Nonlinear Dyn

https://doi.org/10.1007/s11071-022-07225-9

ORIGINAL PAPER

Assessing the potential impact of immunity waning on the

dynamics of COVID-19 in South Africa: an endemic model
of COVID-19
Musa Rabiu · Sarafa A. Iyaniwura

Received: 6 October 2021 / Accepted: 9 January 2022

© The Author(s), under exclusive licence to Springer Nature B.V. 2022

Abstract We developed an endemic model of COVID- Keywords COVID-19 · Endemic model · Backward
19 to assess the impact of vaccination and immunity bifurcation · Non-pharmaceutical intervention ·
waning on the dynamics of the disease. Our model Vaccination · Immunity waning
exhibits the phenomenon of backward bifurcation and
bi-stability, where a stable disease-free equilibrium
coexists with a stable endemic equilibrium. The epi- 1 Introduction
demiological implication of this is that the control
reproduction number being less than unity is no longer The Severe Acute Respiratory Syndrome Coronavirus
sufficient to guarantee disease eradication. We showed (SARS-CoV) that infected millions of people around
that this phenomenon could be eliminated by either the world in 2002 and the Middle East Respiratory Syn-
increasing the vaccine efficacy or by reducing the dis- drome Coronavirus (MERS-CoV) in 2012 are respon-
ease transmission rate (adhering to non-pharmaceutical sible for many cases of the common cold (Flu) in recent
interventions). Furthermore, we numerically investi- years. In mid-December 2019, twenty individuals suf-
gated the impacts of vaccination and waning of both fering from pneumonia and seven other critically ill
vaccine-induced immunity and post-recovery immu- patients were admitted to the Wuhan Municipal Health
nity on the disease dynamics. Our simulation results Commission (WHC). The majority of them had pro-
show that the waning of vaccine-induced immunity longed exposure to animals such as snakes, poultry
has more effect on the disease dynamics relative to and bats at the Huanan Seafood Market in Wuhan [1].
post-recovery immunity waning and suggests that more After extensive research by the World Health Orga-
emphasis should be on reducing the waning of vaccine- nization (WHO) [2], these individuals were believed
induced immunity to eradicate COVID-19. to be infected by a strange virus that was later named
the Severe Acute Respiratory Syndrome Coronavirus
2 (SARS-CoV-2) by the WHO on February 11, 2020
[3]. The disease caused by the virus was also named
COVID-19 on the same day [3]. As at August 29, 2021,
M. Rabiu (B) there has been over 216,789,388 confirmed cases of
School of Mathematics, Statistics & Computer Science,
University of KwaZulu-Natal, Durban, South Africa
COVID-19 globally, with 4,508,379 confirmed deaths
e-mail: [email protected] and over 193,721,846 recovered individuals [4].
S. A. Iyaniwura
After infection, the coronavirus attacks the body
Department of Mathematics and Institute of Applied Mathemat- cells and replicate quickly resulting in a severe cases
ics, University of British Columbia, Vancouver, BC, Canada of COVID-19 [5,6]. According to the medical prac-

123
 M. Rabiu, S. A. Iyaniwura

titioners, COVID-19 victims experience mild or mod- of concerns about the need, efficacy and safety of many
erate respiratory problems and difficulty in breathing. available vaccines [26–29].
Other symptoms include but not limited to tiredness, South Africa is one of the first African countries
flu, cough, fever and shortness of breath [7,8]. Infected to receive the COVID-19 vaccine. They received their
individuals go through the incubation period of 2–14 first shipment which contains a million doses of the
days for developing symptoms, while some do not show AstraZeneca/Oxford COVID-19 vaccine, on February
symptoms at all during their infection (asymptomatic 1, 2020 [31,32]. A few days later, the vaccination pro-
individuals) [9,10]. It is believed that asymptomatic cess was put on hold over concerns on the efficacy of
individuals account for more than 80% of the commu- the vaccine against the B.1.351 variant of SARS-CoV-
nity transmissions [11]. The incubation period is gener- 2, which was the dominant strain of the virus in the
ally referred to as asymptomatic stage since the infected country [33,34]. The Johnson and Johnson COVID-19
person can still successfully transmit the disease around vaccine, with an initial 80,000 doses [35], was rolled
the community through effective contact [9–11]. The out on the February 17, 2021. This vaccine is observed
individuals in this category are generally encouraged by to be more effective against the B.1.351 strain [33,34].
WHO to self-isolate for a minimum of 14 days and be Several mathematical models have been developed
in constant contact with the COVID-19 regulatory line and analyzed to study the transmission dynamics of
in their community to forestall further transmission. COVID-19 [36–49]. Bugalia et al. [50] developed a
People with severe illness, difficulty in breathing and COVID-19 model using an endemic framework con-
those with underlying medical conditions are encour- sidering the roles of intervention strategies such as
aged to visit the hospital with immediate alacrity [4]. lockdown, hospitalization and quarantine. They com-
According to the virologists, coronavirus is a single- puted the reproduction number and calculated both the
stranded RNA virus, while HIV is a retrovirus (i.e., disease-free and endemic equilibriums of their model.
it has the ability to carry single-stranded RNA as its They also established the epidemiological significance
genetic material instead of double-stranded DNA car- of the equilibrium stability, transcritical bifurcation and
ried by human cells) [12]. The implication of this is boundedness of solution. In order to validate their sys-
that the treatment used in treating diseases with RNA tem of differential equations representing the COVID-
virus can also be used for COVID-19 treatment [13]. 19 dynamics, they fitted the cumulative and new daily
In an effort to curb the transmission of COVID-19, cases in India after which they estimated the model
many countries and regions around the world imple- parameters and predicted the near future scenario of the
mented the non-pharmaceutical interventions (NPIs) disease. The global sensitivity analysis was also estab-
such as physical distancing, wearing of face mask, lished to observe the impact of different parameters of
quarantine, closure of schools/businesses and travel Ro . They also investigated the dynamics of disease with
restriction [14–19]. The implementation of some of respect to partial lockdown, complete lockdown and no
these NPIs has both physical and mental implica- lockdown. Their analyses conclude that if there is par-
tions on the society. Some of these include loss of tial or no lockdown case, then endemic level would be
jobs, loneliness, lack of access to adequate medical unbelievably high and that India may experience more
treatment, economic meltdown, etc. [18,20–25]. The than six million infections if care is not taken.
availability of COVID-19 vaccines has represented a The effect of an imperfect vaccine in reducing
unique opportunity in the fight against COVID-19. In COVID-19 transmission and eventually eradicating the
addition to reducing disease transmission and burden, disease in the USA was examined in [51]. Their math-
widespread vaccination will also allow countries to ematical model provides an estimate for the minimum
safely lift NPIs and promote the gradual recovery from number of susceptible individuals needed to be vac-
the pandemic. Vaccination has been reported as one of cinated to obtain vaccine-induced herd immunity. In
the best development in the public health sector since addition, they investigated the epidemiological conse-
around 1900 [26–29]. It has helped in the eradication quence of the herd immunity threshold and showed
of several infectious diseases such as polio, measles, that the disease can be eliminated if the herd immunity
rubella and smallpox [30]. Despite the public health threshold is achieved. They simulated their model using
benefits derived from vaccination, there are still a lot COVID-19 parameters related to states like Florida and
New York and estimated the herd immunity thresh-

123
Assessing the potential impact of immunity waning

old for these states. The impact of government actions Table 1 Model variables and descriptions
and individual reactions on the coronavirus dynamics Variable Description
in South Africa was assessed using a mathematical
model by [52]. Their model examines the effects of S Unvaccinated susceptible population
individual behavioral reaction and non-pharmaceutical Sv Vaccinated susceptible population
interventions such as quarantine and lockdown on the E1 Exposed population
dynamics of COVID-19 in South Africa. Based on the E2 Pre-symptomatic infectious population
South African officially published COVID-19 data for I Symptomatic infectious population
the period of March 2020 to early May 2020, they dis- R Recovered population
covered that a detection rate of at least 0.5 per day
Our model assumes that individuals in the exposed compartment
may lead to a significant reduction in the total num- (E 1 ) are not infectious, while those in the pre-symptomatic com-
ber of infected cases. They also observed that in the partment (E 2 ) are infectious but do not show symptoms
absence of intervention strategies, the peak number of
cases could be attained earlier than expected which can
also be delayed if the interventions are implemented. biological interpretations. In Sect. 4, we present numer-
The importance of having a COVID-19 vaccine that is ical computations of the control reproduction number
effective in older individuals was emphasized in [53]. derived in Sect. 3, in terms of the vaccination and immu-
Johnson and Bruce [54] proposed a framework for nity waning parameters. Also in this section, we study
modeling fear of infection and frustration with social the effects of vaccination and immunity waning of the
distancing during COVID-19 epidemic. They estab- dynamics of the disease. A brief discussion and future
lished that the SEIR behavior-perception model has work conclude the paper in Sect. 5.
three principal modes of qualitative behavior: no out-
break, controlled outbreak and uncontrolled outbreak.
They fitted their model to the cumulative cases of 2 Mathematical model
COVID-19 and mortality data for different regions and
showed that their model can produce and sustain waves We develop a six compartment endemic model of
of infections. Their results show that the region with COVID-19 to study the effect of immunity waning on
significant decline after wave of infection may likely the dynamics of the disease. Our model incorporates
survive the second wave, while those with moderate vaccination as a pharmaceutical intervention strategy
success in controlling their initial COVID-19 outbreak and assumes that the immunity provided by the vac-
are most likely to experience substantial second waves cine (vaccine-induced immunity) and immunity due
or devastating outbreaks. to infection (post-recovery immunity) wane over time
In this work, we develop and analyze a novel [55–58]. The implementation of vaccination in our
endemic model of COVID-19 that incorporates vacci- model is such that it reduces the chances of suscepti-
nation as a pharmaceutical intervention strategy (PIS). ble individuals acquiring the infection. The susceptible
This model considers the waning of both vaccine- compartment is divided into two: unvaccinated suscep-
induced immunity and post-recovery immunity. We tible population (S) and vaccinated susceptible popula-
derive the control reproduction number for our model tion (Sv ). We also divided the infected population into
and study the disease-free and endemic dynamics of three: exposed (not infectious) (E 1 ), pre-symptomatic
the model with a focus on the South African COVID- (infectious without symptoms) (E 2 ) and symptomatic
19 scenario. Our goal is to use this model to study the infectious (I ). The recovered population is denoted by
effect of immunity waning on both the control repro- R. We assume that recovered individuals do not have
duction number and the disease dynamics. permanent immunity to the disease, and they become
The rest of the paper is structured as follows. In susceptible again at some rate (post-recovery immu-
Sect. 2, we present the model formulation and state the nity waning). The vaccine and post-recovery immunity
assumptions on the model. Section 3 entails the model waning assumptions in the model can also be inter-
analyses. In the first part, we discuss the nonnegativity preted as vaccinated and recovered individuals becom-
of solution and the invariant region, while in the second ing susceptible to another variant of the virus. Table 1
part we establish the bifurcation analysis and discuss its gives the descriptions of the variables in our model.

123
 M. Rabiu, S. A. Iyaniwura

φ2 d E2
Λ = α1 E 1 − α2 E 2 − μE 2 , (2.5)
μ
dt
S E2 dI
μ = α2 E 2 − ψ I − μI − δ I, (2.6)
λ μ α1 dt
dR
σ
α2 = ψ I − φ2 R − μR, (2.7)
φ1 E1 R dt
ψ μ
λ(1 − f ) under the following initial conditions
Sv I
μ
μ S(0) ≥ 0, Sv (0) ≥ 0, E 1 (0) ≥ 0,
δ
E 2 (0) ≥ 0, I ≥ 0, R ≥ 0, (2.8)
Fig. 1 Model schematic. Compartments are as follows: Unvac-
cinated susceptible (S); vaccinated susceptible (Sv ); exposed
where σ is the vaccination rate and 0 ≤ f ≤ 1 is the
(E 1 ); pre-symptomatic infectious (E 2 ); symptomatic infectious vaccine efficacy. Here, f = 0 implies that the vac-
(I ); and recovered (R). Black solid arrows show the flow of indi- cine does not provide any protection from acquiring
viduals through the compartments of the model from susceptible the disease, while f = 1 implies that it is 100% effec-
to recovered. The green arrows show the return of individuals
to the susceptible compartment due to immunity waning. The
tive in preventing new infection. The vaccine and post-
blue arrow indicates birth and immigration into the susceptible recovery immunity waning rates are φ1 and φ2 , respec-
compartment, while the red arrows show natural death and death tively. The natural death rate is denoted by μ, while
due to COVID-19. The rates of transitioning from one compart- the rates of transitioning from E 1 to E 2 and E 2 to I are
ment to the other are shown beside the arrows (see equations
(2.2)–(2.7) for more details)
denoted by α1 and α2 , respectively. Symptomatic infec-
tious individuals recover from the disease at the rate ψ
and die as a result of COVID-19-related complications
Figure 1 shows a schematic diagram of the model, at the rate δ.
where the black and green arrows indicate the direction In Table 2, we present the model parameters, their
of the flow of individuals between the compartments at descriptions and values. The birth and immigration
the rates indicated beside the arrows. rate  and the death rates μ and δ are related to
The green arrows specifically show the flow of indi- South African demography. These parameters were
viduals due to immunity waning. The red arrows show obtained from the official Republic of South Africa
natural death and death due to COVID-19, while the websites. The vaccine and post-recovery immunity
blue arrow shows birth and immigration into the sus- waning parameters are varied throughout this paper.
ceptible population. Our model assumes that immigra- The exact values used in each scenario will be speci-
tion into the population only happens in the susceptible fied in the figure caption.
population and individuals do not leave the population
through emigration. The transmission rates for the pre-
symptomatic and symptomatic individuals are, respec- 3 Model analysis
tively, denoted by β1 and β2 . Using these two rates, we
define our force of infection, λ as In this section, we analyze our ODE model (2.2)–(2.7)
β1 E 2 (t) + β2 I (t) and show the nonnegativity of its solution and that of
λ= , (2.1)
N (t) the invariant region that describes the solution. We also
where N (t) = S(t) + Sv (t) + E 1 (t) + E 2 (t) + I (t) + analyze the linear stability of disease-free equilibrium
R(t) denotes the total population at any time t. The and the endemic equilibrium.
differential equations of our model are given by

dS 3.1 Nonnegativity of solution and the invariant region

=  − λS − σ S + φ1 Sv + φ2 R − μS, (2.2)
dt
d Sv Here, we establish the nonnegativity of the solutions
= σ S − (1 − f )λSv − φ1 Sv − μSv , (2.3) of the ODE model (2.2)–(2.7) as well as the invariant
dt
d E1 region that describes these solutions.
= λS + (1 − f )λSv − α1 E 1 − μE 1 , (2.4)
dt

123
Assessing the potential impact of immunity waning

Table 2 Model parameters, descriptions and values

Parameter Description Value References

 Birth and immigration rate 2795 day−1 [59]

β1 Transmission rate for pre-symptomatic population 0.2189 [52,61]
β2 Transmission rate for symptomatic population 0.3521 [52,61]
λ Force of infection Computed See (2.1)
σ Vaccination rate Varied
f Vaccine efficacy Varied
φ1 Vaccine immunity waning rate Varied
φ2 Post-recovery waning rate Varied
μ Natural death or emigration rate 0.0070 [59]
α1 Rate of transitioning from E 1 to E 2 1/5 day−1 [41,52,62]
α2 Rate of transitioning from E 2 to I 1/2 day−1 [2,12,52,63]
ψ Recovery rate 1/7 day−1 [12,52,63]
δ COVID-19-induced death rate 0.0014 [60,64–66]

The birth and immigration rate  and the death rates μ and δ are based on South African demography and were obtained from the
official Republic of South Africa government websites [59,60]

 t 
Lemma 1 Let ϒ represents the closed set defined as S(t) exp [σ + μ + λ(η)]dη
 0
 t  t 
ϒ = (S, Sv , E 1 , E 2 , I, R) ∈ R6+ |
≥ S(0) +  exp [σ + μ + λ(η)]dη du,

 0 0
S + Sv + E 1 + E 2 + I + R ≤ , (3.1) which implies that
μ   t 
then equation (3.1) is positively invariant and attract- S(t) ≥ S(0) exp −(σ + μ)t − λ(η)dη
ing with regard to (2.2)–(2.7). 0
 t   u 
Proof Using (2.8) as the initial conditions of the ODE + exp (σ + μ)u + λ(η)dη du
0 0
system (2.2)–(2.7), from (2.2), we have   t 
× exp −(σ + μ)t − λ(η)dη . (3.3)
dS 0
= − λS − σ S + φ1 Sv + φ2 R − μS, Since S(0) is a nonnegative constant (see (2.8)), (3.3)
dt
≥ − σ S − μS − λS =  − (σ + μ + λ)S. ensures that the solution S(t) is nonnegative.
(3.2) Following the same approach, the nonnegativity of
Sv , E 1 , E 2 , I and R can be established provided that
Observe that equation (3.2) is a first-order linear dif- the initial conditions exist for all time t > 0.
ferential equation. The integrating factor is chosen as More so, summing up equations (2.2)–(2.7) gives
 t  d N (t)
=  − μN − δ I ≤  − μN , (3.4)
Z (t) = exp [σ + μ + λ(η)] dη , dt
0
whose solution is given by
where λ(η) ≡ λ(S, Sv , E 1 , E 2 , I, R) so that the solu-  
 
tion to (3.2) with equality is N (t) ≤ + N (0) − exp(−μt), (3.5)
 t μ μ
t

S(t)Z (t) 0 =  Z (t) dt. where exp(μt) is the integrating factor. Taking the limit
0 of both sides as t → ∞ gives
 
Evaluating the equation above and using the fact that   
S(0) ≥ 0, we have lim N (t) ≤ + lim N (0) − exp(−μt) ≤ .
t→∞ μ t→∞ μ μ

123
 M. Rabiu, S. A. Iyaniwura

This means that the expression in (3.5) is bounded The proof of this theorem is standard and can be estab-
above by  μ in the domain described in Lemma 3.1. lished using Theorem 2 of [68]. Recall that the parame-
Hence, the ODE system (2.2)–(2.7) is well-posed epi- ter Rcon is used to measure the average number of new
demically and meaningful mathematically since all the COVID-19 cases generated by a single infected indi-
state variables are nonnegative for all t > 0. Thus, we vidual introduced into a susceptible population where
proceed to analyze the model in ϒ. This completes the a certain fraction of individuals are vaccinated. The
proof.
 biological interpretation of Theorem 3.1 is that if the
control reproduction number Rcon < 1, the influx of a
new COVID-19 case in the community will not over-
3.2 Local stability of disease-free equilibrium (DFE). whelm the community, but the whole community will
be endemic if Rcon > 1.
The disease-free equilibrium of (2.2)–(2.7) is given by In Figure 2, we present numerical simulations of
the ODE system (2.2)-(2.7) showing the disease-free
C1 = (S c , Svc , E 1c , E 2c , I c , R c ) dynamics. The initial conditions used for this simula-

(φ1 + μ) σ tion are given in Table 4, and the parameters are given
= , , 0, 0, 0, 0 ,
μ(σ + φ1 + μ) μ(σ + φ1 + μ) in Table 2. The vaccination rate is set as σ = 0.85
(3.6) with vaccine efficacy f = 0.85, while the waning
rate for both vaccine-induced and post-recovery immu-
where N c = S c + Svc . We calculate the control repro- nity is φ1 = φ2 = 0.3. For these set of parame-
duction number using the next generation matrix oper- ters, the corresponding control reproduction number
ator [67,68] on (2.2)–(2.7) as used by [44,69]. The is Rcon = 0.9896, which is less than 1. As observed
matrix for the rates of new infections F and that of from the results in the right panel of this figure, the
transfer of infected individuals V at the DFE are given dynamics of the infected compartments (E 1 , E 2 and I )
by tends to zero, while the susceptible populations (S and
⎛ β [S c + (1 − f )S c ] β [S c + (1 − f )S c ] ⎞ Sv ) initially decrease and then stabilize at some positive
1 v 2 v
0
⎜ N c N c ⎟ numbers (see left panel), which leads to the disease-free
F = ⎝0 0 0 ⎠ equilibrium (DFE), as stated in Theorem 3.1.
0 0 0 Using the same set of parameters but increasing the
⎛ ⎞ immunity waning rates to φ1 = φ2 = 0.5, the cor-
α1 + μ 0 0
V = ⎝ −α1 α2 + μ 0 ⎠, responding control reproduction number increases to
0 −α2 ψ + μ + δ Rcon = 1.2322. The results for this scenario are shown
in Fig. 3. As stated by Theorem 3.1, the DFE is unstable
The next generation matrix is defined by FV −1 . Com- for this scenario since Rcon > 1. From the right panel
puting the spectral radius (dominant eigenvalue) of the of Fig. 3, we observe that unlike in Fig. 2, the dynam-
next generation matrix, we obtain the control reproduc- ics of the infected compartments do not tend to zero,
tion number
  rather they saturate as some positive numbers. Similar
α1 σ (1 − f ) + φ1 + μ behaviors can be seen in the inset of the plot in the
Rcon =  
(α1 + μ)(α2 + μ) σ + μ + φ1 left panel of Fig. 3 for the susceptible populations (S
 
β2 α2 and Sv ) as they reduced drastically but do not tend to
β1 + . (3.7) zero. This result shows that the DFE is unable when
(ψ + μ + δ)
Rcon > 1, and as a result, the disease cannot be eradi-
The first term in the control reproduction number
cated from the population. The interpretation of this is
(3.7) accounts for the infections caused by the pre-
that the endemic equilibrium remains stable whenever
symptomatic individuals in E 2 , while the second term
Rcon > 1.
accounts for those caused by symptomatic individuals
in (I ). Using (3.7), we establish the following result.
Theorem 3.2 The endemic equilibrium points of the
Theorem 3.1 The DFE of the COVID-19 model in COVID-19 model in (2.2)–(2.7) is locally asymptoti-
(2.2)–(2.7) is locally asymptotically stable (LAS) if cally stable (LAS) if Rcon > 1 and unstable if Rcon <
Rcon < 1 and unstable if Rcon > 1. 1.

123
Assessing the potential impact of immunity waning

10 7 10 4
5
4
10
5
15 6000

4 2
4000

2000

3 10
0 0
1000 1500 1000 1500

2
5
1

0 0
0 500 1000 1500 0 500 1000 1500

Fig. 2 Disease-free dynamics The time dynamics of the sus- is σ = 0.85, with vaccine efficacy, f = 0.85. The vaccine and
ceptible (S), susceptible vaccinated (Sv ) and recovered (R) post-recovery immunity waning rates are φ1 = φ2 = 0.3, corre-
populations (left panel), and those for the exposed (E 1 ), pre- sponding to a control reproduction number of Rcon = 0.9896.
symptomatic (E 2 ) and symptomatic (I ) populations (right panel), Remaining parameters are given in Table 2
showing the disease-free equilibrium dynamics. Vaccination rate

10 7 10 5
5 10
5 12 10
4
3 6

4 2
10 4

1 2
8
3 0 0
1000 1500 1000 1500
6
2
4
1 2

0 0
0 500 1000 1500 0 500 1000 1500

Fig. 3 Endemic dynamics The time dynamics of the susceptible σ = 0.85, with vaccine efficacy, f = 0.85. The vaccine and
(S), susceptible vaccinated (Sv ) and recovered (R) populations post-recovery immunity waning rates are φ1 = φ2 = 0.5, corre-
(left panel), and those for the exposed (E 1 ), pre-symptomatic sponding to a control reproduction number of Rcon = 1.2322.
(E 2 ) and symptomatic (I ) populations (right panel), for when Remaining parameters are given in Table 2
the disease is endemic in the population. Vaccination rate is

This theorem shows the condition for the stability of the equilibrium and show that our model undergoes the
endemic equilibrium point. The proof of this theorem phenomenon of backward bifurcation.
is standard and can be established using Theorem 2 of
[68]. In the next subsection, we analyze the endemic

123
 M. Rabiu, S. A. Iyaniwura

3.3 The backward bifurcation and endemic where the newly defined parameters g1 , g2 , . . . , g7 are
equilibrium (EE)
g1 = σ + μ, g2 = 1 − f,
Bifurcation is the change in the qualitative structure g3 = φ1 + μ, g4 = α1 + μ,
of the flow of a differential equation with respect to g5 = α2 + μ, g6 = ψ + μ + δ, g7 = ψ2 + μ.
parameter values. The point where bifurcation occurs is Solving (3.9)–(3.14) simultaneously gives
called the bifurcation point. Examples of bifurcations
ψI α2 E2
in dynamical systems include transcritical, pitchfork, R= , I= ,
Hopf, saddle node, forward, backward, etc. [70–72]. g7 g6
However, the bifurcation type that is related to our study α1 E1
E2 = ,
is the backward bifurcation, which is explained in [73] g5
based on the provision of the center manifold theory. λc (S + g2 Sv )
E1 = , (3.15)
The COVID-19 model presented in (2.2)–(2.7) has a g4
positive and unique endemic equilibrium point, which σS  + φ1 Sv + φ2 R
Sv = , S=
can be interpreted as the point where at least one g2 λc + g3 λc + g1
infected population exists and is nonzero. Let  − δI
N = ,
C1 = (S, Sv , E1 , E2 , I, R), μ

be the endemic equilibrium point. We seek to solve the Again, solving the equations in (3.15) simultaneously
ODE system (2.2)–(2.7) in terms of the parameter λc gives
as follows: The force of infection at steady state can be ψα1 α2 (S + g2 Sv )λc
R= . (3.16)
written as g4 g5 g6 g7
β1 E2 + β2 I Substituting the expression for Sv in (3.15) into (3.16)
λc = . (3.8)
N and factorizing, we have
The model equations (2.2)–(2.7) can be re-expressed ψ α1 α2 (g2 λc + g3 + g2 σ )λc S
as R= . (3.17)
g4 g5 g6 g7 (g2 λc + g3 )
Similarly, we get

g4 g5 g6 g7 (g2 λc + g3 )
S= . (3.18)
g4 g5 g6 g7 [(g2 λc + g3 )(λc + g1 ) − φ1 σ ] − φ2 ψα1 α2 [(g2 λc + g3 ) + σ g2 ]λc

Substituting (3.18) into (3.17) and factorizing give

 ψ α1 α2 (g2 λc + g3 + g2 σ )λc
R= . (3.19)
g4 g5 g6 g7 [(g2 λc + g3 )(λc + g1 ) − φ1 σ ] − φ2 ψα1 α2 [(g2 λc + g3 ) + σ g2 ]λc

 − (λ + g1 )S + φ1 Sv + φ2 R = 0, (3.9)
σ S − g2 λSv − g3 Sv = 0, (3.10)
λ(S + g2 Sv ) − g4 E 1 = 0, (3.11)
α1 E 1 − g5 E 2 = 0, (3.12)
α2 E 2 − g6 I = 0, (3.13)
ψ I − g7 R = 0. (3.14)

123
Assessing the potential impact of immunity waning

Similarly, we have

 σ g4 g5 g6 g7
Sv = ,
g4 g5 g6 g7 [(g2 λc + g3 )(λc + g1 ) − φ1 σ ] − φ2 ψα1 α2 [(g2 λc + g3 ) + σ g2 ]λc
 g5 g6 g7 (g2 λc + g3 + g2 σ )λc
E1 = ,
g4 g5 g6 g7 [(g2 λc + g3 )(λc + g1 ) − φ1 σ ] − φ2 ψα1 α2 [(g2 λc + g3 ) + σ g2 ]λc
 α1 g6 g7 (g2 λc + g3 + g2 σ )λc
E2 = ,
g4 g5 g6 g7 [(g2 λc + g3 )(λc + g1 ) − φ1 σ ] − φ2 ψα1 α2 [(g2 λc + g3 ) + σ g2 ]λc
 α1 α2 g7 (g2 λc + g3 + g2 σ )λc
I= , (3.20)
g4 g5 g6 g7 [(g2 λc + g3 )(λc + g1 ) − φ1 σ ] − φ2 ψα1 α2 [(g2 λc + g3 ) + σ g2 ]λc
 
 δ  α1 α2 g7 (g2 λc + g3 + g2 σ )λc
N = − . (3.21)
μ μ g4 g5 g6 g7 [(g2 λc + g3 )(λc + g1 ) − φ1 σ ] − φ2 ψα1 α2 [(g2 λc + g3 ) + σ g2 ]λc

By substituting all the expressions in (3.18), (3.19), expansion is cumbersome and thus ignored. More-so,
(3.20) and (3.21) into (3.8), it can be proved that the x4 is positive if and only if Rcon < 1 and negative if
nonzero equilibria of the model satisfy and only if Rcon > 1. This shows that the number of
x0 λc 4 + x1 λc 3 + x2 λc 2 + x3 λc + x4 = 0, (3.22) positive real roots of (3.22) is dependent on the signs
of x1 , x2 and x3 . These signs can be analyzed using the
where Descartes rule of signs on (3.22). The possible roots of
equation (3.22) are given in Table 3. Hence, we have
x0 = g2 g4 g5 g6 (g4 g5 g6 g7 − φ2 α1 α2 ψ) the following theorem.
(g2 g4 g5 g6 g7 − ψ φ2 α1 α2 g2 − δ α1 α2 g2 g7 ), Theorem 3.3 The COVID-19 model (2.2)–(2.7)
x1 = (k2 g1 g2 + g3 k2 + φ2 ko g3 + φ2 ko σ g2 )
1. has a unique endemic equilibrium if Rcon > 1 and
(m 1 m 4 − δα1 α2 k2 g2 )
whenever cases 1, 2, 6 and 8 of Table 3 are satisfied;
+ g2 (k2 − φ2 ko )[m 2 m 4 − δα1 α2 k2 g3
2. could have more than one endemic equilibrium if
− δα1 α2 g2 σ k2 − μ(β1 α1 g6 + β2 α1 α2 )g2 k2 ],
Rcon > 1 and cases 3, 4, 5 and 7 of Table 3 are
x2 = k2 m 3 [m 1 m 4 − δα1 α2 k2 g2 + m 1 g4 g5 g6 ] satisfied;
+ (k2 g1 g2 + g3 k2 + φ2 ko g3 + φ2 ko σ g2 ) 3. could have 2 or more endemic equilibria if Rcon <
[m 2 m 4 − δα1 α2 k2 g3 − δα1 α2 g2 σ k2 − μm 7 ] 1 and cases 2–8 of Table 3 are satisfied.
− μα1 k2 m 1 (g3 + σ g2 )(β1 g6 + β2 α2 ),
x3 = m 3 [m 2 m 4 − δα1 α2 k2 (g3 − g2 σ ) + m 2 m 4 − μm 7 ] Lemma 2 The COVID-19 model has at least one posi-
− μm 2 α1 k2 (g3 + σ g2 )(β1 g6 + β2 α2 ),
tive endemic equilibrium whenever Rcon > 1 but could
have zero, two or four positive endemic equilibrium
x4 = k22  g4 g5 g6 (g1 g3 − φ1 σ )2 (1 − Rcon ).
whenever Rcon < 1.

and Theorem 3.2 can be used to establish the stability of

these endemic equilibrium points. It is worth noting that
K 0 = ψα1 α2 , k2 = g4 g5 g6 g7 , m 1 = g2 (k2 − φ2 k0 ), the exact expression for the endemic equilibria can be
m 2 = k2 g1 g2 + g3 k2 + φ2 k0 g3 + φ2 k0 σ g2 , obtained by solving for λc in (3.22) to obtain the roots
m 3 = k2 (g1 g3 − φ1 σ k2 ), m 4 = g4 g5 g6 , and then substitute its positive value into the expres-
m 7 = k2 g2 α1 (β1 g6 + β2 α2 ). sions in (3.18), (3.19), (3.20) and (3.21). The existence
of multiple endemic equilibria when the control repro-
We can easily establish the positiveness of the coef- duction number is less than unity established using The-
ficient x0 by substituting the respective value of all the orem 3.3, Lemma 2 and Table 3 suggests the possibility
parameters in x0 and subject the resultant expression of backward bifurcation [70,72]. As previously men-
into full expansion. The negative signs will disappear, tioned, backward bifurcation is a phenomenon where a
and the expression will become positive. The complete stable DFE coexists with a stable endemic equilibrium

123
 M. Rabiu, S. A. Iyaniwura

Fig. 4 Bifurcation diagrams. The bifurcation diagrams for dif- shows the bifurcation diagram for f = 0.65 and σ = φ1 =
ferent values of vaccine efficacy f and different transmission φ2 = 0.5, while other parameter values remain unchanged. The
rates β1 and β2 . The top left panel is the original bifurcation bottom right panel shows the bifurcation diagram f = 0.55,
diagram for f = 0.55 and σ = φ1 = φ2 = 0.5, while other σ = φ1 = φ2 = 0.5 and β1 = 0.0219, β2 = 0.0352 (which is
parameter values remain unchanged. The top right panel is the 10% of its baseline value), while other parameter values remain
bifurcation diagram for f = 0.15 and σ = φ1 = φ2 = 0.5, while unchanged
other parameter values remain unchanged. The bottom left panel

when the control reproduction number is less than 1. while the red-dashed line shows that the disease-free
We present the bifurcation diagrams in Fig. 4. equilibrium is unstable for Rcon > 1. We observe from
The bifurcation diagram on the top left panel of Fig. this result that there are two steady states of the system
4 shows the behavior of the endemic model (2.2)–(2.7) (DFE and endemic equilibrium) when Rcon > 1. As
when the control reproduction number Rcon is less Rcon decreases, we have the emergence of an endemic
than, equal and greater than 1. The parameter values equilibrium (red-circled line) at Rcon = 1, which is
used are given in Table 1 with σ = φ1 = φ2 = 0.5 and unstable, that coexists with a stable endemic equilib-
vaccine efficacy set at f = 0.55. The red-circled line rium (blue-circled line) and a stable disease-free equi-
on the diagram shows that the endemic equilibrium is librium (blue-dashed line). The backward movement
unstable for Rcon < 1, while the blue-circled part of of the equilibrium that emerges at Rcon = 1 clearly
the diagram shows the stability of the endemic equi- shows that our model exhibits the phenomenon of back-
librium for Rcon > 1. The blue-dashed line shows the ward bifurcation, where a stable disease-free equilib-
stability of the disease-free equilibrium for Rcon < 1, rium point coexists with a stable endemic equilibrium

123
Assessing the potential impact of immunity waning

point. The epidemiological implication of this is that rate σ for vaccine efficacy f = 0.5 (left panel) and
the control reproduction number being less than unity f = 0.85 (right panel).
is no longer sufficient to guarantee the disease erad- We observe from the results in this figure that the
ication. More control measures would be required to control reproduction number Rcon decreases as the
eradicate the disease. For this reason, we considered vaccination rate increases and increases as the vaccine
the effects of increase in vaccine efficacy and reduc- immunity waning rates increase. When the vaccination
tion in transmission rates on the bifurcation diagram. rate is relatively small, vaccine immunity waning has
The top right panel shows the behavior of the bifur- little effect on the reproduction number, compared to
cation diagram when the vaccine efficacy f is reduced when a large fraction of the population is vaccinated.
to 0.15, while other parameter values remain the same. In this case, an increase in the waning rate increases
We can see from this result that bi-stability persist and the control reproduction number. In addition, compar-
now exist for all Rcon < 1. The bottom left panel shows ing the results in the left panel of this figure for vaccine
the behavior of the model when the vaccine efficacy f is efficacy f = 0.5 to that of the right panel for f = 0.85,
increased to 0.65, while other parameter values remain we notice that the vaccine immunity waning rate has
the same. We can easily notice the disappearance of bi- more effect on the control reproduction number when
stability. Also, we observe that the endemic equilibrium the vaccine is more effective.
is unstable for some values of Rcon > 1. For these val- Similar results are shown in Fig. 6 for the trans-
ues of Rcon , the DFE equilibrium is stable (not shown). mission and vaccination rates. In the top left panel
This clearly underlines the need for a highly efficacious of this figure, we show the control reproduction num-
vaccine in controlling the spread of the disease and its ber as a function of the pre-symptomatic transmission
eventual eradication. The bottom right panel shows the rate (β1 ) and the vaccination rate (σ ) for φ1 = 0.1.
behavior of the model when the transmission rates β1 For the same vaccine waning rate, we present the
and β2 are reduced to 0.02189 and 0.03521, respec- reproduction number as a function of the symptomatic
tively (which is 10% of its baseline value), while other transmission rate (β2 ) and the vaccination rate (σ ).
parameter values remain the same. We observe from The remaining parameters are given in Table 2. As
this result that there is also no bi-stability in the system expected, the control reproduction number increases
for this scenario. This result emphasizes the importance as the transmission rates β1 and β2 increase, although
of reduction in community transmission in the fight the symptomatic transmission rate has more effect on
against the disease. The lower/higher the transmission the control reproduction number compared to the pre-
rates, the lower/higher the disease prevalence. On the symptomatic transmission rate. The results in the lower
other hand, the higher/lower the vaccine efficacy, the panel of Fig. 6 are similar to those in the top panel but
higher/lower the protection against the disease. These for a higher waning rate of vaccine-induced immunity
two parameters are important in controlling the disease (φ1 = 0.5). Comparing these results with those in the
spread and its eventual eradication. top panel, we notice that an increase in φ1 increases
the dependence of the control reproduction number on
the vaccination rate. Similarly, when φ1 is small (say
4 Numerical simulations φ1 = 0.1 as shown in the top panel of Fig. 6), the
transmission rates have more effect on the reproduc-
We present the numerical simulations of the endemic tion number when a small fraction of the population
model (2.2)–(2.7) for different scenarios to study the is vaccinated compared to when the vaccination rate
effect of immunity waning on the dynamics of COVID- is high. However, the effect of the transmission rates
19. The parameters used are related to the demogra- on the control reproduction number when vaccination
phy of South Africa. These parameters are given in rate is higher increases as the immunity waning rate
Table 2. We studied the effect of immunity waning on increases. This can easily be seen by comparing the
the control reproduction number and the entire diseases results in the top panel to those in the bottom panel for
dynamics. high vaccine efficacy.
In Fig. 5, we present the contour plots of the con- Next, we study the effect of immunity waning on the
trol reproduction number Rcon in terms of the vaccine dynamics of COVID-19 using South Africa as a case
immunity waning parameter φ1 and the vaccination study. The results of this study are presented in terms

123
 M. Rabiu, S. A. Iyaniwura

Table 3 Possible positive real roots of equation (3.22) for Rcon < 1 and Rcon > 1
Cases A1 A2 A3 A4 A5 R0 No of times No of positive real roots
sign changed (endemic equilibrium point)

1 + + + + + <1 0 0
+ + + + − >1 1 1
2 + + + − + <1 2 0,2
+ + + − − >1 1 1
3 + + - + + <1 2 0,2
+ + − + − >1 3 1,3
4 + − + + + <1 2 0,2
+ − + + − >1 3 1,3
5 + − − + + <1 2 0,2
+ − − + − >1 3 1,3
6 + − − − + <1 2 0,2
+ − − − − >1 1 1
7 + − + − + <1 4 0,2,4
+ − + − − >1 3 1,3
8 + + − − + <1 2 0,2
+ + − − − >1 1 1

1 1
2.5 2.5
2.2

1.8
0.8 2
0.8 2
6
1.
2
2.6
2.6

2.4
2.4

0.6 0.6
2.2

1.5 1.4 1.5

2.2

1.8
2

1.8 1.
6 1.2
0.4 2 1 0.4 1
1.4
1
1.8 1.2
0.2 0.5 0.2 0.5
2.6
2.6

.4
2.2 2.4

1.8 .2

1.6 1
2 2 2

6 0.8
2 1..4
1.8 1.6 1.4 111.2 0.8 0.6 0.6
1.4 0 0
0 0.2 0.4 0.6 0.8 0 0.2 0.4 0.6 0.8

Fig. 5 Effect of vaccination and vaccine immunity waning on the control reproduction number. Left panel: vaccine efficacy, f = 0.5.
Right panel: vaccine efficacy, f = 0.85. Remaining parameters are given in Table 2

of the time dynamics of the symptomatic infectious also related to the demography of South Africa and are
population. presented in Table 2.
The initial conditions used are presented in Table 4 Figure 7 shows the dynamics of the symptomatic
and are based on the South African COVID-19 statistics population over time for different vaccine immunity
as of August 29, 2021, and were obtained from the waning rates, φ1 (left panel), and post-recovery immu-
Department of health COVID-19 report, Republic of nity waning rate, φ2 (right panel). We observe from
South Africa [74]. In addition, the parameters used are these results that vaccine immunity waning has much
more effect on the disease dynamics than the post-

123
Assessing the potential impact of immunity waning

1 4 1 4

2
3

2
0.8 1 0.8
3.5

5
1.

2.5 3
3
2.5

3
1
0.6 0.6
1.5

1 2
2

2
2

1
3

1.5
0.4 0.4
0.5
2.5

1 1

1
0.5
0.2 0.2
1.5

0.5
1

0 0 0 0
0 0.2 0.4 0.6 0.8 0 0.2 0.4 0.6 0.8

1 4 1

6
4 3 6
3

0.8 0.8
3.5

5
2

3 5
3
2.5

0.6 0.6 4 4
2
2
3
3

3 2
2

0.4 1.5
0.4
2
2.5

2 1
1
0.2 0.2 1
1
5 1
1.
2

0 0 0 0
0 0.2 0.4 0.6 0.8 0 0.2 0.4 0.6 0.8

Fig. 6 Effect of vaccination rate and disease transmission rates rate β2 . In the first row, the vaccine immunity waning rate is
on the control reproduction number. Left panel: pre-symptomatic φ1 = 0.1, and φ1 = 0.5 in the second row. Vaccine efficacy is
transmission rate β1 , and right panel: symptomatic transmission set as f = 0.85. Remaining parameters are given in Table 2

Table 4 Model initial conditions. These initial conditions are COVID-19 report, Republic of South Africa [74]. The unvac-
based on the South African COVID-19 statistics as of August cinated susceptible population was computed using S(0) =
29, 2021, and were obtained from the Department of health N (0) − (Sv (0) + E 1 (0) + E 2 (0) + I (0) + R(0))

Variable Initial population Value References

N (0) Total population 60, 142, 978 [75]

S(0) Unvaccinated susceptible population 45, 414, 814 Computed
Sv (0) Vaccinated susceptible population 12, 021, 608 [74,76]
E 1 (0) Exposed population 15, 480 Inferred [74]
E 2 (0) Pre-symptomatic population 7,740 Inferred [74]
I (0) Symptomatic population 157, 137 [74]
R(0) Recovered population 2, 526, 199 [74]

123
 M. Rabiu, S. A. Iyaniwura

10 6 10 6
4 4

3 3

2 2

1 1

0 0
0 200 400 600 800 1000 0 200 400 600 800 1000

Fig. 7 Effect of immunity waning on disease dynamics. The reproduction numbers are Rcon = 0.9896, 1.2322, 1.4504 and
dynamics of the symptomatic population over time for different 1.5516, respectively, while in the right panel, Rcon = 1.2322 for
vaccine immunity waning rates (left panel) and post-recovery all values of φ2 . Vaccination rate used is σ = 0.85, with vaccine
immunity waning rates (right panel). In the left panel, for the wan- efficacy, f = 0.85. Remaining parameters are given in Table 2
ing rates φ1 = 0.3, 0.5, 0.75 and 0.9, the corresponding control

recovery immunity waning. For this scenario, we used φ1 used. This result suggests that we need to keep the
φ1 = 0.3, 0.5, 0.75 and 0.9, which give the cor- vaccination rate high in order to eradicate the disease.
responding control reproduction numbers Rcon = The result in Fig. 9 is for when the vaccine efficacy
0.9896, 1.2322, 1.4504 and 1.5516, respectively. By is reduced from f = 0.85 used in Fig. 7 to f = 0.65,
varying the vaccine waning rate from φ1 = 0.3 through while the vaccination rate remains σ = 0.85. Similar
φ1 = 0.9, the control reproduction number goes from to the results in Fig. 8, the disease is endemic in the
Rcon = 0.9896 (Rcon < 1) to Rcon = 1.5516 population for all values of the waning parameter φ1 .
(Rcon > 1) changing the disease dynamics from The increase in infections observed in Fig. 9 relative
disease-free to endemic (See the left panel of Fig. 7). to Fig. 7 is due to a decrease in the efficacy of the
However, this is not the case for the post-recovery vaccine, since the same vaccination rate is used for the
immunity waning rate φ2 . This parameter has little two scenarios. This suggests that in addition to high
effect on the dynamics of the disease (right panel vaccination rate, a high vaccine efficacy is also essential
of Fig. 7). In addition, since this control reproduc- to eradicate the disease.
tion number is independent of this parameter, Rcon = Lastly, we study the effect of the transmission rates
1.2322 for all values of φ2 . on the disease dynamics. In the top panel of Fig. 10,
In Fig. 8, we study the effect of the vaccination rate, we present the results for when the pre-symptomatic
σ and the immunity waning rates φ1 and φ2 on the dis- transmission rate is increased to β1 = 0.3284, while
ease dynamics. In this figure, we reduce the vaccination the symptomatic transmission rate remains the same
rate, from σ = 0.85 used in Fig. 7 to σ = 0.5. Our goal as in the remaining figures (β2 = 0.3521). On the
is to study how a reduction in vaccination rate would other hands, in the lower panel we increased the symp-
affect the disease dynamics. Comparing the results in tomatic transmission rate to β2 = 0.4049, while the
Fig. 7 to those in Fig. 8, we observe that a decrease in pre-symptomatic transmission rate remains the same
vaccination rate leads to more infections in the popu- as in the other figures (β1 = 0.2189). As expected,
lation. Unlike the case where vaccination rate is high, there are more infections as the immunity waning rates
when vaccination rate is σ = 0.5, there are no disease- increase. Comparing the results in this figure to those in
free dynamics for all values of the waning parameter Fig. 7, we observe that an increase in the transmission
rates will increase the disease spread in the community

123
Assessing the potential impact of immunity waning

10 6 10 6
6 6

4 4

2 2

0 0
0 200 400 600 800 1000 0 200 400 600 800 1000

Fig. 8 Effect of vaccination rate and immunity waning on dis- the corresponding control reproduction numbers are Rcon =
ease dynamics. The dynamics of symptomatic population over 1.2474, 1.5230, 1.7442 and 1.8392, respectively, while in the
time for different vaccine immunity waning rates (left panel) right panel, Rcon = 1.5230 for all values of φ2 . Vaccination
and post-recovery immunity waning rates (right panel). In the rate is reduced to σ = 0.5, with vaccine efficacy, f = 0.85.
left panel, for the waning rates φ1 = 0.3, 0.5, 0.75 and 0.9, Remaining parameters are given in Table 2

making it difficult to eradicate the disease. In addition, partment. Both immunity waning described above can
an increase in the symptomatic transmission rate sig- also be interpreted as being susceptible to another strain
nificantly affects the disease dynamics relative to the of COVID-19. Our model implements a sterilizing vac-
pre-symptomatic transmission rate. Note that for all cine that reduces the chances of getting infected. We
the examples considered, the post-recovery immunity apply our model to the South African COVID-19 sce-
waning rate has little effect on the disease dynamics nario.
compared to the waning rate of vaccine-induced immu- We analyzed our model rigorously by first comput-
nity. ing the control reproduction number for the disease-
free equilibrium, after which we analyzed the endemic
equilibrium and show that our model exhibits the back-
5 Discussion ward bifurcation phenomenon. We examined the condi-
tion for the existence of bifurcation and endemic equi-
We have developed and analyzed an endemic model librium. These conditions are stated in Theorem 3.3.
for COVID-19 to study the effect of immunity wan- Furthermore, Lemma 2 provides the condition for the
ing on the dynamics of this disease. Our model has existence of multiple endemic equilibria with respect
six compartment: unvaccinated susceptible (S), vacci- to the conditions specified in Table 3. This leads to
nated susceptible (Sv ), exposed (not infectious) (E 1 ), the phenomenon where stable DFE coexists with a sta-
pre-symptomatic (infectious but no symptoms), symp- ble endemic equilibrium when the control reproduction
tomatic (I ) and recovered (R). We considered two number is less than unity. Figure 4 shows the bifurca-
types of immunity waning in our model: vaccine immu- tion diagrams for different scenarios. It can be under-
nity waning and post-recovery immunity waning. Vac- stood from this diagrams that if the vaccine efficacy
cine immunity waning occurs when a vaccinated indi- is reduced, bi-stability persists but when increased, the
vidual in the Sv compartment loses their immunity and bi-stability disappears and the disease may die out. A
returns to the unvaccinated susceptible compartment better result is achieved when the transmission rates
S. On the other hand, post-recovery immunity waning are reduced, thereby eliminating the bi-stability totally.
occurs when recovered individuals lose their immu- This clearly underlines the need for a highly effica-
nity and return to the unvaccinated susceptible com-

123
 M. Rabiu, S. A. Iyaniwura

10 6 10 6
6 6

4 4

2 2

0 0
0 200 400 600 800 1000 0 200 400 600 800 1000

Fig. 9 Effect of vaccine efficacy and immunity waning on dis- the corresponding control reproduction numbers are Rcon =
ease dynamics. The dynamics of symptomatic population over 1.3768, 1.5623, 1.7292 and 1.8065, respectively, while in the
time for different vaccine immunity waning rates (left panel) right panel, Rcon = 1.5623 for all values of φ2 . Vaccination
and post-recovery immunity waning rates (right panel). In the rate is σ = 0.85, with vaccine efficacy reduced to f = 0.65.
left panel, for the waning rates φ1 = 0.3, 0.5, 0.75 and 0.9, Remaining parameters are given in Table 2

cious vaccine and reduction in disease transmission to cination on the reproduction number. Our results show
achieve disease eradication. that the control reproduction number increases as the
We numerically studied the effects of vaccination transmission rates increase, although the symptomatic
and the waning of vaccine immunity on the control transmission rate has more effect on the control repro-
reproduction number. These results were presented in duction compared to the pre-symptomatic transmis-
the form of contour plots and show that the control sion rate. This may be because infectious individuals
reproduction number decreases as the vaccination rate spend fewer days (2 days) in the pre-symptomatic com-
increases. However, it increases as the vaccine immu- partment compared to the symptomatic compartment,
nity waning rate increases. As the vaccine efficacy where they spend 7 days. In addition, the transmission
increases, the effects of these parameters (vaccination rates have little effect on the reproduction number when
rate and vaccine immunity waning rate) become more vaccination rate is high, and this effect increases as the
pronounced. There is also a decrease in the control vaccination rate drops. In other words, when we vacci-
reproduction number, although the decrease is more nate a large fraction of the population with a relatively
when the vaccination rate is high. When a small frac- effective vaccine, the transmission rate has a little effect
tion of the population is vaccinated, vaccine immunity on the reproduction number. However, if we vaccinate
waning has little effect on the control reproduction only a small fraction of the population with the same
number, compared to when vaccination rate is high. vaccine when the transmission rate is higher, the con-
This is because when a large fraction of the popula- trol reproduction number increases too.
tion is vaccinated, we expect a reduction in infection. Furthermore, we investigated the effects of different
But when these people lose their immunity at a faster parameters of the model on the disease dynamics. Our
rate, the expected reduction in infection will not be main focus is studying how vaccination and the waning
realized, leading to a significant spread of the disease. of immunity affect the disease dynamics. The results
This is not the case when only a small fraction of the of these investigations were presented in terms of the
population are vaccinated from the onset. In this case, dynamics of the symptomatic population over time and
there will be no much difference whether they lose their show that vaccine immunity waning has more effect on
immunity slowly or at a faster rate. We also studied the disease dynamics compared to the post-recovery
the effect of the transmission rates together with vac- immunity waning. By varying the vaccine immunity

123
Assessing the potential impact of immunity waning

10 6 10 6
5 5

4 4

3 3

2 2

1 1

0 0
0 200 400 600 800 1000 0 200 400 600 800 1000

10 6 10 6
5 5

4 4

3 3

2 2

1 1

0 0
0 200 400 600 800 1000 0 200 400 600 800 1000

Fig. 10 Effects of disease transmission rates and immunity wan- the corresponding control reproduction numbers are Rcon =
ing on disease dynamics. The dynamics of symptomatic popula- 1.0680, 1.3297, 1.5652 and 1.6744, respectively, while in the
tion over time for different vaccine immunity waning rates (left) top right panel, Rcon = 1.5623 for all values of φ2 . Similarly,
and post-recovery immunity waning rates (right). In the top panel, in the bottom left panel, for the waning rates φ1 = 0.3, 0.5, 0.75
the pre-symptomatic transmission rate is β1 = 0.3284 and the and 0.9, the corresponding control reproduction numbers are
symptomatic transmission rate is β2 = 0.3521, while in the lower Rcon = 1.1146, 1.3877, 1.6336 and 1.7475, respectively, while
panel, the pre-symptomatic transmission rate is β1 = 0.2189 and in the bottom right panel, Rcon = 1.3877 for all values of φ2 .
the symptomatic transmission rate is β2 = 0.4049. In the top Remaining parameters are given in Table 2
left panel, for the waning rates φ1 = 0.3, 0.5, 0.75 and 0.9,

waning parameter, we realized that when this parame- and a moderate vaccine waning parameter. By reducing
ter is low, the diseased-free equilibrium (DFE) of the the vaccination rate by ≈ 40%, we noticed that there
model is linearly asymptotically stable. This means that is no value of vaccine immunity waning rate for which
the disease can be eradicated when people retain their the DFE is linearly stable. As a result of this, the dis-
vaccine immunity for a longer time. However, as this ease cannot be eradicated when the vaccination rate is
waning rate is increased, the DFE becomes unstable, low. In summary, keeping the vaccination rate high and
thereby making the disease endemic in the population. preventing loss of immunity are required to eradicate
On the other hand, the DFE remains unstable for all val- COVID-19.
ues of the post-recovery immunity waning parameter

123
 M. Rabiu, S. A. Iyaniwura

An interesting future direction for this work is to parameters to be considered to effective eradicate the
fit our model to the reported cases of COVID-19 for disease.
a specific location/country. In this way, the parame-
ters would be more related to that location and the Acknowledgements No funding was received for this project.
predictions would be more representative of the loca-
Declarations
tion. Another interesting future direction is to extend
the model to have explicit cycles of infection or loss Competing interest The authors declare that there is no com-
of immunity. For this case, there will be all six com- peting interest.
partments of our current model in each cycle such that
Data Availability All used data are properly referenced in the
individuals that lose their immunity in one cycle will body of the work.
go to the susceptible compartment of the next cycle
rather than the susceptible compartment of their cur-
rent cycle as it is in our current model, and the disease References
will progress from there to the other compartments of
1. Cheng, Z.J., Shan, J.: 2019 novel coronavirus where we are
the next cycle. An advantage of structuring the model
and what we know. Infection 48(2), 155–163 (2020)
this way is that one can easily track the number of peo- 2. Li, Q., Guan, X., Wu, P., Wang, X., Zhou, L., Tong, Y.,
ple getting infected at each infection cycle. In addition, Ren, R., Leung, K.S.M., Lau, E.H.Y., Wong, J.Y., Xing, X.,
different cycles of infection or loss of immunity can Xiang, N., Wu, Y., Li, C., Chen, Q., Li, D., Liu, T., Zhao,
J., Liu, M., Tu, W., Chen, C., Jin, L., Yang, R., Wang, Q.,
be interpreted as being susceptible to different vari- Zhou, S., Wang, R., Liu, H., Luo, Y., Liu, Y., Shao, G., Li, H.,
ants of COVID-19, where different vaccine efficacies Tao, Z., Yang, Y., Deng, Z., Liu, B., Ma, Z., Zhang, Y., Shi,
could also be used. This will increase the flexibility of G., Lam, T.T.Y., Wu, J.T., Gao, G.F., Cowling, B.J., Yang,
the model. It would also be worthwhile to stratify the B., Leung, G.M., Feng, Z.: Early transmission dynamics in
Wuhan, China, of novel coronavirus-infected pneumonia. N.
population by age and extend the model to have age Engl. J. Med. 382(13), 1199–1207 (2020). https://doi.org/
structure. This is particularly important since individ- 10.1056/NEJMoa2001316
uals younger than 12 years have not been approved for 3. World Health Organization (WHO) et al. Naming the
vaccination yet. It would also make the model more coronavirus disease (COVID-19) and the virus that
causes it. 2020, 2020. https://www.who.int/emergencies/
suitable for answering age-related questions. diseases/novel-coronavirus-2019/technical-guidance/
Some limitations of the modeling framework pre- naming-the-coronavirus-disease-(covid-2019)-and-
sented in this paper include modeling vaccination as a the-virus-that-causes-it#:∼:text=The%20International
single dose despite the fact that majority of the COVID- %20Committee%20on%20Taxonomy,two%20viruses
%20are%20different
19 vaccines available today require more than one dose 4. Worldometer. Corona-virus data, (accessed August 29,
to get a good level of protection. In addition, our model 2021). https://www.worldometers.info/coronavirus/
assumes that vaccine efficacy is the same, either one is 5. Pathak, N.: The great invader: how covid attacks every organ
getting the vaccine for the first time or one had already (2020)
6. Sheshe, S., Abdullahi Nazifi, A.M., Labbo, G.M.K., Yahya,
experienced waning of immunity. This may not be the A., Muhammad, U., Haruna, A.M.: Mechanism of antiviral
case in reality as getting the vaccine the second time immune response and covid-19 infection. Asian J. Immunol.
may lead to a better/stronger protection. Also, the effi- 9, 1–8 (2020)
cacy of the vaccines for different variants of COVID- 7. Li, L., Huang, T., Wang, Y., Wang, Z., Liang, Y., Huang,
T., Zhang, H., Sun, W., Wang, Y.: Covid-19 patients’ clini-
19 may be different. Another limitation of our model is cal characteristics, discharge rate, and fatality rate of meta-
assuming that the population is well mixed. This may analysis. J. Med. Virol. 92(6), 577–583 (2020)
not be so in reality as contact rates and mixing patterns 8. Wollina, U., Karadağ, A.S., Rowland-Payne, C., Chiriac,
may vary between individuals depending on their age A., Lotti, T.: Cutaneous signs in covid-19 patients a review.
Dermatol. Ther. 33(5), e13549 (2020)
and activity level. We would also like to include lack 9. Lauer, S.A., Grantz, K.H., Bi, Q., Jones, F.K., Zheng, Q.,
of stochasticity as a limitation to our modeling frame- Meredith, H.R., Azman, A.S., Reich, N.G., Lessler, J.: The
work. Despite all these limitations, we believe that our incubation period of coronavirus disease 2019 (covid-19)
modeling framework has provided some insights into from publicly reported confirmed cases: estimation and
application. Annal. Intern. Med. 172(9), 577–582 (2020)
the endemic dynamics of COVID-19 and has also pro- 10. Lei, S., Jiang, F., Wating, S., Chen, C., Chen, J., Mei, W.,
vided some reasonable suggestions on the important Zhan, L.-Y., Jia, Y., Zhang, L., Liu, D., et al.: Clinical char-
acteristics and outcomes of patients undergoing surgeries

123
Assessing the potential impact of immunity waning

during the incubation period of covid-19 infection. EClini- 25. Saqib, M.A.N., Siddiqui, S., Qasim, M., Jamil, M.A.,
calMedicine 21, 100331 (2020) Rafique, I., Awan, U.A., Ahmad, H., Afzal, M.S.: Effect of
11. New york times. They Were Infected With the Coro- covid-19 lockdown on patients with chronic diseases. Dia-
navirus. They Never Showed Signs., (accessed August betes Metabolic Syndr. Clinic. Res. Rev. 14(6), 1621–1623
29, 2021). https://www.nytimes.com/2020/02/26/health/ (2020)
coronavirus-asymptomatic.html 26. Henderson, M., Bragg, A., Fahim, G., Shah, M., Hermes-
12. Ganyani, T., Kremer, C., Chen, D., Torneri, A., Faes, C., DeSantis, E.R.: A review of the safety and efficacy of vac-
Wallinga, J., Hens, N.: Estimating the generation interval cines as prophylaxis for clostridium difficile infections. Vac-
for coronavirus disease (covid-19) based on symptom onset cines 5(3), 25 (2017)
data, March 2020. Eurosurveillance 25(17), 2000257 (2020) 27. Kash, Natalie, Lee, Michael A., Kollipara, Ramya, Down-
13. Mukoro Duke George: A novel corona virus (covid-19) ing, Christopher, Guidry, Jacqueline, Tyring, Stephen K.:
pandemic, pathogenesis, clinical features and management Safety and efficacy data on vaccines and immunization to
options, public health measures. Int. J. Tropical Dis. Health human papillomavirus. J. Clinic med. 4(4), 614–633 (2015)
31(11–21), 2020 (2020) 28. Klinman, D. M., Takeno, M., Ichino, M., Gu, M.,
14. Desvars-Larrive, A., Dervic, E., Haug, N., Niederkroten- Yamshchikov, G., Mor, G., Conover, J.: Dna vaccines: safety
thaler, T., Chen, J., Di Natale, A., Lasser, J., Gliga, D.S., and efficacy issues. In: Springer seminars in immunopathol-
Roux, A., Sorger, J., et al.: A structured open dataset of ogy, vol. 19, 245–256. Springer (1997)
government interventions in response to covid-19. Scien- 29. Krause, P.R., Gruber, M.F.: Emergency use authorization of
tific Data 7(1), 1–9 (2020) covid vaccines-safety and efficacy follow-up considerations.
15. Dzobo M, Chitungo I, Dzinamarira T.: Covid-19: a perspec- N. Engl. J. Med. 383(19), e107 (2020)
tive for lifting lockdown in zimbabwe Pan African Med. J. 30. Greenwood, B.: The contribution of vaccination to global
35(Suppl 2) (2020) health: past, present and future. Philosophical Transactions
16. Haug, N., Geyrhofer, L., Londei, A., Dervic, E., Desvars- R. Soc. B Biol. Sci. 369(1645), 20130433 (2014)
Larrive, A., Loreto, V., Pinior, B., Thurner, S., Klimek, P.: 31. Relief Web. COVAX Expects to Start Sending Millions
Ranking the effectiveness of worldwide covid-19 govern- of COVID-19 Vaccines to Africa in February, (accessed
ment interventions. Nat. Human Behav. 4(12), 1303–1312 on 29 August 2021). https://reliefweb.int/report/world/
(2020) covax-expects-start-sending-millions-covid-19-vaccines-
17. McCoy, L.G., Smith, J., Anchuri, K., Berry, I., Pineda, J., africa-february
Harish, V., Lam, A.T., Yi, S.E., Sophie, Hu., Rosella, L., 32. World Health Organization (WHO). First COVID-19
et al.: Characterizing early canadian federal, provincial, ter- COVAX vaccine doses administered in Africa, (accessed
ritorial and municipal nonpharmaceutical interventions in September 16, 2021). https://www.who.int/news/item/
response to covid-19: a descriptive analysis. CMAJ Open 01-03-2021-first-covid-19-covax-vaccine-doses-
8(3), E545 (2020) administered-in-africa
18. Mhango, M., Chitungo, I., Dzinamarira, T.: Covid-19 lock- 33. Madhi, S.A., Baillie, V., Cutland, C.L., Voysey, M., Koen,
downs: impact on facility-based hiv testing and the case for A.L., Fairlie, L., Padayachee, S.D., Dheda, K., Barnabas,
the scaling up of home-based testing services in sub-saharan S.L., Bhorat, Q.E. , et al.: Safety and efficacy of the chadox1
africa. AIDS Behav. 24, 3014–3016 (2020) ncov-19 (azd1222) covid-19 vaccine against the b. 1.351
19. Singh, S., Shaikh, M., Hauck, K., and Miraldo, M.: Impacts variant in South Africa. MedRxiv (2021)
of introducing and lifting nonpharmaceutical interventions 34. Mahase, E.: Covid-19: South africa pauses use of oxford
on covid-19 daily growth rate and compliance in the united vaccine after study casts doubt on efficacy against variant
states. Proceedings of the National Academy of Sciences, (2021)
118(12) (2021) 35. Maverick. D., South Africa to Give First COVID-19
20. Monika Chaudhary, P.R., Sodani, P.R., Das, S.: Effect of Vaccine Doses to President, Health Workers., (accessed on
covid-19 on economy in India: some reflections for policy 29 August 2021). https://www.dailymaverick.co.za/article/
and programme. J. Health Manag. 22(2), 169–180 (2020) 2021-02-17-south-africa-to-give-first-covid-19-vaccine-
21. Fontanesi, L., Marchetti, D., Mazza, C., Di Giandomenico, doses-to-presidenthealth-workers/
S., Roma, P., Verrocchio, M.C.: The effect of the covid-19 36. Anderson, S.C., Edwards, A.M., Yerlanov, M., Mulberry, N.,
lockdown on parents: a call to adopt urgent measures. Psy- Stockdale, J.E., Iyaniwura, S.A., Falcao, R.C., Otterstatter,
chol. Trauma Theory Res. Pract. Policy 12(S1), S79 (2020) M.C., Irvine, M.A., Janjua, N.Z.: Coombs D Quantifying
22. Inoue, H., Todo, Y.: The propagation of economic impacts the impact of covid-19 control measures using a bayesian
through supply chains: The case of a mega-city lockdown to model of physical distancing. PLoS Comput. Biol. 16(12),
prevent the spread of covid-19. PloS One 15(9), e0239251 e1008274 (2020)
(2020) 37. Bulut, H., Gölgeli, M., Atay, F.M.: Modelling personal cau-
23. Magamela, M.R., Dzinamarira, T., Hlongwa, M.: Covid-19 tiousness during the covid pandemic: a case study for turkey
consequences on mental health: an African perspective. S. and italy. Nonlinear Dyn. 11, 1–13 (2021)
Afr. J. Psychiatry 27(1), 1–2 (2021) 38. He, S., Peng, Y., Sun, K.: Seir modeling of the covid-19 and
24. Pierre, G., Uwineza, A., Dzinamarira, T.: Attendance to hiv its dynamics. Nonlinear Dyn. 101(3), 1667–1680 (2020)
antiretroviral collection clinic appointments during covid- 39. Ivorra, B., Ferrández, M.R., Vela-Pérez, M., Ramos, A.M.:
19 lockdown a single center study in kigali, rwanda. AIDS Mathematical modeling of the spread of the coronavirus dis-
Behav. 24(12), 3299–3301 (2020) ease taking into account the undetected infections the case of

123
 M. Rabiu, S. A. Iyaniwura

china. Commun. Nonlinear Sci. Numer. Simul. 88, 105303 et al.: Waning antibody responses in asymptomatic and
05303 105303 (2020) symptomatic sars-cov-2 infection. Emerg. Infect. Dis. 27(1),
40. Iyaniwura, S. A., Rabiu, M. A., David, J. and Kong, J. D.: 327 (2021)
Assessing the impact of adherence to non-pharmaceutical 56. Goldberg, Y., Mandel, M., Bar-On, Y.M., Bodenheimer, O.,
interventions and indirect transmission on the dynamics of Freedman, L.S., Haas, E., Milo, R., Alroy-Preis, S., Ash, N.,
covid-19: a mathematical modelling study. medRxiv. (2021) Huppert, A.: Waning immunity of the bnt162b2 vaccine: a
41. Kucharski, A.J., Russell, T.W., Diamond, C., Liu, Y., nationwide study from israel. medRxiv (2021)
Edmunds, J., Funk, S., Eggo, R.M., Sun, F., Jit, M., Mun- 57. Hamady, A., Lee, J., Loboda, Z.A.: Waning antibody
day, J.D., et al.: Early dynamics of transmission and control responses in covid-19: what can we learn from the analy-
of covid-19: a mathematical modelling study. Lancet Infect. sis of other coronaviruses? Infection 29, 1–15 (2021)
Dis. 20(5), 553–558 (2020) 58. Shrotri, M., Navaratnam, A.M.D., Nguyen, V., Byrne, T.,
42. Kwuimy, C., Nazari, F., Jiao, X., Rohani, P., Nataraj, C.: Geismar, C., Fragaszy, E., Beale, S., Fong, W.L.E., Patel, P.,
Nonlinear dynamic analysis of an epidemiological model for Kovar, J., et al.: Spike-antibody waning after second dose
covid-19 including public behavior and government action. of bnt162b2 or chadox1. The Lancet 398(10298), 385–387
Nonlinear Dyn. 101(3), 1545–1559 (2020) (2021)
43. Liu, Z., Magal, P., Seydi, O., Webb, G.: A covid-19 epidemic 59. STATISTICS SOUTH AFRICA. STATISTICAL RELEASE,
model with latency period. Infect. Dis. Model. 5, 323–337 Mid-year population estimates, 2019, (accessed August
(2020) 28, 2021) http://www.statssa.gov.za/publications/P0302/
44. Musa, R., Ezugwu, A.E. and Mbah, G.C.: Assessment of the P03022019.pdf
impacts of pharmaceutical and non-pharmaceutical inter- 60. Webometer. Total Coronavirus Deaths in South Africa,
vention on covid-19 in south africa using mathematical (accessed August 28, 2021). https://www.worldometers.
model medRxiv (2020) info/coronavirus/country/south-africa/
45. Peter, O.J., Shaikh, A.S., Ibrahim, M.O., Nisar, K.S., 61. Yang, Z., Zeng, Z., Wang, K., Wong, S.-S., Liang, W., Zanin,
Baleanu, D., Khan, I. and Abioye, A.I.: Analysis and dynam- M., Liu, P., Cao, X., Gao, Z., Mai, Z., et al.: Modified seir
ics of fractional order mathematical model of covid-19 in and ai prediction of the epidemics trend of covid-19 in china
nigeria using atangana-baleanu operator. Computers Mater. under public health interventions. J. Thorac. Dis. 12(3), 165
Continua. 66(2), (2020) (2020)
46. Peter, O.J., Qureshi, S., Yusuf, A., Al-Shomrani, M., Idowu, 62. Zou, L., Ruan, F., Huang, M., Liang, L., Huang, H., Hong,
A.A.: A new mathematical model of covid-19 using real data Z., Jianxiang, Yu., Kang, M., Song, Y., Xia, J., et al.: Sars-
from pakistan. Results Phys. 24, 104098 (2021) cov-2 viral load in upper respiratory specimens of infected
47. Rajagopal, K., Hasanzadeh, N., Parastesh, F., Hamarash, I.I., patients. N. Engl. J. Med. 382(12), 1177–1179 (2020)
Jafari, S., Hussain, I.: A fractional-order model for the novel 63. Tindale, L., Coombe, M., Stockdale, J.E., Garlock, E., Lau,
coronavirus (covid-19) outbreak. Nonlinear Dyn. 101(1), W.Y.V., Saraswat, M., Lee, Y.H.B., Zhang, L., Chen, D.,
711–718 (2020) Wallinga, J.: et al. Transmission interval estimates suggest
48. Chuang, X., Pei, Y., Liu, S., Lei, J.: Effectiveness of non- pre-symptomatic spread of covid-19 MedRxiv (2020)
pharmaceutical interventions against local transmission of 64. Jassat, W., Mudara, C., Ozougwu, L., Tempia, S., Blumberg,
covid-19: an individual-based modelling study. Infect. Dis. L., Davies, M. A., Pillay, Y., Carter, T., Morewane R., Wol-
Model. 6, 848–858 (2021) marans M., et al. Difference in mortality among individuals
49. Zhong, H., Wang, W.: Mathematical analysis for covid-19 admitted to hospital with covid-19 during the first and sec-
resurgence in the contaminated environment. Math. Biosci. ond waves in south africa: a cohort study. The Lancet Global
Eng. 17(6), 6909–6927 (2020) Health (2021)
50. Bugalia, S., Bajiya, V.P., Tripathi, J.P., Li, M.-T., Sun, G.-Q.: 65. Peng, L., Yang, W., Zhang, D., Zhuge, C. and Hong, L.: Epi-
Mathematical modeling of covid-19 transmission: the roles demic analysis of covid-19 in china by dynamical modeling.
of intervention strategies and lockdown. Math. Biosci. Eng. arXiv preprintarXiv:2002.06563 (2020)
17(5), 5961–5986 (2020) 66. Zhao, Z., Li, X., Liu, F., Zhu, G., Ma, C., Wang, L.: Predic-
51. Iboi, E.A., Ngonghala, C.N., Gumel, A.B.: Will an imperfect tion of the covid-19 spread in african countries and implica-
vaccine curtail the covid-19 pandemic in the us? Infect. Dis. tions for prevention and control: a case study in south Africa,
Model. 5, 510–524 (2020) Egypt, Algeria, Nigeria, Senegal and Kenya. Sci. Total Env-
52. Mushayabasa, S., Ngarakana-Gwasira, E.T., Mushanyu, J.: iron. 729, 138959 (2020)
On the role of governmental action and individual reaction 67. Diekmann, O., Heesterbeek, J.A.P., Metz, J.A.J.: On the def-
on covid-19 dynamics in south africa: a mathematical mod- inition and the computation of the basic reproduction ratio
elling study. Inform. Med. Unlocked 20, 100387 (2020) r 0 in models for infectious diseases in heterogeneous pop-
53. Sadarangani, M., Raya, B.A., Conway, J.M., Iyaniwura, ulations. J. Math. Biol. 28(4), 365–382 (1990)
S.A., Falcao, R.C., Colijn, C., Coombs, D., Gantt, S.: Impor- 68. Van den Driessche, P., Watmough, J.: Reproduction num-
tance of covid-19 vaccine efficacy in older age groups. Vac- bers and sub-threshold endemic equilibria for compartmen-
cine 39(15), 2020–2023 (2021) tal models of disease transmission. Math. Biosci. 180(1–2),
54. Johnston, M. D., Pell, B.: A dynamical framework for mod- 29–48 (2002)
eling fear of infection and frustration with social distancing 69. Rabiu, M., Willie, R., Parumasur, N.: Analysis of a
in covid-19 spread. arXiv preprintarXiv:2008.06023 (2020) virus-resistant hiv-1 model with behavior change in non-
55. Choe, P.G., Kang, C.K., Suh, H.J., Jung, J., Song, K.-H., progressors. Biomath 9(1), 2006143 (2020)
Bang, J.H., Kim, E.S., Kim, H.B., Park, S.W., Kim, N.J.,

123
Assessing the potential impact of immunity waning

70. Ayinla, A.Y., Othman, W.A., Rabiu, M.: A mathematical 75. Republic of South Africa Department of health.
model of the tuberculosis epidemic. Acta Biotheoretica 20, COVID-19 Statistics in South Africa, (accessed August
1–31 (2021) 29, 2021). https://sacoronavirus.co.za/2021/08/29/
71. Guckenheimer, J., Holmes, P.: Nonlinear Oscillations, update-on-covid-19-sunday-29-august-2021/
Dynamical Systems, and Bifurcations of Vector Fields, vol. 76. Republic of South Africa Department of health. COVID-19
42. Springer, New York (2013) Total Vaccination Summary, (accessed August 29, 2021).
72. Rabiu, M., Willie, R., Parumasur, N.: Mathematical analysis https://sacoronavirus.co.za/latest-vaccine-statistics/
of a disease-resistant model with imperfect vaccine, quaran-
tine and treatment. Ricerche di Matematica 69(2), 603–627
(2020) Publisher’s Note Springer Nature remains neutral with regard
73. Castillo-Chavez, C., Song, B.: Dynamical models of tuber- to jurisdictional claims in published maps and institutional affil-
culosis and their applications. Math. Biosci. Eng. 1(2), 361 iations.
(2004)
74. Republic of South Africa Department of Statistics.
Mid-year population estimates, 2021, (accessed August
29, 2021). http://www.statssa.gov.za/publications/P0302/
P03022021.pdf

123