Optimal Treatment Strategy for Infectious Diseases with Two Treatment Stages ()
1. Introduction
Contagious disease epidemics, mostly caused by viruses, such as influenza, hepatitis, acquired immune deficiency syndrome (AIDS), coronavirus disease 2019, and so on, has always been a major threat to global public health. To understand the transmission dynamics of the infectious diseases, several mathematical models have been proposed and analyzed [1] [2] [3] [4] [5] . A central idea in mathematical epidemiology is that the threshold for many epidemiology models is the basic reproduction number
, i.e., the threshold
is the dividing line between the infection dying out and the onset of an epidemic. Here the basic reproduction number is defined as the average number of secondary infections produced when one infected individual is introduced into a host population where everyone is susceptible [6] .
After virus infection, the host usually displays several different stages according to the course of disease [7] [8] [9] , such as the asymptomatic and symptomatic stage, or the acute and chronic infection stage. At each stage, drug therapy is effective, although the effectiveness of treatment may vary at different stages, and recovered infectives usually are immune against reinfection. According to the above mentioned characteristics, we can divide the population into four compartments, named as susceptible (S), the asymptomatic/acute infected (A), the symptomatic/chronic infected (I), and recovered (R) individuals. Figure 1 illustrates the flowchart of the four compartments, which describes how individuals can move among the states.
Based on the flowchart in Figure 1, the dynamic model can be described by the following autonomous differential equations:
(1)
in which
means the recruitment rate of individuals,
denotes natural death rate,
is death rate due to infection. Corresponding to classes A and I respectively,
and
are the infection coefficient,
and
are the infectious period,
and
are the treated fractions, and
and
are the corresponding effectiveness. Obviously,
and
are connected with the medical resource. Hence, from the perspective of treatment strategy, we can also interpret
and
as medical resource for classes A and I, respectively. Note that natural birth and death rates are incorporated in our model (1) because many viral diseases, such as influenza, hepatitis and AIDS, cannot be eliminated and have persisted decades or more than one hundred years.
Due to the independence of variable R in system (1), the
equation can be eliminated from the system. Thus, in the following text, we only consider the reduced system:
Figure 1. Flow diagram for the SAIR model (1).
(2)
In this study, we will focus on how the treatment strategy
and
should be arranged to optimize the effect of limited medical resource. The rest of this paper is organized as follows. In Section 2, we present dynamic analysis, including the basic reproduction number
of system (2), the existence and globally stability of equilibria. Analytical and numerical results for optimal treatment strategy are given in Section 3. Finally, a short discussion completes the paper.
2. Global Dynamics Analysis
Since we are interested in the dynamics of infectious diseases, and not the initial processes of infection, we assume that the initial condition of (2) has the form
,
and
. Based on the initial conditions, it is easy to show that the solutions of system (2) are non-negative and ultimately bounded. The equilibria of (2) are the solutions of the following algebraic equations:
(3)
Here, for the sake of simplicity of notation, we always let
and
in the following text.
Clearly, the disease free equilibrium
of (2) always exists. Using the next generation matrix terminology in [6] , we can easily obtain the basic reproduction number of system (2) as
(4)
To obtain endemic equilibrium of model (2), solving the algebraic Equations (3), we can obtain that model (2) exists unique endemic equilibrium
if and only if
, in which
(5)
In order to obtain the stability of above mentioned equilibria, we first give the Jacobian matrix J of system (2) at
,
(6)
So we have the following result on the global stability of the disease free equilibrium E0.
Theorem 1. The disease free equilibrium E0 is globally asymptotically stable if
, and it is unstable when
.
Proof. According to (6) and the expression of E0, we know that
is the eigenvalue of
, and the other two eigenvalues of
are determined by the following sub-matrix
Simple calculations imply
and
if
. Thus, the real parts of the eigenvalues of J1 are negative, i.e., E0 is locally asymptotically stable if
. When
, we have
, so E0 is unstable in this case.
Let
Here k1 and k2 are non-negative constants which will be determined. By the first two equations of system (2), it follows
Taking the time derivative of V along the solution of system (2), and using the above-mentioned equations, we have
When
, we know
. Setting
after some algebraic calculations, we have
Thus,
if
, and
only if
and
. In this case, it is easy to obtain that the maximal invariant subset in
is the singleton
. As a result, E0 is globally asymptotically stable based on the LaSalle’s invariance principle [10] .
Note that system (1) has same structure as system (1) in Tien and Earn [11] . Tien and Earn [11] have proved that endemic equilibrium is globally asymptotically stable if it exists. Thus, we have the following result on the global stability of the endemic equilibrium E1.
Theorem 2. When
, the endemic equilibrium E1 is globally asymptotically stable.
3. Optimal Treatment Strategy
According to Theorem 1 and Theorem 2, we know that
can predict the disease. When
, the disease is an endemic disease, and the endemic equilibrium E1 is globally asymptotically stable. In this case, E1 gives the dynamical final size for different subpopulations. Therefore, the dynamical final infective size
can indicate the seriousness of the disease. So then, we will show how treatment strategies
and
should be arranged to press
and Z small if limited medical resources are given in this section. Here, the term “dynamical final infective size” is different from “final size” in [12] [13] , where the model did not include the natural birth and death. The term “dynamic” means that the rates of entering into and leaving from infective subpopulation are equal.
Firstly, we give the optimal treatment strategy to control the basic reproduction number
. In this case
will be considered as a function of
and
. Simple calculations imply
which means that when medical resource is abundant enough, whether increasing
or
can reduce
. However, when medical resources are limited, what is the optimal treatment strategy?
Let
. After some algebraic operations, we have
(7)
in which
Obviously, curve (7) divides the first quadrant of the
plane into following two regions:
(8)
Therefore, based on the size of
and
in the region, corresponding optimization strategies of reducing the basic reproduction number
as quickly as possible can be obtained. For more intuitive interpretation, a graphical representation of the regions is shown in Figure 2 under the following fixed artificial parameters:
(9)
Next, we give the optimal treatment strategy to control the dynamical final infective size Z. In this case,
is always valid and Z will be considered as a function of
and
. Simple calculations imply
(10)
Clearly,
if
. Otherwise, when
, we have
Furthermore, using the expression (4) of the basic reproduction number, after some algebraic operations, we have
Thus,
Figure 2. Illustration of optimal treatment strategy to control
as the function of treatment coefficients
and
. Here
and
in (A),
and
in (B), and the other parameters are taken in (9). The red line is the curve
in (7),
and
are defined in (8).
Because
and
, both
and
are valid. As a result, we know
is always valid.
According to (10), we know
is always valid. Thus, whether increasing
or
will always be beneficial for controlling the total scale of the disease if there are abundant medical resource.
After some algebraic operations, we know
is equivalent to the following equation:
(11)
Using the implicit function theorem, we know that (11) contains the function
, although the analytical expression of
cannot be obtained because (11) is too complex with respect to
and
.
For more intuitive interpretation, we give some numerical simulations under the fixed artificial parameters (9). Similar to (7), we find that the curve
may still divide the first quadrant of the
plane into two regions, with
in
and the opposite in
. As a result, the corresponding optimal treatment strategy of reducing the dynamical final infective size Z as quickly as possible can be obtained in different regions, which is shown in Figure 3.
Figure 3. Illustration of optimal treatment strategy to control Z as the function of treatment coefficients
and
. Here
in (A),
in (B), and the other parameters are taken in (9). The red line is the curve
implied by (11),
and
are defined based on the size of
and
.
4. Discussion
In this paper, a disease transmission model with two treatment stages is proposed and analyzed. When medical resources are limited, the following optimization strategies from the perspectives of basic reproduction number and dynamical final infective size can be drawn:
● When the parameters lie in region
, because increasing
can decrease
(or Z) more greatly than increasing
, the optimal treatment strategy is priority treatment I class until
(or
).
● On the contrary, when the parameters lie in region
, the optimal treatment strategy is priority treatment A until
(or
).
● The curve
(or
) is the optimal treatment strategy under limited medical resources.
In practical application, if the goal is to eliminate diseases, it is better to optimize the control of the basic reproduction number and ultimately make it less than 1. When the disease has become an endemic disease and it is difficult to achieve the target of the basic reproduction number less than 1, then optimal control of the dynamic final effective size should be carried out to minimize the scale of the endemic disease as much as possible. Furthermore, in order to better prevent and control the epidemic, further studies are warranted to optimizing the control of the basic reproduction number and dynamic final effective size simultaneously.
Acknowledgements
This work was supported by the National Natural Science Foundation of China (grant number: 11901576).