#install.packages("deSolve")
library(deSolve)
##
## Attaching package: 'deSolve'
##
## The following object is masked from 'package:graphics':
##
## matplot
seir_model = function (current_timepoint, state_values, parameters)
{
# create state variables (local variables)
S = state_values [1] #Initial number of susceptible individuals
E = state_values [2] #Initial number of exposed individuals
I = state_values [3] #Initial number of infectious individuals
R = state_values [4] #Initial number of recovered individuals
with (
as.list (parameters), # variable names within parameters can be used
{
#compute derivatives
dS = (-beta * S * I)
dE = (beta * S * I) - (delta * E)
dI = (delta * E) - (gamma * I)
dR = (gamma * I)
# combine results
results = c (dS, dE, dI, dR)
list (results)
}
)
}
Defining parameters
infectious_period = 4.5 # infectious period
latent_period = 1.9 # latent period
transmission_rate = beta_value = 0.7 # Transmission rate
vaccine_compliance = 0 # Vaccine compliance at the base case
vaccine_efficacy = 0.5 # Vaccine efficacy
vac.total = vaccine_compliance * vaccine_efficacy
Compute values for delta and gamma
beta_value = transmission_rate
delta_value = 1 / latent_period
gamma_value = 1 / infectious_period
Compute R0
R0 = beta_value / gamma_value
Disease dynamics parameters
parameter_list = c (beta = beta_value, gamma = gamma_value, delta = delta_value)
Defining the initial condition
X = 9999 #Initial number of susceptible individuals
V = 0 #Initial number of exposed inidviduals
Y = 1 #Initial number of infectious individuals
Z = 0 #Initial number of recovered individuals
Compute total population
N = X + V + Y + Z
Initial state values for the differential equations
initial_values = c (S=(X*(1-vac.total))/N,E=V/N,I=Y/N,R=(Z+X*vac.total)/N)
Output timepoints
timepoints = seq (1, 100, by=1)
Simulate the SEIR epidemic
output = lsoda (initial_values, timepoints, seir_model, parameter_list)
Base case plot
plot (S~time, data=output, type="l", lwd=2, col=1, ylab="", ylim=c(0,1), main="SEIR Epidemic: Base Case")
lines (E~time, data=output, type="l", lwd=2, col=2, ylab="")
lines (I~time, data=output, type="l", lwd=2, col=3, ylab="")
lines (R~time, data=output, type="l", lwd=2, col=4, ylab="")
legend ("topright", legend=c("S","E","I","R"), col=c(1,2,3,4), lwd=2, cex=0.7)
Attack rate and total health care cost of treatment (base case)
treatment_cost = 500 # per infected person
Attack_rate = sum (output[,4]) / (1/gamma_value)
basecase_cost = treatment_cost * Attack_rate * N
#Summary
print (paste("Base case attack rate=", round(Attack_rate,3)))
## [1] "Base case attack rate= 0.95"
print (paste("Base case cost = $", round(basecase_cost,3)))
## [1] "Base case cost = $ 4748966.747"
COMMENT
The attack rate of the base case in the outbreak is 0.95 (95%) and the total treatment cost is $4,748,966.747
Vaccine Intervention
vaccine_compliance = 0.4
vaccine_efficacy = seq (0.4, 0.6, 0.2) # Vaccine efficacies
cost_per_vaccine = 20 # average vaccination cost per person
par(mfrow=c(2,2))
for (i in vaccine_efficacy) {
vaccine_efficacy = i
vac.total = vaccine_efficacy * vaccine_compliance
initial_values = c(S=(X*(1-vac.total))/N,E=V/N,I=Y/N,R=(Z+X*vac.total)/N)
output = lsoda (initial_values, timepoints, seir_model, parameter_list)
# plot intervention
plot (S~time, data=output,type="l", lwd=2, col=1, ylab="", ylim=c(0,1), main=paste("vaccine efficacy=",i))
lines (E~time, data=output, type="l", lwd=2, col=2, ylab="")
lines (I~time, data=output, type="l", lwd=2, col=3, ylab="")
lines (R~time, data=output, type="l", lwd=2, col=4, ylab="")
legend ("topright", legend=c("S","E","I","R"), col=c(1,2,3,4), lwd=2,cex=0.7)
# Cost
vaccine_cost = cost_per_vaccine * vac.total * X
new_Attack_rate = sum (output[,4])/(1/gamma_value)
new_treatment_cost = treatment_cost * new_Attack_rate * N
total_cost = vaccine_cost + new_treatment_cost
# Incremental cost effectiveness ratio
ICER = (total_cost - basecase_cost) / {(new_Attack_rate * N) - (Attack_rate * N)}
#Summary
print (paste("Base case attack rate =", round(Attack_rate,3),"Base case cost =",round(basecase_cost,3)))
print (paste("Vaccine compliance =",vaccine_compliance, "Vaccine efficacy =",i,", Attack_rate=", round(new_Attack_rate,3), "Total Healthcare Cost= $", round(total_cost,3), "Incremental cost effectiveness ratio = ", round(ICER,3)))
}
## [1] "Base case attack rate = 0.95 Base case cost = 4748966.747"
## [1] "Vaccine compliance = 0.4 Vaccine efficacy = 0.4 , Attack_rate= 0.764 Total Healthcare Cost= $ 3852833.06 Incremental cost effectiveness ratio = 482.763"
## [1] "Base case attack rate = 0.95 Base case cost = 4748966.747"
## [1] "Vaccine compliance = 0.4 Vaccine efficacy = 0.6 , Attack_rate= 0.666 Total Healthcare Cost= $ 3378568.928 Incremental cost effectiveness ratio = 483.081"
COMMENT
We register an attack rate of 0.764 (76.4%) when the vaccine efficacy is 40%. Total health care cost is $3,852,833.06 and ICER is $482.76 per case averted. When the vaccine efficacy is 60%, we register an attack rate of 0.666 (66.6%). In this case, the total health care cost is $3,378,568.92 and ICER is $483.08 per case averted.
vaccine_compliance = seq (0.5, 0.6, 0.1) # Vaccine compliances
vaccine_efficacy = seq (0.4, 0.6, 0.2) # Vaccine efficacies
cost_per_vaccine = 20 # Average vaccination cost per person
par(mfrow=c(2,2))
for (i in vaccine_efficacy)
for (ii in vaccine_compliance) {
vaccine_efficacy = i
vaccine_compliance = ii
vac.total = vaccine_efficacy * vaccine_compliance
initial_values = c(S=(X*(1-vac.total))/N,E=V/N,I=Y/N, R=(Z+X * vac.total)/N)
output = lsoda (initial_values, timepoints, seir_model, parameter_list)
# plot intervention
plot (S~time, data=output, type="l", lwd=2, col=1, ylab="", ylim=c(0,1), main=paste("vaccine efficacy=",i, "compliance=",ii))
lines (E~time, data=output, type="l", lwd=2, col=2, ylab="")
lines (I~time, data=output, type="l", lwd=2, col=3, ylab="")
lines (R~time, data=output, type="l", lwd=2, col=4, ylab="")
legend ("right", legend=c("S","E","I","R"), col=c(1,2,3,4), lwd=2, cex=0.7)
# Cost
vaccine_cost = cost_per_vaccine * vac.total * X
new_Attack_rate = sum(output[,4])/(1/gamma_value)
new_treatment_cost = treatment_cost * new_Attack_rate * N
total_cost = vaccine_cost + new_treatment_cost
# Incremental cost effectiveness ratio
ICER = (total_cost - basecase_cost) / {(new_Attack_rate * N) - (Attack_rate * N)}
#Summary
print (paste("Basecase attack rate =", round(Attack_rate,3), "Basecase cost =", round(basecase_cost,3)))
print (paste("Vaccine compliance=",ii, "Vaccine efficacy=",i, "Attack_rate=",round(new_Attack_rate,3), "Total Healthcare Cost= $", round(total_cost,3),"Incremental cost effectiveness ratio = ", round(ICER,3)))
}
## [1] "Basecase attack rate = 0.95 Basecase cost = 4748966.747"
## [1] "Vaccine compliance= 0.5 Vaccine efficacy= 0.4 Attack_rate= 0.716 Total Healthcare Cost= $ 3618629.858 Incremental cost effectiveness ratio = 482.913"
## [1] "Basecase attack rate = 0.95 Basecase cost = 4748966.747"
## [1] "Vaccine compliance= 0.6 Vaccine efficacy= 0.4 Attack_rate= 0.666 Total Healthcare Cost= $ 3378568.928 Incremental cost effectiveness ratio = 483.081"
## [1] "Basecase attack rate = 0.95 Basecase cost = 4748966.747"
## [1] "Vaccine compliance= 0.6 Vaccine efficacy= 0.6 Attack_rate= 0.504 Total Healthcare Cost= $ 2590156.109 Incremental cost effectiveness ratio = 483.864"
COMMENT
We register an attack rate of 0.716 (71,6%) when the vaccine efficacy is 40% and compliance 50%, with total health care cost of $3,618,629.85 and ICER of $482.91 per case averted. We register an attack rate of 0.666 (66,6%) when the vaccine efficacy is 40% and compliance 60%, with total health care cost of $3,378,568.92 and ICER of $483.08 per case averted
We register an attack rate of 0.504 (50.4%) when the vaccine efficacy is 60% and compliance 60%, with total health care cost of $2,590,156.1 and ICER is $483.86 per case averted.
#Summary
print(ICER <- structure (list('vaccine efficacy 40%' = c(482.76, 482.91, 483.08),'vaccine efficacy 60%' = c(483.08, 483.38, 483.83)), .Names = c("vaccine efficacy 40%", "vaccine efficacy 60%"), row.names = c("compliance 40%", "compliance 50%", "compliance 60%"), class="data.frame"))
## vaccine efficacy 40% vaccine efficacy 60%
## compliance 40% 482.76 483.08
## compliance 50% 482.91 483.38
## compliance 60% 483.08 483.83
print(Attack_rate <- structure (list('vaccine efficacy 40%' = c(0.764, 0.712, 0.666),'vaccine efficacy 60%' = c(0.666, 0.588, 0.504)), .Names = c("vaccine efficacy 40%", "vaccine efficacy 60%"), row.names = c("compliance 40%", "compliance 50%", "compliance 60%"), class="data.frame"))
## vaccine efficacy 40% vaccine efficacy 60%
## compliance 40% 0.764 0.666
## compliance 50% 0.712 0.588
## compliance 60% 0.666 0.504