (1-q)Ve, where q was sampled uniformly in the range of 10-50% for each vaccinated individual. This parameterization was based on estimated reductions in influenza vaccine effectiveness in comorbid individuals and the elderly (37).
Model scenarios
We assumed a 5% level of pre-existing immunity accrued prior to vaccination based on recent seroprevalence studies (8). To distribute this level of immunity, we first simulated the model in the absence of vaccination and derived the infection rates in different age groups when the overall attack rate reached 5% of the population. We then used the age-specific attack rates to initialize the proportion of the population with immunity for the vaccination model.
Model implementation
We calibrated the transmission probability per contact to the effective reproduction number Re=1.2 to account for the effect of current public health measures in Ontario, Canada (38,39). With parameters provided in Table 1, each simulation was seeded with one initial case in the latent stage in a population of 10,000 individuals, and the results for incidence, hospitalizations, and deaths were averaged over 1000 independent Monte-Carlo realizations. The model was coded in Julia language and is available at: https://github.com/thomasvilches/vac_covid_ontario.
Description of model parameters and their estimates.
Results
Without vaccination, we projected an attack rate of 5.5% (95% CrI:4.8% - 6.2%) under the current mitigation measures. The average number of hospital admissions was 12.2 (95% CrI: 10.9 - 13.7) per 10,000 population. We projected a cumulative 2.0 (95% CrI: 1.8 - 2.3) deaths per 10,000 over the course of the outbreak.
Vaccination campaign
We assumed that 0.3% of the population was vaccinated per day under the existing distribution capacity. Vaccination substantially reduced the magnitude and duration of outbreaks in all scenarios of vaccine efficacy against infection (Figure 1). For a vaccination campaign with a 21-day interval between the first and second doses, we projected the relative reduction of attack rate, compared to no vaccination, to range from 44.6% (95% CrI: 34.5% - 54.3%) when the vaccine offered no protection against infection to 61.8 (95% CrI: 54.4% - 68.7%) when vaccine efficacy against infection was equal to its efficacy against disease (Figure 2A).
Projected incidence of infection in vaccination scenarios, compared to no vaccination, with a vaccine roll-out of 30 vaccine doses per 10,000 population per day. The second dose of vaccine was offered 21 (A) and 28 (B) days after the first dose. Vaccine efficacy against infection was assumed to be 50% lower than (orange) or the same as (blue) vaccine efficacy against COVID-19 disease.
Projected relative reduction of attack rate, hospitalizations, and deaths in vaccination scenarios with a vaccine roll-out of 30 vaccine doses per 10,000 population per day. The second dose of vaccine was offered 21 (A) and 28 (B) days after the first dose. Vaccine efficacy against infection was assumed to be 50% lower than (orange) or the same as (blue) vaccine efficacy against COVID-19 disease. Boxplots represent the mean and range of reduction in attack rate, hospitalizations, and deaths attributable to vaccination. Circles indicate the median of the simulated data, and lines are the extended range from the minimum (25th percentile – 1.5 IQR) to maximum (75th percentile + 1.5 IQR).
Vaccination was more effective in reducing adverse outcomes. When the vaccine only offered protection against disease but not infection, vaccination reduced hospitalizations and deaths by 63.4% (95% CrI: 56.1% - 69.9%) and 70.0% (95% CrI: 62.6% - 75.8%), respectively, compared to no vaccination. These relative reductions increased to 71.4% (95% CrI: 65.9% - 76.7%) and 77.0% (95% CrI: 71.8% - 81.6%) for hospitalizations and deaths, respectively, when vaccine efficacy against infection was assumed to be the same as vaccine efficacy against symptomatic disease. In all scenarios, the highest benefit of vaccination was achieved in reducing deaths (Figure 2). We observed a similar range of reduction in attack rates, hospitalizations and deaths in the corresponding scenarios with a 28-day time interval between vaccine doses (Figure 2B).
Hospital bed utilization
In the absence of vaccination, we projected that an outbreak would require an average of 143.6 (95% CrI: 126.3 - 163.6) hospital bed-days per 10,000 population (Figure 3). Vaccination significantly reduced hospital bed-days to 52.6 (95% CrI: 44.2 - 60.2), 46.9 (95% CrI: 40.4 - 54.4), and 40.6 (95% CrI: 35.0 - 46.6) per 10,000 population, when vaccine efficacy against infection was: 0%, 50% lower than its efficacy against disease, and equal to its protection against disease, respectively. The reduction of hospital bed-days were similar in the corresponding scenarios when the time interval between vaccine doses was 28 days (Figure 3).
Distribution of the mean cumulative number of hospital bed-days per 10,000 population during the outbreak in vaccination scenarios with a vaccine roll-out of 30 vaccine doses per 10,000 population per day, compared to no vaccination. Boxes indicate interquartile range (IQR), and horizontal lines are the extended range from minimum (25th percentile – 1.5 IQR) to maximum (75th percentile + 1.5 IQR).
Discussion
Results of phase III clinical trials, although encouraging and well above initial expectations, have only reported on vaccine efficacy against disease and its severity. Currently, there is no data on the efficacy of vaccines against infection. We therefore evaluated vaccination strategies by varying the vaccine efficacy in preventing infection. Our results show that the impact of a COVID-19 vaccine on reducing incidence and outcomes can be substantial, reducing hospitalizations and deaths by over 60%, even with a vaccine which offers no protection against infection. This impact is founded upon a relatively strong vaccine efficacy against disease and severe outcomes, which masks the potential consequences of a leaky vaccine with imperfect protection against infection.
Our estimates rely on achieving a vaccine coverage of 40% over the course of 40 weeks and vaccination of a relatively large proportion of high-risk individuals. This coverage and timeline, however, is contingent upon the manufacturing capacity, storage, and distribution of vaccines. Moreover, despite proven safety and efficacy of current vaccines in phase III clinical trials, vaccine hesitancy and concerns about the novelty of these platforms, as well as the speed of vaccine development may result in a lower uptake than the coverages assumed in our model. Therefore, a highly efficient national adverse event reporting system will be critical to supporting public confidence and improving vaccine uptake.
Our results should be interpreted within the context of study limitations. We assumed that the efficacy of a vaccine would be reduced by 10-50% in comorbid and elderly individuals, based on estimated reductions for influenza vaccine efficacy (37). If vaccination performed equally well in comorbid individuals, we would expect a higher reduction for adverse clinical outcomes. We did not consider drop out over the course of the vaccination program, which would affect the level of protection conferred by vaccination. Finally, we did not consider children under 18 years of age for vaccination in these scenarios. Phase III trials are mainly performed in adults and it may be several months before trial results in children are available to support vaccination programs in this population.
This study indicates that, while vaccination could significantly mitigate the severity of outbreaks, it is unlikely to eliminate the need for other non-pharmaceutical interventions before reaching a sufficiently high level of population immunity. Nevertheless, our results show that vaccination is a key public health measure in the fight against COVID-19.
Data Availability
The model was coded in Julia language and is available at: https://github.com/thomasvilches/vac_covid_ontario.
Competing Interest
JML’s institution has received funding for research studies from Sanofi Pasteur, GlaxoSmithKline, Merck, Janssen and Pfizer. JML holds the CIHR-GSK Chair in Paediatric Vaccinology. Other authors declare no competing interests.
Funding
SMM gratefully acknowledges the Canadian Institutes of Health Research OV4 – 170643, COVID-19 Rapid Research; and Natural Sciences and Engineering Research Council of Canada. TNV acknowledges the support of the São Paulo Research Foundation, grant #2018/24811-1.
References
