Abstract
Introduction As SARS-COV-2 and the disease COVID-19 is sweeping through countries after countries around the globe, it is critical to understand potential burden of a future outbreak in cities of Colombia. This pandemic has affected most of the countries in the world because the high global movement of individuals and excessive cost in interventions.
Objective Using demographic data from city of Cali, disease epidemiological information from affected countries and mathematical models, we estimated the rate of initial exponential growth of new cases and the basic reproductive rate for a potential outbreak in city of Cali in Colombia.
Materials and methods We used dynamical models with different modeling assumptions such as use of various types of interventions and/or epidemiological characteristics to compare and contrast the differences between Colombian cities and between Latin American countries.
Results Under the assumption of homogeneously mixing population and limited resources, we predicted expected number of infected, hospitalized, in Intensive Care Units (ICU) and deaths during this potential COVID-19 outbreak. Our results suggest that on a given day in Cali there may be up to around 73000 cases who might need hospitalization under no intervention. However, this number drastically reduces if we carry out only-isolation intervention (with 16 days of symptomatic infection; ~13,000 cases) versus both quarantining for 6 days and isolation within 16 days (~3500 cases). The peak in Cali will reach in 2-3 months.
Conclusions The estimates from these studies provides different scenarios of outbreaks and can help Cali to be better prepared during the ongoing COVID-19 outbreak.
1. Introduction
An outbreak of pneumonia with unknown etiology in the Wuhan city, Hubei’s province in Mainland China was first reported on December 31, 2019 by World Health Organization (WHO). On January 7, 2020 a virus associated with this outbreak was isolated [1] and found to be novel strain of Coronavirus that may have crossed over to human population from its sylvatic host. On January 30, 2020, WHO declared the Public Health emergency of International importance [2] and denominated this coronavirus as SARS-COV-2 while the disease was referred as COVID-19 on February 12, 2020 [3]. This virus belongs to the genus betacoronavirus of the subfamily coronavirus and family Coronaviridae. 2002-2004 SARS epidemic and 2012-2016 MERS epidemic were also caused by the viruses belonging to the family coronavirus, however, both showed lower transmissibility and absolute lethality [4].
In order to protect people from this unknown outbreak of COVID-19 it is critical to understand its trends and estimate potential burden. Mathematical models have been used before to this purpose for various epidemic diseases and can be extremely helpful in disease preparedness. An important threshold quantity associated with a disease transmissibility is the basic reproduction number, denoted by R0 (pronounced “R naught”). The epidemiological definition of R0 is the average number of new cases of the disease that will be generated by one contagious person during his/her infectious period. It specifically applies to a population of people who were previously free of infection and not vaccinated. Three possibilities exist for the potential spread or decline of a disease, depending on its R0 value: (i). If R0 is less than 1, each existing infection causes less than 1 new infection. In this case, the disease will decline and eventually disappear. (ii). If R0 equals 1, the disease will stay alive, but there won’t be an epidemic. (iii). If R0 is greater than 1, cases could grow exponentially and cause an epidemic or even a pandemic [5]. A preliminary R0 estimate of 1.4 - 2.5 was presented on Jan 23, 2020 in a WHO’s statement regarding the outbreak of 2019-nCoV [1]. S. Zhao et al. [6] estimated the mean R0 for 2019-nCoV in the early phase of the outbreak ranging from 3.3 to 5.5 (likely to be below 5 but above 3 with rising report rate) [6,7], which appeared slightly higher than those of SARS-CoV (R0:2−5) [4]. J. Read et al. [44] estimated the R0 for 2019-nCoV to be in the range 3.6 - 4.0, indicating that 72-75% of transmissions must be prevented in order to stop the increasing trend. The explosive and dramatic behavior of the spread of this virus SARS-CoV-2 in mainland China, has forced health and administrative authorities take severe and strict control measures. Until January 31, 2020, 11950 cases had been reported and 259 deaths but the number of cases had increased to 75,184 and 2009 deaths on February 18, 2020 [8]. The pharmacological treatment with antivirals is unknown and the vaccine is in the early stages of investigation.
Cali is a city with a population of about 2.5 million (2,492,442 people, 98.4% in urban areas and 1.6% in rural areas) [9], located at 3°27’26’’ of latitude North and 76°31’42’’ of longitude West (Greenwich Meridian), at an altitude of 1,070 meters above the sea level, with an average temperature of 24.7°C, an annual precipitation of 1,019.2 mm 12. Cali presents two rainy seasons during the year [10], the first one between the months March and May, and then the second one rainy period occurs between September and November when the temperatures come down. Other strains of Coronavirus are known to be significantly related to climatic conditions with colder weather more suitable for their spread. It is worth considering that the covid-19 epidemic in Mainland China began in winter, the period of year with lower temperature and that in Cali the first winter season is approaching now in April/May when the temperature drops, and the population tends to be more crowded. Cali is a touristic city and the number of passengers moved by international routes in Colombia [11] reached 11,527,416 people alone in the first half-year. The countries of origin of these flights were of North America with 3,108,146 people, Europe 4,024,519 people, South America 2,608,066 people. Between the factors [12] explain the emergence or re-emergence of infectious diseases the modern aeronautical technology allows the connection between countries in very short time and thus sick people carry infectious agents over long distances [13]. This information will allow prepare plans of education to community that interrupts the transmission of the virus.
The objectives of this study are to (i) use available current COVID-19 information from other countries and estimate, for the city of Cali, Colombia, the expected number of cases that needs to be hospitalized over an outbreak, maximum daily number of cases that can be expected, time needed to reach the epidemic peak and mean duration of an epidemic for the current outbreak in the presence of different control actions by the public health department and different modeling assumptions, (ii) compare situation in Cali with other four major cities of Colombia, and (iii) compare the situation in Colombia with other similar countries.
2. Materials and Methods
We considered different modeling assumptions to capture COVID-19 outbreak in Cali and four other cities of Colombia. In order to understand the ongoing burden of the disease, a systematic procedure was implemented. The procedure in this study primarily involves four steps: (i) Using reported estimates of R0 and model-derived formulation of R0 (corresponding to different model and under the assumption of initial exponential growth rate of epidemic), estimate effective transmission rate in the population, (ii) Use estimated effective transmission rate, demographic data from city of Cali and epidemic models, to predict different outbreak scenarios for Cali, and (iii) From these outbreak scenarios (in step (ii)), estimate the total number of cases, hospitalizations, time needed to reach the peak and expected duration of the outbreak.
Basic Modeling Framework
Mathematical Model to Capture Potential Epidemic in Cali, Colombia
For the mathematical modeling of the transmission dynamics of the new SARS-CoV-2 coronavirus, we have established the structure compartmental of SEIR model (Susceptible, Exposed, Infectious, Recovered) and applying the method of differential equations 5 based on the COVID-19 epidemic data in continental China [14,15].
Model variables and parameters are defined in Table 1. The rest of the models are defined in Table 2 and in the Appendix, section B.
Variables and parameters of the SEIR model
Formulas to the estimate basic reproduction number
Model-related metrics
Consider a simple epidemic model, Susceptible-Infectious-Recovered (SIR) or Susceptible-Exposed-Infectious-Recovered (SEIR) model with I individuals are only infectious. The number of susceptible individuals who are infected per unit time is given by the product of the force of infection, λ(t), and the number of susceptible individuals at time t (i. e., λ(t)S(t)) under the assumption that individuals contact each other randomly. Let λ(t) is the force of infection or rate at which a susceptible individual is infected or rate at which they move to an infected compartment then it can be written as:
where β is the rate at which two specific individuals come into effective contact per unit time and I(t) is the number of infectious individuals at time t. The formula for basic reproduction number of the model is R0 = βND and can used it to estimate effective contact rate as
where N is the population size at the beginning of an outbreak and D is the length of the infectious period.
Now if Exposed state (defined as individuals who are infected not infectious) is present in the model then the model will be referred as SEIR model. Suppose f is the rate at which the individuals move from exposed compartment to infectious compartment. The rate f = 1/D′, where D′ is the duration of the latent or pre infectious period. Let r be the rate at which the individuals move from infectious compartment to recovered or immune compartment, i.e., r = 1/D, where D is the duration of the infectious period.
Since R0 is a metric computed early in the epidemic, it can be used to relate it with the initial growth rate of an epidemic. Different formulae are used to estimate R0 using data from the early stages of an epidemic, each of them requires estimates of something which is often referred to as exponential growth of rate of the epidemic sometimes denoted by Λ (capital lambda). During the early stages of an epidemic, the number of infectious individuals increases at an approximately constant exponential growth rate.
Estimation of Various Relevant Quantities
Initial Exponential Growth Rate
We can estimate this growth rate as follows. Consider an exponential growing function of number of reported cases I(t) where Λ is the exponential growth rate parameter, i.e.,
where I(0) is the initial number of infectious individuals at the beginning of infection. If we take natural logarithm of this equation, we obtain the following equation
relating the number of infectious individuals and Λ. Note, this is the equation of a straight line with slope Λ, suggesting that if we plot the natural logarithm of the number of infectious individuals against time, we should obtain a straight line with slope Λ.
On the other hand, we can also write an expression of R0, in terms of Λ. This expression of R0 will depend on the type of model considered and other model parameters including D’ and D the average durations of the pre-infectious and infectious periods respectively. Sometimes R0 can also be expressed in terms of the serial interval, Ts (defined as the time between the start of symptoms in the primary patient (infector) and onset of symptoms in the patient receiving that infection from the infector (the infected) [7]. In Table 2 the different formulas are shown. For example, in case of the SEIR model with only I infectious, the basic reproduction number can be written as
and in case of the SIR model, R0 = 1 + ΛD.
Furthermore, we can use equation (8) to obtain the relationship between the growth rate Λ and the doubling time Td of an epidemic, which is defined as the time until the number of cases in the population doubles, relative to that at some other time. Suppose that there is only one infectious individual at time t = 0 (I(0) = 1) and there are two infectious individuals at time t = Td (I(Td) = 2). Substituting for I(0) = 1 and I(Td) = 2 into equation (8) we obtain ln2 = ln1 + ΛTd which implies
Since R0 can be written in the form of Λ and the doubling time can be computed as in equation (10), we can find Td in terms of R0. For example, in case of SIR model,
and in case of SEIR model with only I infectious, it can be written as
where 
, D1= D + D′ and D2 = D * D′. This formula can be used to study the role of epidemiological factors such as transmission rate (β), asymptomatic and symptomatic periods.
Estimation of Model Parameters
In groups of Japanese migrants who were repatriated, the proportion of positive asymptomatics to PCR test, was estimated by Hiroshi Nishiura et al, in 41.6% (CI 95%: 16.7-66.7) [16]; Anne Kimball et al, [17] found 57.0% and Kenji Mizumoto et al 51.7% [18]. For purpose of our study we will take 50 % as parameter value [19]. It has also been established that 14% of the symptomatic people are hospitalized for complications related to pneumonia and respiratory distress [20,21]. From those hospitalized, 15% die [22-24]. Because the covid-19 is novel disease, some parameters are unknown or are known with less precision, and hence, they are estimated. Using case reports, outbreaks’ studies and some based on the behavior of coronavirus in past epidemics, these parameters are estimated. Multiple research articles report incubation period estimates obtained from different methods. The most of them report incubation period mean between 5 – 7 days and the range between 2 – 14 days [25-34]. We have used the incubation period to be 6.4 days based on most of the studies. Although the virus SARS-2 can be detected in nasopharyngeal swab between 2.5 days before and 18 days after starting symptoms, infectivity is considered very low after 7 days. The infectious period is assumed for this study, an average of 7.6 days based on reported 14 days quarantine period COVID-19 [35-39].
Estimation of R0
The data used for calculating R0 were collected from published research in the New England journal of Medicine on January 29, 2020 [14]. The novel coronavirus SARS-CoV-2, causes COVID-19, which is characterized mainly by fever, muscle pain and cough and it can move along to severe phases of pneumonia or even die. Because it is a new disease, there is a small knowledge about the time of latent period, infectious period in the natural history of disease, though some estimates have been made based on cluster of cases and hospitalized.
Models
In this study, we consider various modeling assumptions leading to seven different models, six of them are defined in Table 2 and their details are given in Appendix B. The last (seventh) model is referred as Age of infection model, whose some information are given below and other details are given in Appendix B. Suppose S,E,I,R,T,J, Q1, Q2 and Q3 represent Susceptible, Exposed, Infectious, Recovered, Treated, Hospitalized, Susceptible-Quarantined, Exposed-Quarantined and Infectious-Isolated. The models are (i) simple SIR, (ii) simple SEIR with I only infectious, (iii) simple SEIR with E and I infectious, (iv) SITR, (v) SIR with general distribution of age of infection (vi) SEIR-Q1Q2Q3, and (vii) SEIJR.
Age of Infection Model
S(t) and φ(t) denote the number of susceptibles and total infectivity at time t. The total infectivity is measured as the average of product of the number of infected individuals and the mean infectivity of the infected individuals at their age of infection τ where τ ∈ [0,t]. Denoting the mean infectivity at time τ by A(τ), we can write the epidemic model with age of infection as follows
Then the basic reproduction number is defined as
The basic reproduction number in terms of initial growth rate is as follows,
where Λ is a positive eigenvalue satisfying the relation
Interventions
The modeling assumptions in the models were related to epidemiological characteristics and/or type of interventions. The interventions that we have considered include quarantine, isolation, treatment, or infectious hospitalized individuals with mild infection.
We also consider three different types of interventions [40] such as
“Wuhan-style” Containment: It assumes closed borders with full shutdown and no movement of individuals (first reporting Jan 1 in Wuhan and this type of lockdown started on Jan 23). In order to capture this intervention, the R0 was assumed as follows
“Shelter-in-place” Containment: It assumes voluntary community-wide home quarantine with only essential services open. In order to capture this intervention, the R0 was assumed as follows
“Social Distancing” Order: It assumes ban on events over 50 people and public advocacy around “social distancing”. In order to capture this intervention, the R0 was assumed as follows
Using these estimates of R0 and the formula for R0 for the SEIR model with both E and I infectious, it is possible to readily obtain corresponding β estimates (transmission rate over time). For intervention β will be also piecewise defined function.
Three different types of transmission metrics corresponding to the three different interventions
3. Results
We assume that the initial values for simulations were I0 = 1,S0 = N − I0 and other state variables zero at the starting time. The parameter estimates of the models were collected in Table 3 and Table 4.
Date of reporting starts and first exponential day
SEIR model parameters estimate for R0 = 4.95
Initial Exponential Growth Rate
Exponential growth model was used to estimate initial exponential growth rate in city of Cali. The exponential function (Equation (7) or (8)) is fitted to reported cases from different countries. The early estimated growth rate using data from China is found to around 0.17 (Figure 2) and this is used for further simulation scenarios. The confidence interval is given by 95% CI (0.162, 0.185). The comparison of different exponential growth and estimated start of epidemics in some countries are computed (Table 3).
Growth rate of Covid-19 in initial epidemic phase
Outbreak in City of Cali, Colombia
SEIR model with I Infectious
Using literature, we can estimate some of the model parameters. We obtained Th (the hospitalization rate) around 14%, Lh (the lethality in hospitalized individuals) around 15%, the estimated Ts = 7.5 days (95% CI, 5.3 to 19) and the estimated Td = 7.4 days (95% CI, 4.2 to 14) (6).
It is assumed that the individuals contact randomly and furthermore the controls are not implemented. We carried out simulations with the population size of Cali and it was applied the hospitalization rate (Th) and the lethality rate (Lh) in hospitalized individuals reported from Wuhan.
Based on the data for confirmed cases from December 08, 2019 through January 21, 2020 from city of Wuhan, China [14], we estimated the growth rate [15] of natural logarithm of the incidence accumulate be 0.175 and the corresponding R0 was 4.95.
Comparison of SIR, SEIR with I only infectious and SEIR with E and I both Infectious models
Various outbreak related-metrics including number of cases and deaths for the city of Cali are given in Table 5 when there are no interventions. If we assume that infections can be generated by asymptomatic then the outbreak will persist for relatively longer time (around 8 months versus 7 months). The total number of hospitalized cases during the outbreak between the three models are presented in Table 6 for 30, 60 and 90 days after start of epidemic, respectively.
Description of parameters and values for all other models in Table 2.
Scenario for different models, SIR, SEIR with only I infectious, and SEIR with E and I both Infectious. For each model, the first line represents 30 days after start of epidemic, the second line 60 days from beginning and the third line have values after 90 days in epidemic. The number of hospitalized is calculated using the hospitalization rate, Th, of 14% [15] and number of deaths is calculated using the COVID-19 mortality rate, Lh, of 15% among hospitalized individuals [15].
Simulation for SEIR model with E and I both infectious
Comparison of Outbreak burden using SEIR Model with no intervention, only isolation and both quarantine and isolation interventions
We consider SEIR-Q1Q2Q3 model (see Appendix, section B6 for details) under three different scenarios: baseline (no intervention), with only isolation and with both quarantine and isolation (Figure and Table shows the statistics and trends).
SEIR-Q1Q2Q3 Model results. Showing total cases and effective reproduction number
SEIR-Q1Q2Q3 Model results
SEIR-J.
Comparison Between 5 Big Cities of Colombia
Comparison of SEIR with E and I both Infectious and quarantine and isolation interventions models for 5 different cities of Colombia
The difference between outbreaks under similar level of quarantine and isolation interventions are shown in Table 8 and Figure 6. Cartagena and Barranquilla will have roughly similar type of an outbreak and Cali and Bogota will have similar. Maximum number of a beds that will be needed on a given day in Bogota and Cartagena will be 3 and 1/2 times, respectively, the corresponding number in Cali. Cali might need up to 3502 beds a day for the COVID-19 patients at the time of the peak of the outbreak, which will last for 7 to 8 months.
Expected number and 95% confidence interval
Comparison of outbreaks in 6 major cities of Colombia using SEIR_Q1Q2Q3 model
SEIR-Q1Q2Q3 Model results for Different Cities. Not all symptomatic will be needing hospitalization as some will recover themselves. In all the cities roughly 52% will be infected either as mild or severe infections during the outbreak
4. Discussion
The estimates of R0 considered here might be slightly higher than those published by other authors, from Asia and Europe [21,32,34] but it is noteworthy that during the SARS epidemic in 2002 the estimates of the average R0 varied between 2.24 (95 % CI: 1.96–2.55) and 3.58 (95% CI: 2.89–4.39) [4] and 8,000 cases were reported while for the new SARS-Cov-2 which is a similar coronavirus, originated in the same country the estimates of R0 are similar or even lower than 2.2 (95% CI, 1.4 to 3.9) [14], and 80,000 cases have already been reported in the first 11 weeks of the epidemic.
We have observed the application of different techniques to calculate R0, that could explain the different scenarios of a potential outbreak in Colombia. While Li et al. (14) used the duration of the serial interval, whose result is similar to ours when we use the same technique, our highest estimate was based on the growth rate and the duration of similar pre-symptomatic and infectious periods.
We estimated that in the city of Cali the outbreak under current intervention of isolation and quarantine will last for 5-6 months and will need around 3500 beds on a given during the peak of the outbreak. Outbreak in Cali will be similar to outbreak in Medellin but at least 1/3 times less cases will be observed as compared to the outbreak in Bogota. The cases in the current outbreak in Cali was reported about a week late.
5. Conclusions
COVID-19 is a highly transmissible virus with the capacity to produce outbreaks and with high repercussions on lethality in the vulnerable (individuals with pre-health conditions or elderly) population and with comorbidities.
Mathematical models and simulation applied to data ongoing epidemics allow anticipation in the phase of preparing mitigation plans.
Based on the estimates, the distribution of hospital beds and intensive care units in the city available for response to an eventual emergency have been planned. Also, the inclusion of 500,000 Euros in municipal budget for mitigating the epidemic if it reaches.
Data Availability
We used data obtained from different documents already published online.
Conflict of interest statement
Authors declare do not have any conflict of interest
Appendix
A. Some other figures
Simulation for SIR model
Simulation for SEIR model with only I infectious
SEIR with only I infectious
SEIR with E and I infectious
Exponential fitting for data from South Korea and South Africa
Exponential fitting for data from Mexico and Panama
Exponential fitting for data from six South America countries
B. Models
B.1. SIR Model
We start with a simple SIR epidemic model
with initial conditions S(0) = S0, I(0) = I0, S0 +I0 = N. It is known that the basic reproduction number for this model is 
and in terms of the initial exponential growth rate Λ
For this simple SIR model, if t is small, S ≈ N, and the equation for I is approximately
and grow exponentially with growth rate (R0 − 1)α. The exponential growth rate A can be measured and then we have an estimate
More complicated models are approximated for small t by linear systems, whose solutions have an exponential growth rate given by the largest eigenvalue of the coefficient matrix. Thus for the SEIR model, the initial exponential growth rate Λ < α(R0 − 1) is the (unique if R0 > 1) largest positive eigenvalue of Jacobian at disease free equilibrium.
The final size relation in terms of the initial exponential growth rate is
B.2. SEIR Model
Now we add an exposed class and we have the following model
The basic reproduction number for this model is 
and in terms of the initial exponential growth rate Λ takes the form
For this SEIR model, the initial exponential growth rate A is the largest positive eigenvalue of
The final size relation for this model in terms of the initial exponential growth rate is
If we estimate R0 from other formulas obtain
We have estimated the expected number of 255 infected, 9 hospitalized and 1 death, during the first month of the epidemic if one infected individual with SARS-COV-2 enters Cali and no control would be taken.
B.3. SEIR Model with infectivity in exposed stage
Now we consider the last model with infectivity in exposed stage
In this case the basic reproduction number is 
and in terms of the initial exponential growth rate Λ takes the form
The final size relation in terms of the initial exponential growth rate is
B.4. A model with treatment
Now we add treatment to the basic model at a rate γ. We consider the following assumptions
Treatment moves infectives to a class T with infectivity decreased by a factor δ and with a recovery rate η).
Treatment continues so long as an individual remains infective.
Treatment is beneficial, η > δα.
Integration of the first equation, sum of the of the first two equations and the third equation gives
where the basic reproduction number is 
and in terms of the initial exponential growth rate Λ takes the form
The final size relation in terms of the initial exponential growth rate is
B.5 Age of infection model
The age of infection plays a crucial role to determine the rate of disease propagation due to infecting the susceptible individuals. Let S(t) and φ(t) denote the number of susceptibles and total infectivity at time t. The total infectivity is measured as the average of product of the number of infected individuals and the mean infectivity of the infected individuals at their age of infection τ where τ ∈ [0, t]. Denoting the mean infectivity at time τ by A(τ), we can write the epidemic model with age of infection as follows
Then the basic reproduction number is defined as
From (44) and (45) we can write
Integrating above expression from 0 to ∞ we get,
Interchanging the order of integration in the second term at right hand side, we find
Finally using the straight forward result 
, we get the final size relation as
Note that the initial measure of total infectivity at the age of infection t is given by φ0(t) = (N − S0)A(t).
This part was done following the book by Brauer, Castillo-Chavez and Feng [41].
The age of infection model is given by
And the basic reproduction number is defined as in equation (46). Now, we linearizing the system (47) – (48) around (S,φ) = (N,0) and find,
Assuming solution of the above system of equations of the form u(t) = eλt, v(t) = eλt, we find one negative eigenvalue λ = −βN and the other eigenvalue can be obtained by solving
Eliminating βN between above equation and (46), we find the basic reproduction number in terms of initial growth rate as follows,
where Λ is a positive eigenvalue satisfying the relation (49).
To derive the final size relation, we can divide equation (47) by S(t) and then integrating we get,
Interchanging the order of integration and using the fact that S(−τ)= N, we can write
B.6. First model with quarantine
Now we consider a model for the transmission of an emerging infectious disease like, the population in the dynamic model is divided into seven classes: Susceptible (S), Exposed (E), Infectious (I), Recovered (R), Quarantined (Q1,Q2), and Isolated (Q3). The flow chart for our model is
Then the corresponding system of equations is
The control reproduction number for this model is
Suppose if 
in our system then the control reproduction number becomes
The control reproduction number (58) in terms of the initial exponential growth rate Λ is
B.7. Model with hospitalization and exposed infectiousness
We consider a model for the transmission of an emerging epidemic disease consisting with susceptible (S), exposed (E), infected (I), quarantined (J), and recovered (R), class as follows,
The control reproduction number for this model is
This can be calculated following the approach outlined by Watmogh and P. Ven Den Driesse in Math. Biosci. With the help of following two relevant matrices,
The control reproduction number is the largest eigenvalue of the matrix FV−1. In order to find the final size relation, integrating both sides of equation (60), we find
Integrating equation (63) and using the fact that I0 = 0, I∞ = 0, we get
Next we integrate (64), use J0 = 0, J∞ = 0, and with the help of (68), we find
Adding the equations (61) and (62) and then integrating, we find
Finally, substituting the results (68) and (69) in (67) and then using (70), after some algebraic calculations, we find
Part of the characteristic equation of the Jacobian matrix for the model (60) – (64) evaluated at disease free equilibrium point which is involved with the determination of initial exponential growth rate Λ takes the form
The initial exponential growth rate Λ satisfies above equation and hence we find
From the controlled basic reproduction number we can write 
, where
with A11 = η + κ + μ, A22 =γ1 + α + δ, A33 =γ2 + δ. Using these expressions now we can write (71) as follows,
Substituting 
in above equation and then after some algebraic calculation, we find
References
