Professional Documents
Culture Documents
Modelling of COVID-19 Vaccination Strategies and Herd Immunity
Modelling of COVID-19 Vaccination Strategies and Herd Immunity
Vaccine
journal homepage: www.elsevier.com/locate/vaccine
a r t i c l e i n f o a b s t r a c t
Article history: Several vaccines for SARS-CoV-2 are expected to be available in Australia in 2021. Initial supply is limited
Available online xxxx and will require a judicious vaccination strategy until supply is unrestricted. If vaccines have efficacy as
post-exposure prophylaxis (PEP) in contacts, this provides more policy options. We used a deterministic
Keywords: mathematical model of epidemic response with limited supply (age-targeted or ring vaccination) and
COVID-19 mass vaccination for the State of New South Wales (NSW) in Australia. For targeted vaccination, the effec-
SARS-CoV-2 tiveness of vaccinating health workers, young people and older adults was compared. For mass vaccina-
Vaccines
tion, we tested varying vaccine efficacy (VE) and distribution capacities. With a limited vaccine stockpile
Epidemic
Australia
enough for 1 million people in NSW, if there is efficacy as PEP, the most efficient way to control COVID-19
will be ring vaccination, however at least 90% of contacts per case needs to be traced and vaccinated.
Health worker vaccination is required for health system resilience. Age based strategies with restricted
doses make minimal impact on the epidemic, but vaccinating older people prevents more deaths. Herd
immunity can only be achieved with mass vaccination. With 90% VE against all infection, herd immunity
can be achieved by vaccinating 66% of the population. A vaccine with less than 70% VE cannot achieve
herd immunity and will result in ongoing risk of outbreaks. For mass vaccination, distributing at least
60,000 doses per day is required to achieve control. Slower rates of vaccination will result in the popu-
lation living with COVID-19 longer, and higher cases and deaths.
Ó 2021 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND
license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
https://doi.org/10.1016/j.vaccine.2021.04.042
0264-410X/Ó 2021 The Author(s). Published by Elsevier Ltd.
This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
Please cite this article as: C. Raina MacIntyre, V. Costantino and M. Trent, Modelling of COVID-19 vaccination strategies and herd immunity, in scenarios of
limited and full vaccine supply in NSW, Australia, Vaccine, https://doi.org/10.1016/j.vaccine.2021.04.042
C. Raina MacIntyre, V. Costantino and M. Trent Vaccine xxx (xxxx) xxx
other frontline responders at high risk of disease transmission, and eral possible vaccination strategies under limited or unlimited
sociodemographic groups at significantly higher risk of severe dis- supply assumptions, as described in more detail below. We used
ease, such as older adults or people with high risk chronic health the NSW population and age distribution from 2020 [24], stratified
conditions [13,14]. However, prioritizing young people may impact into 16 five-year age groups from 0 to 74 years old and the last age
transmission more, since vaccines are more effective in younger group comprising people 75 years and over.
people, and transmission is highest in young adults [15]. Alterna- The model moves the population in 17 mutually exclusive com-
tively, a ring vaccination strategy could be utilized, which involves partments, each divided by the 17 age groups: susceptible (S),
identifying the close contacts of a confirmed case and vaccinating latent, not infectious (E), pre-symptomatic infectious and diag-
them. This strategy was used effectively against Ebola, and also nosed (Et), pre-symptomatic infectious and undiagnosed (Eu), con-
smallpox in settings where mass vaccination was not possible, tacts traced (C) symptomatic, high infectiousness, diagnosed (I1)
despite efficacy being half that of primary prevention [16–19]. and undiagnosed (I2), asymptomatic, high infectiousness (A1),
Many vaccines including measles, hepatitis A and smallpox, are symptomatic, lower infectiousness, previous diagnosed (I11) and
effective as post-exposure prophylaxis (PEP) and can be given to undiagnosed (I22), asymptomatic low infectiousness (A2), home
contacts during an outbreak, albeit with lower efficacy than pri- isolated (Q), hospitalized (H), individuals admitted to intensive
mary prevention [20]. Whether COVID-19 vaccines will be effective care (ICU) , recovered (R), successfully vaccinated (V) and COVID-
as PEP is unknown as yet, but may well be given the long incuba- 19 deceased people (D) (Fig. 1). The duration of stay in each com-
tion period [21]. partment determines the rate at which people move from one
In Australia, the stated priority groups for early COVID-19 vac- compartment to another for the epidemiological disease states.
cination are: 1) individuals with increased risk of severe disease, The duration of the latent period is assumed to be 5.2 days [25],
such as older adults, Aboriginal and Torres Strait Islander people, of which the last two days before symptoms onset are considered
and those with high risk chronic conditions; 2) individuals at infectious [26].
higher risk of disease, such as health and aged care workers; and The transmission was distributed over the infectious period
3) individuals working in critical services [22]. Vaccination is being with 44% of transmissions occurring in the last two days of the
rolled out to frontline health workers, aged care residents and pre-symptomatic state. We assume that the viral load is very high
hotel quarantine workers first. The aim of this paper was to model on the first day of symptoms and then decreases to a lower infec-
the impact of various COVID-19 vaccine strategies under a limited tious level, spread over the following 6 days [26], so 36% of trans-
supply scenario, as well as varied vaccine efficacy and speed of vac- missions occur in the first day of symptoms and 20% in the
cination for mass vaccination, on COVID-19 case numbers and following 6 days of symptoms [26]. Of the total people infected,
mortality in New South Wales, the most populous state of 35% are considered asymptomatic and never develop symptoms,
Australia. however we assumed that asymptomatic cases are as equally
infectious as symptomatic ones [27–29]. In conjunction with vac-
cination, other disease control interventions such as case finding,
2. Methods contacts tracing, quarantine, isolation and hospital, ICU treatments
were included, however we did not include additional measures
A previously published age structured deterministic compart- such as universal masking or lockdowns, neither of which are in
mental model for Covid-19 spread [23] was modified to test sev- effect in NSW.
2
C. Raina MacIntyre, V. Costantino and M. Trent Vaccine xxx (xxxx) xxx
The force of infection that moves people from susceptible to Fig. 1 shows model diagrams, and in Table 1 are described all
infected is an age specific rate calculated as a combination of aver- parameters used in the model, while the differential equations
age age specific number of close contacts per day in Australia [30], are described in the supplementary materials.
the proportion of infectious people and the probability of infection To test the most effective vaccination distribution for outbreak
per contact, which is estimated to reproduce an R0 of 2.5 [28]. control, we used a hypothetical epidemic in NSW with 100 symp-
Once infected, a person enters the latent compartment, and after tomatic people and 250 untraced latent infected people as the
3.2 days will become infectious and pre-symptomatic for 2 days, starting conditions, so that we could test the effectiveness of differ-
where if traced will be quarantined with a 50% reduction in trans- ent vaccination strategies in an epidemic scenario.
missions [31] and upon symptoms onset will take 1 day on average The vaccination strategies tested are:
to get isolated with no more transmissions after isolation. How-
ever, if an infected person is not traced in latency, it will take 5 days 1. Limited vaccine supply (2 million doses for 1 million people)
to get isolated following symptom onset [32]. Hospitalizations and given to targeted age groups;
admission to ICU follow age-specific rates [33]. The model assumes 2. Limited vaccine supply, ring vaccination (contact tracing and
that 80% of all close contacts are traced and quarantined for 14 days vaccination of contacts);
and 90% of the symptomatic cases are isolated after 5 days. 3. Unlimited supply and mass vaccination assuming enough doses
There are no data on COVID-19 vaccines as PEP, but early data to vaccinate the entire NSW population.
from phase three trials indicate efficacy in excess of 90% as primary
prevention of mRNA vaccines [8]. The vaccine efficacy (VE) deter- In the limited supply scenario, we assumed a limited stock of
mines the percentage of people fully protected, and all VE esti- only 2 million doses (for 1 million people in a two-dose schedule)
mates in the model are against all infection (symptomatic and for NSW as an estimate of initial supply. For mass vaccination we
asymptomatic). VE against all infection is assumed to prevent assumed a delivery capacity of 50,000–300,000 doses per day
transmission, as infection is prevented altogether. Both the [46]. We explored 3 different scenarios with a limited vaccine
BNT162b2 mRNA and ChAdOx1 nCoV-19 vaccines are used in Aus- supply:
tralia. Vaccine efficacy was set at 90% against all infection in the
base case scenario if a person is susceptible, based on effectiveness Vaccine delivered in 8 days to 1 million young people (age
estimates for the BNT162b2 mRNA vaccine against all infection of group 10–29);
92% from Israel [34]. A sensitivity analysis incorporating lower effi- Vaccine delivered in 8 days (first dose) to 1 million older people
cacy such as with the ChAdOx1 nCoV-19 (AZD1222) was done as (age group 65 + );
described below [35]. Our model assumed people vaccinated as Vaccinate 125,000 HCWs with first dose and over the following
PEP has a 50% reduction in protection from vaccination (from 7 days 875,000 people aged 10–29.
90% to 45%). Successful vaccination is considered to give immunity
to the infection for at least 12 months. Duration of immunity The two dose schedule was assumed to be given with 21–
beyond 12 months is not considered, as long term follow up data 28 days between dose one and two. [10,12] The health care work-
are not available yet to inform the modelling of waning immunity. ers (HCWs) for NSW were estimated from the 2019 estimation
Table 1
Model parameters used and their values.
3
C. Raina MacIntyre, V. Costantino and M. Trent Vaccine xxx (xxxx) xxx
State and Territory Statistics – 2019 [47] and were assumed to be 3 3. Results
times more susceptible to infection with COVID-19 [48]. The num-
ber of medical practitioners, nurses, midwives, and pharmacists in A restricted supply (for 1 million people) will not be enough to
NSW is 135,379 – we assumed 80–90% would be in the clinical control the epidemic, with all scenarios resulting in a large number
workforce and require vaccination (125,000 vaccinations). As their of cases and deaths after 500 days. Fig. 2 shows results for the
average age is respectively 47, 44 and 41 years old [47], we dis- choice of using the available doses in the age group 10–29 years
tributed them through 8 age groups from 25 to 64 years old, fol- old, vaccinating HCWs and then 10–29 years old, or vaccinating
lowing the age distribution of HCWs. the 65 + years old. Targeting the younger age group will have more
For ring vaccination we assumed varying proportions of vacci- impact on reducing the number of cases, whilst vaccinating people
nation of all the traced contacts 70%, 80% and 90%. For use as 65 + will have more impact on deaths, as this age group is at much
PEP, we assumed a 50% reduction in efficacy (from 90% to 45%) higher risk of death.
when given to latent pre-symptomatic people. If enough vaccine is available to vaccinate the entire population
For mass vaccination in NSW, assuming that the vaccine supply of NSW, the important variables influencing the results will be vac-
is available for the population, we tested the effect of speed of cine efficacy and the speed of vaccination. In Fig. 3 we show results
achieving high coverage by comparing the capacity to deliver of varying vaccination capacity from 50,000, to 300,000 people per
50,000, 75,000, 100,000, 125,000, and 300,000 doses per day to day. Vaccinating 50,000 people per day will take about 177 days to
vaccinate the NSW population. We also varied the efficacy of the vaccinate the entire population with one dose, will result in a much
vaccine to determine the minimum vaccine efficacy to achieve longer and larger epidemic, compared to 125,000 or 300,000 peo-
herd immunity, and the epidemic scenario with low and high effi- ple per day. However, if the health system has capacity to vacci-
cacy vaccines. We used the following formula to calculate the nate at least 75,000 people per day the epidemic could be
required vaccine coverage (Vc) for herd immunity assuming terminated rapidly (Fig. 3).
R0 = 2.5. [28] In the base case, vaccine efficacy (Ve) = 0.9 (90%). With a vaccine efficacy (VE) of 90% against all infection, vacci-
nating 66% of the total population (approximately 5.384 million
1 1 people) will provide herd immunity and reduce the R0 to less than
Vc ¼ ð1 Þ 1, while efficacy below 60% cannot achieve herd immunity. A min-
Ve R0
imum VE of 85% is required to achieve herd immunity at 70% pop-
ulation coverage. A VE of 80% will require 75% of the population to
To test the effect of varied efficacy, vaccine coverage was held be vaccinated, and VE of 60% will require 100% of the population to
constant at 70%, reflecting a feasible mass vaccination goal. We be vaccinated to achieve herd immunity. Supplementary Table 1
tested varied vaccine efficacy of 38%, 50%, 60%, 70%, 80%, 90% and shows the required vaccine coverage for varying VE, and that effi-
95%. These estimates were selected based on published upper effi- cacy below 60% cannot achieve herd immunity.
cacy of vaccines of 94–95% [9,10,12], and the remaining values to Fig. 4 shows the effect of varied vaccine efficacy on the epi-
reflect a plausible efficacy range of other potential vaccines. The demic, with vaccine coverage of 70% (5.25 million people) with
lower estimate of 38% is the efficacy against any laboratory con- 125,000 doses delivered per day. When vaccine effectiveness drops
firmed infection (symptomatic or asymptomatic) calculated from to 70% the number of new cases a day (incidence) will not increase
ChAdOx1 nCoV-19 vaccine (AZD1222) published data, excluding however will take up to one year to reduce the number of new
the data from the Brazilian component of the study, because cases per day to less than 50. At VE of 60% or less, transmission
weekly testing for infection was not done in Brazil. [10] and the risk of outbreaks will continue, as the number of new cases
Model outputs are the epidemic curve (cases), total number of keeps increasing for more than 3 months following the start of vac-
cases and deaths and, for ring and mass vaccination, there is the cination campaign.
added output of total number of vaccine doses used and days to In Fig. 5 the results for ring vaccination with varying rates of
vaccinate the entire susceptible population, respectively. contact tracing per day (70%, 80% or 90% of contacts of cases) are
Fig. 2. Targeted vaccination strategies with a restricted supply for 1 million people: From left to right is the epidemic curve, the cumulative case and deaths numbers by
targeted age group vaccinated.
4
C. Raina MacIntyre, V. Costantino and M. Trent Vaccine xxx (xxxx) xxx
Fig. 3. The impact of speed of mass vaccination with a vaccine of 90% efficacy. From left to right is the epidemic curve, the cumulative case and deaths numbers by number of
number of people vaccinated per day in NSW.
Fig. 4. From left to right is the epidemic curve, the cumulative case and deaths by vaccine efficacy (VE), with 125,000 people per day vaccinated until 70% of NSW population
is vaccinated.
Fig. 5. The impact of contact tracing and ring vaccination on epidemic control for a vaccine with 45% efficacy as post-exposure prophylaxis: From left to right is the epidemic
curve, the cumulative case and deaths numbers by percentage of contacts traced and vaccinated.
5
C. Raina MacIntyre, V. Costantino and M. Trent Vaccine xxx (xxxx) xxx
shown. The higher the proportion of contacts traced and vacci- resulted in community transmission, and without herd immunity,
nated from the start, the lower the required vaccine doses. If only this risk will continue and will hamper economic recovery [52].
70% of contacts are traced, the required vaccine doses are much During the period of initial restricted supply, health and aged
higher and the epidemic impact much greater than 80 or 90% care workers should be the highest priority, as protecting health
(Fig. 3). workers and other first responders is essential for health system
Table 2 shows all numerical results for the different scenario resilience in a pandemic. During a large second wave in the state
tested. of Victoria, over 3500 health and aged care workers became
In conclusion, with a limited vaccine stockpile the most efficient infected [53], and we demonstrated ongoing risk to health workers,
way to delivery vaccine doses during an epidemic will be ring vac- even during periods of low community incidence [48]. After vacci-
cination, assuming a 45% efficacy when used as PEP, and at least nating health and aged care workers, the evidence supports vacci-
90% of contacts per case traced and vaccinated. If supply of vaccine nation of older adults. Age is the strongest predictor of morbidity
is adequate to cover the entire NSW population, mass vaccination and mortality, far more than any chronic disease [54], so protecting
with a high efficacy vaccine and rapid uptake (at least 75,000 peo- older adults will prevent the most deaths from COVID-19. Includ-
ple vaccinated per day) will be the only strategy able to achieve ing younger people in the prioritisation group can reduce transmis-
herd immunity and prevent community transmission. sion and the overall infection rates, although this benefit is negated
by the higher mortality rates in older people. Those findings are
supported from a preprint study, which explored the most effective
age-group to prioritize for vaccination in order to reduce morbidity
4. Discussion and mortality from Covid-19 [55].
If COVID-19 vaccines are effective as PEP, this will be the best
We showed that herd immunity through vaccination is the only approach for epidemic control using a limited vaccine supply and
means to prevent sustained community transmission and requires requires capacity for high levels of contact tracing. If community
mass vaccination of 66% of the population vaccinated with a vac- transmission of COVID-19 remains low, complete contact tracing
cine of 90% efficacy against all infection. Vaccines of lower efficacy and ring vaccination will be a highly cost-effective, dose-sparing
require higher vaccination coverage, but a VE below 70% against all strategy. Whilst there are no data yet on the efficacy of COVID-
infection may be inadequate for herd immunity. Mass vaccination 19 vaccines as PEP, we showed that if a 50% reduction of efficacy
with a high efficacy vaccine represents the best strategy for eco- is seen with these vaccines, that ring vaccination can be an effec-
nomic recovery, as herd immunity through effective vaccination tive strategy, and may be a better use of a restricted supply ini-
will provide the best prospect for lifting of societal restrictions. tially, if rates of contact tracing are high. Many vaccines have
Early results from Israel, which used the BNT162b2 mRNA vaccine efficacy as post-exposure prophylaxis in outbreak settings, includ-
are very promising, showing efficacy of 92% against asymptomatic ing measles, hepatitis A, varicella zoster and smallpox, with an effi-
infection, suggesting herd immunity is feasible [34]. Data from cacy of 50–90% of that for primary prevention [20]. This is a
Scotland on use of the ChAdOx1 nCoV-19 show reduction in hospi- research gap for COVID-19 vaccines, which needs to be addressed.
talization, but do not report on asymptomatic infection [49]. Elim- If a high (>80%) percentage of contacts cannot be traced (such as
ination cannot be achieved and sustained without vaccination [50]. when an epidemic is very large and the scale of contact tracing is
We show that for a population of 7.5 million people, if faced unmanageable, as occurred in the second wave in the State of Vic-
with an initial restriction in vaccine supply, sustained epidemic toria [56], mass vaccination becomes more urgent.
control cannot be achieved by vaccination alone, and other non- We showed that for mass vaccination, a slow trickle approach
pharmaceutical interventions will need to continue to mitigate will leave us living with COVID-19 longer with all the associated
the risk of outbreaks [2]. If COVID-19 vaccines have efficacy as societal and economic implications - rapid speed of vaccination
PEP, contact tracing and ring vaccination is the best way to achieve will be the best strategy to reduce morbidity and mortality. The
epidemic control, but the majority of the population will remain logistics of rapid delivery is key, and we show that slower uptake
non-immune and susceptible to outbreaks. Recurrent breaches in will result in more cases and deaths and a longer epidemic risk.
hotel quarantine [51] and associated with cruise ships have Rapid vaccine uptake requires concerted planning for vaccination
infrastructure, including training more accredited vaccinators,
planning mass vaccination clinics, and testing the capacity to deli-
Table 2
ver 300,000 or more vaccine doses a day, which provides the best
Total cases and deaths for targeted and mass vaccination scenarios, and total cases,
deaths and vaccination used for ring vaccination scenarios. outcomes. We assumed in the base case scenario for mass vaccina-
tion that the health system will be able to deliver 125,000 doses
Targeted vaccination: age groups Total cases Total
per day, which is similar to the distribution capacity estimated in
deaths
Victoria of about 117,000 doses per day [44]. We showed that
1) 10–29 2,536,040 56,708
the speed of Covid-19 vaccine uptake is critical to rapid epidemic
2) 65+ 3,415,521 45,379
3) HCWs + 10–29 2,554,442 58,409
control, however the logistics of vaccine distribution is complex
[57], particularly for a country such Australia, where an estimated
Mass vaccination: people vaccinated per day Total cases Total
29% of the population lives in rural and remote areas [58]. There-
(total 5,250,000 doses, 70% coverage) deaths
fore, development of detailed logistical plans and tools to support
50,000 407,198 6,845 increased capacity to vaccinate, as well as effective vaccine trans-
75,000 108,233 1,766
100,000 46,073 752
port, storage, and continuous cold-chain monitoring is required.
125,000 26,197 431 No country has attempted mass vaccination since the eradica-
300,000 5,507 100 tion of smallpox. Currently in Australia, vaccination is delivered
mainly in primary care. Under the National Immunisation Program
Contacts Total doses used Total cases Total
about 5 million doses are delivered across the country each year in
traced/vaccinated deaths
the infant, adolescent and adult vaccination programs. Mass vacci-
70% 2,494,500 1,194,140 15,850
nation requires this to be scaled up by five times. Nurse vaccinators
80% 1,849,300 736,254 8,608
90% 794,720 266,164 2,707 can add to the primary care capacity of general practitioners but
require accredited training to be able to vaccinate. In addition,
6
C. Raina MacIntyre, V. Costantino and M. Trent Vaccine xxx (xxxx) xxx
pharmacists can vaccinate for influenza in Australia and may be an then vaccinating health workers and ring vaccination would be
additional source of accredited vaccinators but giving a new vac- the preferred use of a restricted supply. This remains a gap in
cine with uncertain safety profile for rare adverse events would research, and we recommend that clinical trials urgently be con-
be best done in general practice or in public vaccination clinics ducted to study the efficacy of COVID-19 vaccines as PEP given
staffed by medical and nursing personnel. Modelling of GP capacity to close contacts of cases. When vaccine supply is available for
to rapidly vaccinate 70–80% of the population, and the impact on mass vaccination, the quicker a high uptake can be achieved, the
other medical care, is needed, and logistic support and resourcing greater the impact on epidemic control, morbidity, and mortality,
for GPS to expand their capacity to achieve this will be required. but herd immunity cannot be achieved with vaccines of less than
This includes cold chain infrastructure for vaccines that require 60% efficacy. The VE of the ChAdOx1 nCoV-19 vaccine, the largest
deep freezing [12]. Decisions need to be made on public vaccina- component of the Australian planned stockpile, is less than 40%
tion clinics, staffing, training of more nurse vaccinators and other against any laboratory confirmed infection (symptomatic or
measures to increase vaccination capacity. asymptomatic) based on interim analysis. [10] The use of high effi-
Australia does not have a goal of herd immunity through mass cacy vaccines provides the only prospect of achieving herd immu-
vaccination at this stage, which means uptake will not likely be nity against SARS-COV-2, and such vaccines should be procured if
adequate for herd immunity. A vaccine with 90% VE requires 66% herd immunity is a goal. There are no data on serial dosing with
coverage, which is feasible, even without children being vacci- two different vaccines, which is a possible scenario if Australia
nated. At this stage, because of lack of phase three clinical trial data starts with a lower efficacy vaccine. This is another reason to
in children, initial vaccine roll out will be restricted to adults, who ensure adequate supply of high efficacy vaccines. The logistics of
comprise 80% of the population. This also makes age-based tar- mass vaccination, including cold chain, delivery, capacity to vacci-
geted vaccination only feasible for older adults. However, adult nate and addressing vaccine hesitancy should be a high priority for
vaccination is more challenging than infant vaccination, as adults planning and resourcing.
are a mobile population – for people over 65 years, about 70%
receive influenza vaccine annually. [59] Achieving whole-of-life 5. Data statement**a
vaccination rates above 70% may be challenging, and impossible
while COVID-19 vaccines remain unlicensed in children, who com- This research is a mathematical model. All data used in the
prise 20% of the population. This is another impetus for choosing a model were from published papers listed in the reference list.
vaccine with at least 90% VE, as the required coverage for herd
immunity (66%) is far more feasible, even if children are not Declaration of Competing Interest
vaccinated.
Vaccine hesitancy may also affect the ability to achieve rapid, The authors declare the following financial interests/personal
high uptake and must also be addressed well in advance of vacci- relationships which may be considered as potential competing
nation programs to ensure that programmatic competency to interests: CR MacIntyre has been on an advisory board for COVID-
achieve high uptake rapidly will not be met by unexpected social 19 vaccine for Seqirus and consulted for Astra Zeneca Australia on
or behavioral obstacles. Australia has very high rates of vaccination COVID-19 vaccines. She has been on advisory boards for influenza,
for National Immunization Program vaccines [47], especially with RSV and pneumococcal vaccines in the last 3 years for Seqirus,
the use of financial disincentives, and has a more accepting culture Sanofi, Pfizer and Janssen. The other authors have no conflicts to
of public health interventions, so large scale vaccine refusal is not declare.
expected. We also identified that vaccine hesitancy about COVID-
19 vaccines mostly does not reflect fixed objection to vaccination,
Acknowledgements
and that there is scope and need for health promotion and positive
messaging to ensure high vaccine uptake [60].
Chandini Raina MacIntyre is supported by a NHMRC Principal
The main limitation of this study is the uncertainty of the
Research Fellowship, grant number 1137582.
parameters used and the assumptions. The model assumes that
symptomatic and asymptomatic infected people have the same
infectiousness, based on several published studies that show the Appendix A. Supplementary data
proportion of asymptomatic cases and the same viral load in both
types of infection [27–29]. If symptomatic infection is more trans- Supplementary data to this article can be found online at
missible, our model may overestimate the impact of a COVID-19 https://doi.org/10.1016/j.vaccine.2021.04.042.
outbreak. However, symptomatic people may reduce their contacts
with others, while asymptomatic people will continue with normal References
behaviors, which may result in similar rates of transmission. We
[1] WHO Coronavirus Disease (COVID-19) Dashboard | WHO Coronavirus Disease
used a contact matrix estimated for Australia, based on pre- (COVID-19) Dashboard. Available at: https://covid19.who.int/. (Accessed: 2nd
COVID mixing patterns, however reduced contacts are considered December 2020).
in the model when people are latent or traced. As this study [2] MacIntyre CR. Case isolation, contact tracing, and physical distancing are
pillars of COVID-19 pandemic control, not optional choices. Lancet Infect Dis
focuses on the relative differences between vaccine scenarios in 2020;20:1105–6.
an unmitigated epidemic, we did not consider social distancing [3] Kissler SM, Tedijanto C, Goldstein E, Grad YH, Lipsitch M. Projecting the
or lockdown. Further the exact performance and characteristics transmission dynamics of SARS-CoV-2 through the postpandemic period.
Science (80-.). 2020;368:860–8.
of the different vaccines vary and there no data as yet around vac- [4] Kucharski AJ et al. Effectiveness of isolation, testing, contact tracing, and
cines as PEP. The ability of vaccines to prevent transmission (by physical distancing on reducing transmission of SARS-CoV-2 in different
preventing any infection, symptomatic or asymptomatic) is uncer- settings: a mathematical modelling study. Lancet Infect Dis 2020;20:1151–60.
[5] Brüssow H. Efforts towards a <scp>COVID</scp> -19 vaccine. Environ
tain, but data from Israel, which conducted a rapid mass vaccina-
Microbiol 2020;22:4071–84.
tion campaign using the Pfizer vaccine, do show a reduction in [6] Florian K. SARS-CoV-2 vaccines in development. Nature 2020;586:516–27.
transmission [61]. [7] Draft landscape of COVID-19 candidate vaccines. Available at: https://www.
In conclusion, restricted vaccine supply can protect first respon- who.int/publications/m/item/draft-landscape-of-covid-19-candidate-
vaccines. (Accessed: 2nd December 2020).
ders but will not have an impact on epidemic control, unless the [8] Evan C. What Pfizer’s landmark COVID vaccine results mean for the pandemic.
vaccine has some efficacy as PEP. If vaccines are effective as PEP, Nature 2020.
7
C. Raina MacIntyre, V. Costantino and M. Trent Vaccine xxx (xxxx) xxx
[9] Evan C. COVID vaccine excitement builds as Moderna reports third positive [36] Lin Y, Xie B, et al. C. D. Asymptomatic Novel Coronavirus Pneumonia Patient
result. Nature 2020. Outside Wuhan: The Value of CT Images in the Course of the Disease. Clin
[10] Voysey M. et al. Safety and efficacy of the ChAdOx1 nCoV-19 vaccine Imaging 2020;63:7–9.
(AZD1222) against SARS-CoV-2: an interim analysis of four randomised [37] Rothe M, Sothmann P, et al. C. S. Transmission of 2019-nCoV Infection from an
controlled trials in Brazil, South Africa, and the UK. Lancet (London, England) Asymptomatic Contact in Germany. N Engl J Med. 2020. http://doi.org/
0, (2020). 10.1056/NEJMc2001468.
[11] Normile D. Development of unique Australian COVID-19 vaccine halted. [38] Zou L et al. SARS-CoV-2 Viral Load in Upper Respiratory Specimens of Infected
Science 80- 2020. https://doi.org/10.1126/science.abg1208. Patients. N Engl J Med 2020. https://doi.org/10.1056/NEJMc2001737.
[12] Polack FP et al. Safety and Efficacy of the BNT162b2 mRNA Covid-19 Vaccine. N [39] Du Z et al. Serial Interval of COVID-19 among Publicly Reported Confirmed
Engl J Med 2020;NEJMoa2034577. https://doi.org/10.1056/NEJMoa2034577. Cases. Emerg Infect Dis 2020;26.
[13] WHO SAGE Roadmap For Prioritizing Uses Of COVID-19 Vaccines In The [40] Mizumoto K, Chowell G. Transmission potential of the novel coronavirus
Context Of Limited Supply. Available at: https://www.who.int/publications/m/ (COVID-19) onboard the diamond Princess Cruises Ship, 2020. Infect Dis Model
item/who-sage-roadmap-for-prioritizing-uses-of-covid-19-vaccines-in-the- 2020;5:264–70.
context-of-limited-supply. (Accessed: 2nd December 2020). [41] Estimating the asymptomatic ratio of 2019. . . | Oxford Martin School. Available
[14] Schmidt H, Pathak P, Sönmez T, Ünver MU. Covid-19: How to prioritize worse- at: https://www.oxfordmartin.ox.ac.uk/publications/estimating-the-
off populations in allocating safe and effective vaccines. BMJ 2020;371. asymptomatic-ratio-of-2019-novel-coronavirus-onboard-the-princess-
[15] Giubilini A, Savulescu J, Wilkinson D. COVID-19 vaccine: vaccinate the young cruises-ship/. (Accessed: 24th March 2020).
to protect the old?. J Law Biosci 2020;7:1–13. [42] MacIntyre Valentina, Kunasekaram, Mohana Priya, CRC. Health system
[16] Strassburg MA. The global eradication of smallpox. AJIC Am J Infect Control capacity in Sydney, Australia in the event of a biological attack with
1982;10:53–9. smallpox. PLoS One 2019;14.
[17] MacIntyre CR, Costantino V, Mohanty B, Nand D, Kunasekaran MP, Heslop D. [43] Team, T. N. C. P. E. R. E. Vital Surveillances: The Epidemiological Characteristics
Epidemic size, duration and vaccine stockpiling following a large-scale attack of an Outbreak of 2019 Novel Coronavirus Diseases (COVID-19) — China, 2020.
with smallpox. Global Biosecurity. 2019;1(1):74–81. http://doi.org/10.31646/ China CDC Wkly. 2020;2:113–22.
gbio.13. [44] Garg S et al. Hospitalization Rates and Characteristics of Patients Hospitalized
[18] Kucharski AJ et al. Effectiveness of ring vaccination as control strategy for with Laboratory-Confirmed Coronavirus Disease 2019 — COVID-NET, 14
Ebola virus disease. Emerg Infect Dis 2016;22:105–8. States, March 1–30, 2020. MMWR Morb Mortal Wkly Rep 2020;69:458–64.
[19] Wells C et al. Harnessing Case Isolation and Ring Vaccination to Control Ebola. [45] Richardson S et al. Presenting Characteristics, Comorbidities, and Outcomes
PLoS Negl Trop Dis 2015;9. Among 5700 Patients Hospitalized With COVID-19 in the New York City Area.
[20] Gallagher T, Lipsitch M. Postexposure Effects of Vaccines on Infectious JAMA 2020. https://doi.org/10.1001/jama.2020.6775.
Diseases. Epidemiol Rev 2019;41:13–27. [46] Mass Vaccination Clinics during an Influenza Pandemic Policy manual,
[21] Wassie GT, Azene AG, Bantie GM, Dessie G, Aragaw AM. Incubation Period of Infectious Diseases Population Health-Communicable Diseases. (2018).
Severe Acute Respiratory Syndrome Novel Coronavirus 2 that Causes [47] Health Workforce Data | Summary Statistics. Available at: https://hwd.
Coronavirus Disease 2019: A Systematic Review and Meta-Analysis. Curr health.gov.au/summary.html#part-3. (Accessed: 15th October 2020).
Therapeut Res - Clin Exper 2020;93. [48] Quigley AL, Stone H, Nguyen PY, Chughtai AA, MacIntyre CR. Estimating the
[22] COVID-19 vaccination prioritisation | Australian Government Department of Burden of COVID-19 on the Australian Healthcare Workers and Health System.
Health. Available at: https://www.health.gov.au/news/health-alerts/novel- Int J Nurs Stud 2020;103811. https://doi.org/10.1016/j.ijnurstu.2020.103811.
coronavirus-2019-ncov-health-alert/vaccines-and-treatments/covid-19- [49] Torjesen I. Covid-19: First doses of vaccines in Scotland led to a substantial fall
vaccination-prioritisation. (Accessed: 2nd December 2020). in hospital admissions. BMJ 2021;372:n523.
[23] Costantino V, Heslop DJ, MacIntyre CR. The effectiveness of full and partial [50] Heywood AE, Macintyre CR. Elimination of COVID-19: what would it look like
travel bans against COVID-19 spread in Australia for travellers from China and is it possible? Lancet Infect Dis 2020;20:1005–7.
during and after the epidemic peak in China. J Travel Med 2020;27. [51] Blakely T et al. The probability of the 6-week lockdown in Victoria
[24] ABS. National, state and territory population, March 2020 | Australian Bureau (commencing 9 July 2020) achieving elimination of community transmission
of Statistics. (2020). Available at: https://www.abs.gov.au/statistics/people/ of SARS-CoV-2. Med J Aust 2020;213:349–351.e1.
population/national-state-and-territory-population/latest-release. (Accessed: [52] Quigley AL, Yen Nguyen P, Stone H, Lim S, Raina MacIntyre C. Cruise Ship
15th October 2020). Travel and the Spread of COVID-19 – Australia as a Case Study. Int Travel Med
[25] Li Q et al. Early transmission dynamics in Wuhan, China, of novel coronavirus– Cent Iran 2021;9:10–8.
infected pneumonia. Med: N. Engl. J; 2020. [53] Department of Health and Human Services Victoria | Victorian healthcare
[26] He X et al. Temporal dynamics in viral shedding and transmissibility of COVID- worker (clinical and non-clinical) coronavirus (COVID-19) data. Available at:
19. Nat Med 2020;26:672–5. https://www.dhhs.vic.gov.au/victorian-healthcare-worker-covid-19-data.
[27] Bai Lingsheng, Wei Tao, Tian Fei, Jin Dong-Yan, Chen Lijuan, Wang Meiyun YY. (Accessed: 16th December 2020).
Presumed Asymptomatic Carrier Transmission of COVID-19. JAMA Published; [54] Docherty AB et al. Features of 20 133 UK patients in hospital with covid-19
2020. using the ISARIC WHO Clinical Characterisation Protocol: Prospective
[28] COVID-19 Pandemic Planning Scenarios | CDC. Available at: https://www. observational cohort study. BMJ 2020;369.
cdc.gov/coronavirus/2019-ncov/hcp/planning-scenarios.html. (Accessed: 29th [55] Foy BH. et al. Comparing COVID-19 vaccine allocation strategies in India: a
June 2020). mathematical modelling study. medRxiv 2020.11.22.20236091, 2020.
[29] He D et al. The relative transmissibility of asymptomatic COVID-19 infections http://doi.org/10.1101/2020.11.22.20236091.
among close contacts. Int J Infect Dis 2020;94:145–7. [56] Parliament of Victoria - Inquiry into the Victorian Government’s COVID-19
[30] Prem K, Cook AR, Jit M. Projecting social contact matrices in 152 countries Contact Tracing System and Testing Regime. Available at: https://
using contact surveys and demographic data. PLoS Comput Biol 2017;13: www.parliament.vic.gov.au/1005-lsic-lc/inquiry-into-the-victorian-
e1005697. government-s-covid-19-contact-tracing-system-and-testing-regime.
[31] Zhang S et al. Estimation of the reproductive number of novel coronavirus (Accessed: 16th December 2020).
(COVID-19) and the probable outbreak size on the Diamond Princess cruise [57] We modelled how a COVID vaccine roll-out would work. Here’s what we
ship: A data-driven analysis. Int J Infect Dis 2020;93:201–4. found. Available at: https://theconversation.com/we-modelled-how-a-covid-
[32] Wang D et al. Clinical Characteristics of 138 Hospitalized Patients With 2019 vaccine-roll-out-would-work-heres-what-we-found-150544. (Accessed: 2nd
Novel Coronavirus-Infected Pneumonia in Wuhan, China. JAMA 2020. https:// December 2020).
doi.org/10.1001/jama.2020.1585. [58] Aihw. 5.2 Rural and remote populations.
[33] Richardson S et al. Presenting Characteristics, Comorbidities, and Outcomes [59] Trent MJ, Zhang EJ, Chughtai AA, MacIntyre CR. Parental opinions towards the
Among 5700 Patients Hospitalized With COVID-19 in the New York City Area. ‘‘No Jab, No Pay” policy in Australia. Vaccine 2019;37:5250–6.
JAMA 2020;323:2052. [60] Mallory Trent, Holly Seale, Abrar Chinghtai, Salmon Daniel MCR. Trust in
[34] Dagan N et al. BNT162b2 mRNA Covid-19 Vaccine in a Nationwide Mass government, intention to vaccinate and COVID-19 vaccine hesitancy: a
Vaccination Setting. N Engl J Med 2021;NEJMoa2101765. https://doi.org/ comparative survey in the United States, United Kingdom. and Australia.
10.1056/NEJMoa2101765. Unpublished; 2020.
[35] Voysey M et al. Single-dose administration and the influence of the timing of [61] Yamin D, Gerlic M. The rise of SARS-CoV-2 variant B.1.1.7 in Israel intensifies
the booster dose on immunogenicity and efficacy of ChAdOx1 nCoV-19 the role of surveillance and vaccination in elderly. medRxiv
(AZD1222) vaccine: a pooled analysis of four randomised trials. Lancet 2021.02.16.21251819 (2021). http://doi.org/10.1101/2021.02.16.21251819.
2021;397:881–91.