Basic Epidemic Model

Introduction to Infectious Disease Modelling:
Mathematical Epidemiology Approach
Althaf Setyawan
Center for Reproductive Health,

and
Department of Biostatistics, Epidemiology and
Population Health
Althaf Setyawan SEMINAR - UNDIP November 2021 1 / 44

Learning Objectives
Goals
Upon completion of the course unit, participant should be able to:
a. Understand the topic of infectious disease modeling on population level.
b. Develop skills to form and analyze simple mathematical models of
infectious diseases.
c. Further, develop skills to compute the basic reproduction number.

Contents
Table of Contents
1 Background and Introduction
2 Basic Epidemic models.
3 Techniques for analysis of Ordinary Differential Equation systems
4 Building Basic ODE Epidemic disease.
5 Fitting ODE epidemic models to data

Introduction
Classification of Infectious Diseases

 An infectious disease is a clinically evident illness
resulting from the presence of a pathogenic microbial
agent. The microbial agent causing the disease can be
bacterial, viral, fungal, parasitic, or it can be toxic proteins,
called prions.
 Communicable diseases are infectious diseases that can
be transmitted from one infectious person to another,
directly or indirectly. Often, we do not make a distinction
between infectious diseases and communicable
diseases, since many of the infectious diseases are in
fact communicable diseases.
 Transmittable diseases are infectious diseases that can
be transmitted from one person to another through
unnatural routes.

Introduction
Transmission of Infectious Diseases

 Person-to-person transmitted diseases are diseases that
require direct or indirect contact.
 Airborne transmission occurs on inhalation of infected air.
 Food- and waterborne diseases are transmitted through
ingestion of contaminated food or water.
 Vector-borne diseases are transmitted by a vector, most
often an arthropod such as a mosquito or tick, or a
mollusk such as a snail.
 Vertical transmission occurs when a disease is
transmitted through the placenta from a mother to a child
before or at birth.

Introduction
Transmission of Infectious Diseases

For modeling purposes, we distinguish four types of
transmission:
 direct, when the causative pathogen is transmitted from
one person to another;
 vector-transmitted, when the causative agent is
transmitted from a vector to a human;
 environmental transmission, when a human becomes
infected through contact with a pathogen present in the
environment; and
 vertical, when the pathogen is transmitted from mother to
child at birth.

Introduction
Basic Definitions in the Epidemiology of Infectious Diseases
These concepts play an important role in the construction of

mathematical models by adding various features to the
model direct, when the causative pathogen is transmitted
from one person to another;
 Exposed Individuals. When a healthy individual who is
vulnerable to contracting a disease makes a potentially
disease-transmitting contact, that individual becomes
exposed. Exposed individuals may or may not develop
the disease. These individuals are typically not infectious.
In mathematical models, we often assume that all
exposed individuals eventually develop the disease.

Introduction

 Infected and Infectious Individuals. If the pathogen
establishes itself in an exposed individual, then that
individual becomes infected. Infected individuals who can
transmit the disease are called infectious. Infected
individuals may not be infectious during the entire time of
being infected.
 Latent Individuals. These are individuals that are
infected but not yet infectious. The latent period is
defined as the time from infection to when the host is able
to transmit the infectious agent to another individual.
 Incubation Period. The incubation period is the period
between exposure to an infectious agent and the onset of
symptoms of the disease. In infectious diseases, the
incubation period is the time required for the infectious
agent to multiply to a threshold necessary to produce
symptoms or laboratory evidence of infection.
Introduction
 Incidence. Incidence is defined as the number of

individuals who become ill during a specified interval of
time (e.g., one year).
 Prevalence. The prevalence of a disease is the number
of people who have the disease at a specific time.
 Case Fatality Proportion (CFP). The case fatality
proportion is given as the ratio of people who die of a
disease to those who contract it.
 Disease-Induced Mortality. Disease-induced mortality is
the number of people who have died from the disease in
one unit of time (e.g., one year) divided by the entire
population.

Introduction
Historical Remarks on Infectious Diseases and Their Modeling
 In the late 1800s, Robert Koch, identified the specific

causative agents of tuberculosis, cholera, and anthrax,
thus giving experimental support to the concept of
infectious disease. He was also famous for the
development of Koch’s postulates. Science could finally
explain the mechanism of how one becomes ill. The
concept of passing a bacterial disease through contact
between an infected individual and a healthy one became
known. This paved the way for the mathematical
modeling of infectious diseases.
 Mathematical epidemiology was raised to a new level by
the model of the spread of infectious diseases, published
by Kermack and McKendrick in 1927. In their joint article
“A contribution to the mathematical theory of epidemics”

Introduction
General Approach to Modeling
 A mathematical model is a description of a system using mathematical

tools and language. The process of developing mathematical models
is called mathematical modeling.
 In this course we will be concerned with modeling infectious diseases
and their spread in populations, but in principle, mathematical
modeling can be applied to any system, biological or otherwise.”
Introduction
Classification of Models
Models can be classified in multiple ways:
 Linear/nonlinear. A model is classified as nonlinear if it
contains a nonlinear dependence on the variables (e.g., a
product of variables). Otherwise, it is classified as linear.
 Static/dynamic. A dynamic model accounts for time-
dependent changes in the state of the system, while a static
model calculates system quantities assuming that it does not
change in time and thus is time-invariant. Dynamic models
typically employ differential equations or difference equations.
 Discrete/continuous. Discrete models treat time or system
states as discrete. Continuous models incorporate time and
system states as continuous.
 Deterministic/stochastic. A deterministic model is one in
which every set of variable states is uniquely determined by the
parameters in the model and the initial state of the variables.
Stochastic models are characterized by randomness, and
variable states are described by probability distributions.

Introduction
Frequency and Cumulative Graphs

 During an outbreak of a disease, such as the coronavirus (COVID-19)
pandemic, the media shows daily graphs that convey the spread of
the disease.
 New cases for each day (or week). This information is usually shown
as a histogram or needle plot.
 The graph is sometimes called a frequency graph.
 The total number of cases plotted against time.
 Usually, this graph is a line graph.
 The graph is sometimes called a cumulative frequency graph.

Introduction
Reading a cumulative frequency graph

 A typical cumulative curve is somewhat S-shaped, as shown to
the right.
 The initial portion of the curve (the red region) is concave up,
which indicates that the number of new cases is increasing.

Introduction
Frequency Graphs
 some scenarios
will overwhelm
the health care
system
whereas others
will not

Epidemic Modeling
Introduction to Epidemic Modeling

Introduction to epidemic modeling is usually made through one of the first
epidemic models proposed by Kermack and McKendrick in 1927, a model
known as the SIR epidemic model.
 The class of individuals who are healthy but can contract the disease.
These are called susceptible individuals or susceptible. The size of this
class is usually denoted by S.
 The class of individuals who have contracted the disease and are now
sick with it, called infected individuals. In this model, it is assumed that
infected individuals are also infectious (there are distinction between
infected and infectious individuals). The size of the class of
infectious/infected individuals is denoted by I.
 The class of individuals who have recovered and cannot contract the
disease again are called removed/recovered individuals. The class of
recovered individuals is usually denoted by R.

Epidemic Modeling
Basic SIR Models
The number of individuals in each of these classes changes with

time, that is, S(t), I(t), and R(t) are functions of time t. The total
population size N is the sum of the sizes of these three classes
𝑁𝑁 𝑡𝑡 = 𝑆𝑆 𝑡𝑡 + 𝐼𝐼 𝑡𝑡 + 𝑅𝑅(𝑡𝑡)
 To formulate a model, we have to make assumptions to simplify
reality.
 The first assumption for the Kermack–McKendrick
model is that infected individuals are also infectious.
 The second assumption of the model is that the total
population size remains constant.

Epidemic Modeling
Simple Definitions – stock of people

 S = Number of susceptible persons. This is the number of people that
can potentially become infected. With a new virus and at time 0 of an
epidemic, this is usually the entire population
 I = Number of infected persons. Number of people who contracted the
infection and are infectious. They can potentially transmit the virus to
others
 R = Number of recovered people. Those who have been infected (I)
and have recovered
 N = Population = S+I+R

Epidemic Modeling
Simple Definitions – shares of population
 N = Population = S+I+R
 s = S/N = Share of the population that is susceptible to the virus
 i = I/N = share of the population that is infected
 r = R/N = share of the population that has recovered
 s+i+r=1

Epidemic Modeling
Basic SIR Models

 Epidemiological models consist of systems of ODEs that describe
the dynamics in each class.
 To derive the differential equations, we consider how the classes
change over time. When a susceptible individual enters into contact
with an infectious individual, that susceptible individual becomes
infected with a certain probability and moves from the susceptible
class into the infected class.
 The susceptible population decreases in a unit of time by all
individuals who become infected in that time. At the same time, the
class of infectives increases by the same number of newly infected
individuals.
 The number of individuals who become infected per unit of time in
epidemiology is called incidence, and the rate of change of the
susceptible class is given by
𝑑𝑑𝑑𝑑
= 𝑆𝑆 ′ (𝑡𝑡) = −incidence
𝑑𝑑𝑑𝑑
Epidemic Modeling
Basic SIR Models
𝛽𝛽 𝛾𝛾
The rate of change of the susceptible class is given by
𝑆𝑆 ′ 𝑡𝑡 = −incidence = −𝛽𝛽𝑆𝑆𝑆𝑆
 The coefficient β is the constant of proportionality called the
transmission rate constant = number of contacts per day per
infected individual = =1/average period an infectious person
makes an infecting contact (measured in unit time = days)
 βS is number of susceptible individuals who become infected per
unit of time per infectious individual.
 βSI is the number of individuals who become infected per unit of
time (incidence).
Epidemic Modeling
Basic SIR Models
 The susceptible individuals who become infected move to the

class I. Those individuals who recover or die leave the infected
class at constant per capita probability per unit of time γ, called
the recovery rate.
 γ = R/I fraction of the infectious group that will recover =
1/average period of infectiousness (measured in days)
 That is, γI is the number of infected individuals per unit of time
who recover. So,
𝐼𝐼 ′ 𝑡𝑡 = 𝛽𝛽𝑆𝑆𝑆𝑆 − 𝛾𝛾𝛾𝛾
 Individuals who recover leave the infectious class and move to
the recovered class
𝑅𝑅′ 𝑡𝑡 = 𝛾𝛾𝛾𝛾
Epidemic Modeling
Basic SIR Models

The whole model is given by the following system of ODEs:
𝑆𝑆 ′ 𝑡𝑡 = −𝛽𝛽𝑆𝑆𝑆𝑆
 To be well defined mathematically, this system is equipped with
given initial conditions S(0), I(0), and R(0).
 We denote by N the total population size at time zero N =
S(0)+I(0)+R(0).
 Adding all three equations in system, we obtain
N’(t)=S’(T)+I’(t)+R’(t)=0. Hence, N(t) is constant and equal to its
initial value, N(t) = N. This model is called the SIR model or SIR
system.

Epidemic Modeling
Basic SIR Models

SIR epidemic model:
𝐼𝐼 ′ 𝑡𝑡 = 𝛽𝛽𝑆𝑆𝑆𝑆 − 𝛾𝛾𝛾𝛾 𝛾𝛾
The Kermack–McKendrickmodel is based on several assumptions:
 (1) There are no births and deaths in the population.
 (2) The population is closed, that is, no one from the outside enters the
population, and no one leaves the population, and finally,
 (3) All recovered individuals have complete immunity and cannot be
infected again.
Diseases that lead to permanent immunity and are well modeled by the
SIR epidemic model are most diseases typical of childhood years, often
called childhood diseases. These include chickenpox, smallpox, rubella,
and mumps.

Epidemic Modeling
The SIR curves plotted using population share

SIR epidemic model:
𝑆𝑆 ′ 𝑡𝑡 = −𝛽𝛽𝑆𝑆𝑆𝑆/𝑁𝑁 𝛾𝛾
𝐼𝐼 ′ 𝑡𝑡 = (𝛽𝛽𝑆𝑆𝑆𝑆 − 𝛾𝛾𝛾𝛾)/𝑁𝑁
𝑅𝑅′ 𝑡𝑡 = 𝛾𝛾𝛾𝛾/𝑁𝑁
 Since the population is equal to 1,when the infection rate increases, the
recovery rate increases and the susceptible rate decreases.
 The rate of increase and decrease of these parameters depend on the
parameters β and γ

Epidemic Modeling
What happens when β varies?

 If the average period an infectious person makes an infecting contact is
increased, β decreases, infections per day decrease and the infection
rate curve is “flattened”.
 This is why social distancing works. It reduces the frequencies of the
contacts between infectious and susceptible people (β ).
 β=1/1; means
that an infected
person makes
an infecting
contact every
days.
 β=1/2; means
that an infected
person makes
an infecting
contact every
two days.

Epidemic Modeling
What happens when γ varies?

 If the average period of infectiousness increases, γ decreases,
infections per day increase and the infection rate
 This is why treatment may help. It can reduce the average period of
infectiousness.
 γ=1/3; means
that 1/3 of the
infectious group
will recover each
day.
 γ=1/4; means
that 1/4 of the
infectious group
will recover each
day.

Epidemic Modeling
Estimating the Recovery Rate (γ)

SIR epidemic model:
𝑅𝑅 ′ 𝑡𝑡 = 𝛾𝛾𝛾𝛾
average periodof infectiousness = 𝛾𝛾1
For influenza, we are sick with it for 3–7 days. Say that the mean time
spent as infectious is 5 days. Thus, the recovery rate, measured in units
of [days]−1, is 1/5.

Epidemic Modeling
Estimating the Transmission Rate (β)

SIR epidemic model:
Estimating β is possible for the Kermack–McKendrick model, because

that model is relatively simple.
𝑆𝑆0
ln( )
𝛽𝛽 = 𝑆𝑆 +𝐼𝐼𝑆𝑆~
−𝑆𝑆~
𝛾𝛾
𝛾𝛾 0 0
The Kermack–McKendrick model is equipped with initial conditions: S0 =

S(0) and I0 = I(0). S~=the final number of susceptible individuals after the
epidemic is over.

Epidemic Modeling
Estimating Parameters from Data

Example: The SIR Model and Influenza at an English Boarding School 1978
• In January and February 1978, an epidemic of influenza occurred in a
boarding school in the north of England. The boarding schoolhouse a total of
763 boys, all of whom were at risk during the epidemic.
• The spring term began on January 10. The boys returned from their New
year vacation spent at many different locations in the world. A boy returning
from Hong Kong exhibited elevated temperature during the period 15–18
January. On January 22, three boys were sick.
• Table below gives the number of boys ill on the nth day beginning January
22 (n = 1).

Epidemic Modeling

• The number of boys who escaped influenza was 19. The average time spent
sick was 5–6 days. However, since boys were isolated in the infirmary, they
spent perhaps about 2 days as infectious. A swab taken from some of the
boys revealed that they were infected with H1N1 influenza A virus. The staff
of the boarding school remained healthy, with only one staff member
displaying symptoms of illness.
• These data give the following values: S3 = 738, I3 = 25, S∞ = 19.
• From the computations:
𝛾𝛾
• We take the infective period to be 2.1 days. This value can be obtained as
the best fit as values around 2 days are tried with the procedure below. After
we fix the duration of the infectious period, we compute α as the reciprocal
of the time spent as an infectious individual (infectious period):
𝛾𝛾
Epidemic Modeling

• Using the value for γ, we can obtain the value for β
• β = 0.004918 γ = 0.004918 ∗ 0.476= 0.002342

Epidemic Modeling
Fitting parameter continuous-time models to data using trajectory matching

with least squares and Maximum Likelihood
• If we assume that the only source of variability in the data is measurement
error, and that this is symmetrically distributed with a constant variance. If
these conditions apply, then least squares is a statistically appropriate basis
for estimation.
• This process is referred to as optimization and, fortunately for us, there are
many robust algorithms available for this purpose. One of them, the Nelder-
Mead algorithm, using computer software. We can use it to return the best
values for β and γ
• You can also use maximum likelihood estimator (MLE) method to get the
best value β and γ, not just the Gaussian case, where MLE and least
squares are equivalent.

Epidemic Modeling
The basic reproduction number R0

• The basic reproduction number is usually denoted by R0 or contact number
for the disease is the number of secondary infections caused by introducing
a single infective into a wholly susceptible population.
𝛽𝛽𝛽𝛽
𝑅𝑅0 =
𝛾𝛾
• The quantity R0 determines whether there is an epidemic.
• If R0<1 , the infection dies out because I′(t) < 0 for all t, and there is no
epidemic.
• If R0>1, I increases initially, and this is interpreted as saying that there is an
epidemic.

Epidemic Modeling
The basic reproduction number R0

A dynamic social context with its interpretation over interventions; initially, an infected individual
introduces an infectious agent into a heterogeneous population with individuals with different
profiles of susceptibility to the infectious agent, who are likely to produce secondary cases from
the primary one, in this context, if the intervention is effective, it will be able to reduce the cases
secondary infections by reducing the basic reproductive number of the disease in the population.

Epidemic Modeling
A Simple SIS Epidemic Model

• We want to relax the assumption for permanent immunity after recovery to
model diseases that can infect us repeatedly, such as influenza. We may
assume in the simplest scenario that individuals who recover become
immediately susceptible again. 𝛾𝛾
• The model takes the form
𝑆𝑆 ′ 𝑡𝑡 = −𝛽𝛽𝑆𝑆𝑆𝑆+ 𝛾𝛾𝛾𝛾
• System (is called an SIS epidemic model and is perhaps the simplest model
in mathematical epidemiology. Here, if N = S+I and we add the two
equations

Epidemic Modeling

𝛾𝛾
The simplest SIS model, due to Kermack and McKendrick is based on the
following assumptions:
• The rate of new infections is given by mass action incidence.
• Infectives leave the infective class at rate γI per unit time and return to the
susceptible class.
• There is no entry into or departure from the population.
• There are no disease deaths, and the total population size is a constant N.

Epidemic Modeling

• We want to relax the assumption for permanent immunity after recovery to
model diseases that can infect us repeatedly, such as influenza. We may
assume in the simplest scenario that individuals who recover become
immediately susceptible again. 𝛾𝛾
• The model takes the form
• System (is called an SIS epidemic model and is perhaps the simplest model
in mathematical epidemiology. Here, if N = S+I and we add the two
equations

Epidemic Modeling
A SEIR Epidemic Model

• The SEIR model is another classical epidemiological model, which
incorporates a compartment of exposed individuals, E(t), where the
individuals are infected but not infectious. With S(t) denoting the number of
susceptible individuals, I(t) the number of infectious individuals, and R(t) the
number of recovered individuals, the model takes the form
𝐸𝐸 ′ 𝑡𝑡 = 𝛽𝛽𝑆𝑆𝑆𝑆 − 𝜎𝜎𝜎𝜎
𝐼𝐼 ′ 𝑡𝑡 = 𝜎𝜎𝜎𝜎 − 𝛾𝛾𝛾𝛾
• Here, if N = S+E+I+R and we add the four equations
𝛽𝛽 𝜎𝜎 𝛾𝛾
S E I R

Epidemic Modeling
S E I R
𝐸𝐸 ′ 𝑡𝑡 = 𝛽𝛽𝑆𝑆𝑆𝑆 − 𝜎𝜎𝜎𝜎
𝐼𝐼 ′ 𝑡𝑡 = 𝜎𝜎𝜎𝜎 − 𝛾𝛾𝛾𝛾
The SEIR model assumptions
• All persons of a population can be assigned to one of these three categories
at any point of the epidemic
• Once recovered, a person cannot become infected again (this person
becomes immune)
• Natural births and natural deaths of the population are assumed to balance
each other out
• International migration is ignored, and the population is considered constant.

Epidemic Modeling
S E I R
Interpretation of parameter (time period= day)
• β = number of contacts per day per infected individual =1/average period an
infectious person makes an infecting contact (measured in days)
• γ = R/I fraction of the infectious group that will recover = 1/average period of
infectiousness (measured in days)
• σ = Rate of infected individuals becoming infectious
• 1/σ = Average duration of incubation
Example:
• β = 1/2 means that an infected person makes an infecting contact every two
day
• γ =1/3 means that 1/3 of the infectious group will recover each time period
(day)
• σ = 0.25 means that the average duration of incubation is 4 days and that
25% of exposed/infected persons become infectious every days.

Epidemic Modeling
What happens when σ varies?

 If the average period of infectiousness increases, γ decreases,
infections per day increase and the infection rate
 This is why social distance may help. It can ensure that people
incubating the virus are not circulating and buys time for the health
authority to prepare.
 σ = 0.5 means that the
average duration of
incubation is 2 days and
that 50% of
exposed/infected
persons become
infectious every time
period (days)
 σ = 0.3 means that the
average duration of
incubation is 3 days and
that 30% of
exposed/infected
persons become
infectious every time
period (days)
Epidemic Modeling
A modified SEIR Epidemic Model for Analysis of Covid-19 transmission

For implementing to the model for COVID-19.
The population is divided into
• Susceptible (S),
• Exposed (E),
• Infected Undetected (IU),
• Infected Detected (directly quarantine) (IQ),
and
• Recovered (R).

Epidemic Modeling
A single-patch SEIQR model for Analysis of Covid-19 transmission
The population was divided into five groups

according to disease state
• Susceptible (S),
• Exposed (E),
• Confirmed and therefore, quarantined
individuals (Q)
• Recovered (R).

Basic Epidemic Model

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Basic Epidemic Model

Uploaded by

Copyright:

Available Formats

Introduction to Infectious Disease Modelling:

Mathematical Epidemiology Approach

Center for Reproductive Health,

Althaf Setyawan SEMINAR - UNDIP November 2021 1 / 44

Althaf Setyawan SEMINAR - UNDIP November 2021 2 / 44

1 Background and Introduction

2 Basic Epidemic models.

3 Techniques for analysis of Ordinary Differential Equation systems

4 Building Basic ODE Epidemic disease.

5 Fitting ODE epidemic models to data

Althaf Setyawan SEMINAR - UNDIP November 2021 3 / 44

Classification of Infectious Diseases

Althaf Setyawan SEMINAR - UNDIP November 2021 4 / 44

Transmission of Infectious Diseases

Althaf Setyawan SEMINAR - UNDIP November 2021 5 / 44

Transmission of Infectious Diseases

Althaf Setyawan SEMINAR - UNDIP November 2021 6 / 44

Basic Definitions in the Epidemiology of Infectious Diseases

These concepts play an important role in the construction of

Althaf Setyawan SEMINAR - UNDIP November 2021 7 / 44

Basic Definitions in the Epidemiology of Infectious Diseases

Basic Definitions in the Epidemiology of Infectious Diseases

 Incidence. Incidence is defined as the number of

Althaf Setyawan SEMINAR - UNDIP November 2021 9 / 44

Historical Remarks on Infectious Diseases and Their Modeling

 In the late 1800s, Robert Koch, identified the specific

Althaf Setyawan SEMINAR - UNDIP November 2021 10 / 44

General Approach to Modeling

 A mathematical model is a description of a system using mathematical

Althaf Setyawan SEMINAR - UNDIP November 2021 12 / 44

Frequency and Cumulative Graphs

Althaf Setyawan SEMINAR - UNDIP November 2021 13 / 44

Reading a cumulative frequency graph

Althaf Setyawan SEMINAR - UNDIP November 2021 14 / 44

Althaf Setyawan SEMINAR - UNDIP November 2021 15 / 44

Introduction to Epidemic Modeling

Althaf Setyawan SEMINAR - UNDIP November 2021 16 / 44

Basic SIR Models

The number of individuals in each of these classes changes with

Althaf Setyawan SEMINAR - UNDIP November 2021 17 / 44

Simple Definitions – stock of people

Althaf Setyawan SEMINAR - UNDIP November 2021 18 / 44

Simple Definitions – shares of population

Althaf Setyawan SEMINAR - UNDIP November 2021 19 / 44

Basic SIR Models

Basic SIR Models

The rate of change of the susceptible class is given by

Basic SIR Models

 The susceptible individuals who become infected move to the

Basic SIR Models

Althaf Setyawan SEMINAR - UNDIP November 2021 23 / 44

Basic SIR Models

Althaf Setyawan SEMINAR - UNDIP November 2021 24 / 44

The SIR curves plotted using population share

Althaf Setyawan SEMINAR - UNDIP November 2021 25 / 44

What happens when β varies?

Althaf Setyawan SEMINAR - UNDIP November 2021 26 / 44

What happens when γ varies?

Althaf Setyawan SEMINAR - UNDIP November 2021 27 / 44

Estimating the Recovery Rate (γ)

Althaf Setyawan SEMINAR - UNDIP November 2021 28 / 44

Estimating the Transmission Rate (β)