 Research
 Open Access
 Published:
Global threshold dynamics of a stochastic epidemic model incorporating media coverage
Advances in Difference Equations volume 2018, Article number: 462 (2018)
Abstract
In this paper, we investigate the global threshold dynamics of a stochastic SIS epidemic model incorporating media coverage. We give the basic reproduction number \(\mathcal{R}_{0}^{s}\) and establish a global threshold theorem by Feller’s test: if \(\mathcal{R}_{0}^{s}\leq 1\), the disease will die out a.s.; if \(\mathcal{R}_{0}^{s}>1\), the disease will persist a.s. In the case of \(\mathcal{R}_{0}^{s}>1\), we prove the existence, uniqueness, and global asymptotic stability of the invariant density of the Fokker–Planck equations associated with the stochastic model. Via numerical simulations, we find that the average extinction time decreases with the increase of noise intensity σ, and also find that the increasing σ will be beneficial to control the disease spread. Thus, in order to control the spread of the disease, we must increase the intensity of noise σ.
Introduction
It is now widely believed that environmental variations have a critical influence on the spread of the disease [1,2,3,4], and stochastic noise plays an indispensable role in the transmission of diseases, especially in a small population. Therefore, it seems more practical to consider stochastic epidemic models [5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20].
In order to understand the role of media coverage towards the disease transmission dynamics in a random environment, based on the results in [21], Cai et al. [22] studied the following stochastic differential equations (SDE) SIS model with a standard incidence rate:
where \(S(t)\) and \(I(t)\) are the number of susceptible and infective individuals at time t, respectively. \(\beta_{1} \frac{\beta_{2}I}{b+I}\) is the effective contact rate, \(\beta_{1}\) is the usual contact rate without considering the infective individuals, \(\beta_{2}\) is the maximum reduced contact rate due to the presence of the infected individuals. \(\sigma_{i}\) (\(i=1,2\)) is a real constant and \(\sigma_{i}^{2}\) is known as the intensity of environmental fluctuations, \(B_{i}(t)\) (\(i=1,2\)) is the standard onedimensional independent Wiener process defined over the complete probability space \(( \varOmega, \mathcal{F}, \{\mathcal{F}_{t}\}_{t\geq 0}, \mathrm{Prob} ) \).
In [22], the authors defined \(R_{0}^{s}:=\frac{ \beta_{1}}{\mu +\gamma }\frac{\sigma_{2}^{2}}{2(\mu +\gamma)}\) and obtained the following results by using the method of stochastic stability [23]:

(A1)
If one of the following conditions
$$ R_{0}^{s}< 1 \quad \text{and} \quad \sigma_{2}^{2}< 2 \beta_{1}, \quad \text{or}\quad \sigma_{2}^{2}\geq 2 \beta_{1}, $$holds, then the disease will die out with probability one (Theorem 3.1, [22]).

(A2)
If \(R_{0}^{s}>1\), then the disease will persist with probability one (Theorem 4.1, [22]).
Furthermore, based on the results in [22], by using the theory of Markov semigroup and asymptotic properties [9, 24], Guo et al. [25] studied the following SDE SIS model with bistable incidence rate:
and obtained that:

(B1)
If
$$ {R}_{0}^{s}:=\frac{\varLambda \beta_{1}}{\mu (\mu +\gamma)}\frac{ \sigma^{2}}{2(\mu +\gamma)}< 1, $$then the disease will die out with probability one (Theorem 3.4, [25]).

(B2)
If
$$ {R}_{0}^{s}>1 \quad \text{and} \quad \sigma^{2}< 2 \mu \min \{1,A\} $$hold, then the stochastic process \((S(t),I(t))\) has a unique stationary distribution (Theorem 3.7, [25]).
It is well known that epidemic threshold theorem holds for most deterministic compartmental epidemic models by the basic reproduction number \(R_{0}\) [26]: if \(R_{0}<1\), there is a diseasefree equilibrium which is globally asymptotically stable; if \(R_{0}>1\), there exists an endemic equilibrium which is globally asymptotically stable. However, in [A1] for SDE model (1), there is an extra condition \(\sigma_{2}^{2}<2\beta_{1}\), and in [B2] for SDE model (2), there is an extra condition \(\sigma^{2}<2\mu \min \{1,A\}\).
There naturally comes a question: Is there any global threshold theorem for a stochastic epidemic model (e.g., SDE model (1) or (2)) incorporating media coverage?
Thanks to the insightful work of Xu [10], in this paper, we will focus on the global threshold dynamics for the following SDE epidemic model incorporating media coverage:
Since \(S(t)+I(t)=N\), which is a constant, SDE model (3) can degenerate into the following onedimensional model:
where
Particularly, if \(\sigma =0\) in (4), i.e., without environmental noise, we can obtain
For simplicity, we call model (6) a deterministic model corresponding to SDE model (4).
One goal of this paper is to establish a global threshold theorem for SDE model (4). We will prove that the basic reproduction number \(\mathcal{R}_{0}^{s}\) can be used to govern the stochastic dynamics of SDE model (4).
The other goal is to further study the invariant density of process \(I(t)\). Many longterm asymptotic properties of dynamical systems or random dynamical systems can be described in terms of invariant measure [27] and the density function with respect to Lebesgue measure of the marginals of an invariant measure that can be called an invariant density [28]. If invariant density is \(\mathcal{L}^{1}\) on a set Ω, it satisfies the Fokker–Planck equations (FPE) in the interior of Ω [29]. Hence, we will investigate the FPE associated with (4) and solve the invariant density.
This paper is organized as follows. In Sect. 2, we present the global stochastic threshold theorem. In Sect. 3, in the case of disease persistence, we derive the existence, uniqueness, global stability, and an explicit formula of an invariant density of the Fokker–Planck equation associated with (4). In Sect. 4, we give some numerical examples to show the complicated stochastic dynamics of the model. And in the last section, Sect. 5, we provide a brief discussion and the summary of our main results.
Stochastic threshold theorem
In this section, we will focus on the stochastic threshold theorem for model (4). First of all, we state the global existence of the uniqueness and boundedness of the positive solution of model (4).
Theorem 2.1
For any given initial value \(I_{0}\in (0, N)\), SDE (4) has a unique global positive solution \(I(t) \in (0, N)\) for all \(t\geq 0\) with probability one, namely
The proof of Theorem 2.1 is similar to that in [22] or [25]. So we omit it here.
Next, similar to [22, 25] and [8], we define the basic reproduction number \(\mathcal{R}_{0}^{s}\) for SDE model (4) as follows:
where \(R_{0}:=\frac{\beta_{1}}{\mu +\gamma }\) is the basic reproduction number of the deterministic model (6).
Theorem 2.2

(i)
If \(\mathcal{R}_{0}^{s}\leq 1\), for any given initial value \(I_{0} \in (0, N)\),
$$ \mathbb{P} \Bigl\{ \lim_{t\rightarrow \infty }I(t)=0 \Bigr\} =1. $$Namely, the disease will go extinct with probability one.

(ii)
If \(\mathcal{R}_{0}^{s}>1\), for any given initial value \(I_{0} \in (0, N)\),
$$ \mathbb{P} \Bigl\{ \sup_{0\leq t< \infty }I(t)=N \Bigr\} =\mathbb{P} \Bigl\{ \inf_{0\leq t< \infty }I(t)=0 \Bigr\} =1. $$
In particular, the process \(I(t)\) is recurrent: for every \(\theta \in (0, N)\), we have
Namely, the disease will persist with probability one.
Proof
Define
then
hence
Using Itô’s formula for \(Y(t)\), we have
Substituting (8) into (10), we obtain
Following [30, 31], the scale function defined for (11) is given by
with
It is easy to see that \(\psi (\infty)=\infty \). While if \(\mathcal{R}_{0}^{s}> 1\), then \(\psi (\infty)= \int_{0}^{\infty }\phi (\xi)\,d\xi =\infty\).
On the other hand, if \(\xi >0\) and \(\mathcal{R}_{0}^{s}\leq 1\), then
Hence, if \(\mathcal{R}_{0}^{s}\leq 1\), \(\psi (\infty)= \int_{0}^{\infty }\phi (\xi)\,d\xi \geq 4 \int_{0}^{\infty }\,d\xi =\infty\).
It follows from [31, Propositions 5.22] and [10, Lemma A.2] that we end the proof. □
The properties of invariant density
In this section, in the case of \(\mathcal{R}_{0}^{s}>1\), we will focus on the properties of the invariant density of the process \(I(t)\) of SDE model (4). Instead of studying SDE model (4) directly, we investigate the associated Fokker–Planck equation (FPE) with (4), which is given by
From (12), we denote by \(\{\mathbb{P}(t)\}_{t\geq 0}\) the Markov semigroup. And the following theorem gives the existence, uniqueness, and asymptotic stability of an invariant density of the Markov semigroup \(\{\mathbb{P}(t)\}\).
Theorem 3.1
If \(\mathcal{R}_{0}^{s}>1\), then there exists a unique invariant probability measure \(\nu_{\sigma }^{s}\) for (12) which has the density \(p_{\sigma }^{s}\) with respect to the Lebesgue measure. Moreover,

(i)
the process \(I(t)\) has the ergodic properties, i.e., for any \(\nu_{\sigma }^{s}\)integrable function G,
$$ \mathbb{P}_{I_{0}} \biggl( \lim_{t\rightarrow \infty } \frac{1}{t} \int_{0}^{t}G(I_{\tau })\,d\tau = \int_{0}^{K}G(y)\nu_{\sigma }^{s}(dy) \biggr) =1 $$for all \(I_{0}\in (0,N)\);

(ii)
the invariant density \(p_{\sigma }^{s}\) is globally asymptotically stable in the sense that
$$ \lim_{t\rightarrow \infty } \int_{0}^{K} \bigl\vert \mathbb{P}(t)g(x)p_{\sigma }^{s}(x) \bigr\vert \,dx=0,\quad \forall g\in { L_{+}^{1}(0,N)}, $$and
$$\begin{aligned} L_{+}^{1}\bigl((0,N)\bigr) :=& \biggl\{ w\in L^{1}(\mathcal{R}): \int_{0}^{N}w(x)\,dx=1, w(x)=0 \textit{ for } x\geq N \textit{ or } x\leq 0, \\ &{} \textit{and } w(x)\geq 0 \textit{ for } x\in \mathcal{R} \biggr\} . \end{aligned}$$ 
(iii)
the unique invariant density \(p_{\sigma }^{s}\) is given by
$$ p_{\sigma }^{s}(x):=CN^{3\frac{2\beta_{2}N}{\sigma^{2}(b+N)}} \frac{x ^{c_{0}(\mathcal{R}_{0}^{s}1)1}(x+b)^{\frac{2\beta_{2}N}{\sigma ^{2}(b+N)}}}{(Nx)^{c_{0}(\mathcal{R}_{0}^{s}1)+3\frac{2\beta_{2}N}{ \sigma^{2}(b+N)}}}e^{\frac{c_{0} x}{Nx}} $$(13)with
$$ C^{1}= \int_{0}^{\infty }\bigl(b+(b+N)e^{\xi } \bigr)^{ \frac{2N\beta_{2}}{\sigma^{2}(b+N)}}\bigl(e^{\xi }+1\bigr)^{2} e^{c_{0}( \mathcal{R}_{0}^{s}1)\xi c_{0}e^{\xi }}\,d\xi, $$(14)and \(c_{0}=\frac{2(\mu +\gamma)}{\sigma^{2}}\).
Proof
First of all, we need to study the following FPE associated with SDE (11):
We next prove assertions (i) and (ii). Since \(Y(t)\) of (11) is conservative and nondegenerate (i.e., \(\frac{1}{2}\sigma^{2}>0\)), by [32, p. 153], there exists a unique classical fundamental solution to (15). Thus it follows from [33, p. 365 and p. 368] that there exists a generalized solution \(u(t,\xi)\in L_{+}^{1}(\mathcal{R})\) for all \(t>0\), \(\xi \in \mathcal{R}\) provided that the initial density \(u_{0}\in L_{+}^{1}( \mathcal{R})\).
Motivated by [10], we define a Lyapunov function V:
where \(\alpha =\frac{\mu +\gamma }{\sigma^{2}}(\mathcal{R}_{0}^{s}1)>0\). Hence by [33, Theorem 11.9.1], we have the uniqueness and global asymptotic stability of the invariant density \(u_{\sigma } ^{s}\). From [10, Theorem A.7], we have assertions (i) and (ii).
We now prove assertion (iii). The stationary FPE associated with SDE (11) is given by
Rewrite system (16) in the following simpler form:
where \(y(\xi)=u_{\sigma }^{s}(\xi)\), c is a constant and
The solution of (17) is
with
where \(c_{0}=\frac{2(\mu +\gamma)}{\sigma^{2}}\).
It follows from the conditions \(u_{\sigma }^{s}(\xi)\geq 0\), \(\int_{0}^{\infty }u_{\sigma }^{s}(\xi)\,d\xi =1\) that \(c=0\) and
Simple computations reveal that
Note that \(\varGamma (c_{0}(\mathcal{R}_{0}^{s}1))>0\) for \(\mathcal{R} _{0}^{s}>1\), and thus C defined in (14) is finite. We can conclude that \(u_{\sigma }^{s}\) is an invariant which has the form
In the same way as in the proof of [10, Theorem A.6], we have
This, together with (18), implies assertion (iii). □
Numerical results via disease dynamics
In this section, we give some numerical results to show complex disease dynamic outcomes of SDE model (4) by using Milstein’s method [34], and the numerical scheme for model (4) under consideration is given by
where \(\eta_{k}\) (\(k=1,2,\ldots,n\)) are independent Gaussian random variables \(\mathbf{N}(0, 1)\), Δt is the time step size.
From (7), we note that \(\mathcal{R}_{0}^{s}=R_{0}\frac{ \sigma^{2}}{2(\mu +\gamma)}< R_{0}\), if \(R_{0}<1\), then \(\mathcal{R} _{0}^{s}<1\). We can know that if \(R_{0}<1\), \(I(t)\) goes to extinction for the deterministic model (4) (see [21]); and from Theorem 2.1, \(I(t)\) almost surely tends to zero exponentially with probability one for the stochastic model (4). Therefore we only consider the case of \(R_{0}>1\).
Following [22, 25], the choice for the following parameters remains unaltered:
and the initial value is \(I_{0}=10\). In this case, \(R_{0}=2.143>1\), and the deterministic model (6) has an unstable diseasefree equilibrium \(E_{0}=0\) and a unique globally stable endemic equilibrium \(E^{*}=34.4494\).
Next we will focus on the role of noise intensity σ on the resulting dynamics for SDE model (4).
Stochastic diseasefree dynamics
First of all, we adopt \(\sigma =0.405\), in this case, \(\mathcal{R} _{0}^{s}=0.97125<1\). From Theorem 2.1(i), we know that the disease \(I(t)\) will go extinct with probability one.
In Fig. 1, we show the stochastic disease dynamics of the evolution of the single path of \(I(t)\) obtained from two different numerical simulations run with the same parameters. One can see that \(I(t)\) is strongly oscillatory at the beginning and finally dies out with probability one. Easy to know that, in these cases above, \(I(t)\) tends to the diseasefree equilibrium \(E_{0}=0\) of the deterministic model (6) almost surely at last. It should be noted that, for model (6), \(R_{0}=2.143>1\), \(E_{0}=0\) is unstable. Hence we can conclude that, in this case, environmental noise can make unstable \(E_{0}\) to a stable one.
Furthermore, we repeat 10,000 simulations with the same parameters as in Fig. 1, we can calculate the average extinction time for \(I(t)\), and the results are shown in Fig. 2. For example, when \(\sigma =0.405\), the average extinction time for \(I(t)\); when \(\sigma =0.45\), it is 278.3943. We can conclude that the average extinction time decreases with the increase of noise intensity σ.
Stochastic endemic dynamics
In this subsection, we will focus on the stochastic endemic dynamics of (4) in the case of \(\mathcal{R}_{0}^{s}>1\). For this reason, we choose \(\sigma =0.01, 0.05, 0.10, 0.15\) implies that \(\mathcal{R}_{0} ^{s}=2.142\), 2.125, 2.071, 1.982, respectively. From Theorem 2.1(ii), we can conclude that the disease will persist almost surely. In Fig. 3, we show the single path of \(I(t)\) for model (4) and its corresponding deterministic model (6) with \(\sigma =0.01, 0.05, 0.10, 0.15, 0.25, 0.35\), respectively, and we can see that the solutions \(I(t)\) of SDE model (4) fluctuate around the endemic equilibrium \(E^{*}=34.4494\) of the deterministic model (6), respectively. In addition, we can find that the bigger noise intensity σ, the stronger oscillatory \(I(t)\).
For the sake of learning the effects of the intensity of noise σ on the stochastic disease dynamics of SDE model (4), we have repeated the simulation 10,000 times, keeping the same parameters as in Fig. 3 and never observing any extinction scenario up to \(t=500\). These results are respectively confirmed by the histograms and the probability density functions in Fig. 4, showing the stationary distributions of \(I(t)\) at \(t=500\) for model (4). And the numerical method for them can be found in [9].
From Fig. 4, one can see that the solution to SDE model (4) in the persistent case also suggests for lower σ (e.g., \(\sigma =0.01\) and 0.05), the amplitude of fluctuation is slightly and the oscillations to be more symmetrically distributed (cf. Figs. 4(a) and 4(b)), and the fluctuations are reflected at the stationary distributions. While for higher σ (e.g., \(\sigma =0.1\) and 0.15), the amplitude of fluctuation is remarkable and the distributions of the solutions are skewed (cf. Figs. 4(c) and 4(d)) and the fluctuations are also reflected at the stationary distributions.
Furthermore, in the case of \(\sigma =0.01\) and 0.05, the distribution of \(I(t)\) appears closer to a normal distribution (see Figs. 4(a) and 4(b)). Simple computations show that, when \(\sigma =0.01\), the distribution of \(I(t)\) closely obeys the normal distribution \(\mathbf{N}(34.4577, 1.6995^{2})\), and the values less than one standard deviation away from the mean account \((32.7582, 36.1572)\) for 67.94% of the set; while two standard deviations from the mean account \((31.1530, 37.7624)\) for 94.92%; and three standard deviations account \((30.0517, 38.8637)\) for 98.99% (see Fig. 5(a)). Simple calculations show that the skewness (i.e., the measure of the asymmetry of the probability distribution of a realvalued random variable about its mean) in this case is 0.1662598, which is a positive skew. And in the case of \(\sigma =0.05\), the distribution of \(I(t)\) closely obeys \(\mathbf{N}(33.7969, 8.8664^{2})\), and the values less than three standard deviations away from the mean account \((8.0924, 59.5015)\) for 98.75% (see Fig. 5(b)), the positive skewness is 1.045055.
In order to understand the effect of the intensity of noise σ on the skewness of the distribution of \(I(t)\), we show the four probability density functions with \(\sigma =0.01, 0.05, 0.1\), and 0.15 in Fig. 6. Easy to see that, as σ increases, the means of \(I(t)\) become smaller and smaller, and the positive skewness of the distributions of \(I(t)\) becomes bigger and bigger.
Concluding remarks
In this paper, we investigate the global threshold dynamics of the SDE SIS epidemic model (4) incorporating media coverage. After defining the basic reproduction number \(\mathcal{R}_{0}^{s}\), we establish a stochastic threshold theorem by using Feller’s test for explosions of solutions to onedimensional SDE model (4) (see Theorem 2.2). It is worthy to note that in the proof of this main result, we employ Feller’s test [10, 30, 31]. This is very different from the previous results by constructing Lyapunov function to prove the threshold theorem (see, [8, 22, 25]). And in the case of \(\mathcal{R}_{0}^{s}\), i.e., the disease persists with probability one, by studying the FPE associated with SDE model (4), we prove the existence, uniqueness, and global asymptotic stability of the invariant density of the FPE (see Theorem 3.1), which can be useful for us to understand the profile of the distribution of the process \(I(t)\).
Via numerical simulations, in the case of \(\mathcal{R}_{0}^{s}\leq 1\), we find that the average extinction time decreases with the increase of noise intensity σ. And in the case of \(\mathcal{R}_{0}^{s}>1\), we find that the solutions \(I(t)\) of SDE model (4) fluctuate around the endemic equilibrium \(E^{*}=34.4494\) of the deterministic model (6), and finally the distribution of \(I(t)\) seems like a normal stationary distribution. More precisely, for lower σ (e.g., \(\sigma =0.01\) and 0.05), the distribution of \(I(t)\) appears closer to a normal distribution (see Figs. 5(a) and 5(b)); while for higher σ (e.g., \(\sigma =0.1\) and 0.15), the amplitude of fluctuation is remarkable and the distribution of the solutions is positively skewed (cf. Figs. 4(c), 4(d), and 6). Obviously, variance increases with the increase of noise intensity σ, this is the main cause of positive skew distribution (see Figs. 5, 6 and also Figs. 4(c) and 4(d)). In this sense, we can claim that bigger noise σ will be beneficial to make \(I(t)\) stay away from the endemic equilibrium \(E^{*}\) of the deterministic model (6). In other words, increasing σ will be beneficial to control the disease spread. Thus, in order to control the spread of the disease, we must increase the intensity of noise σ.
On the other hand, in Theorem 2.2, we give the global threshold dynamics by Feller’s test for the explosions of solutions to onedimensional SDE model (4). Unfortunately, this method cannot be used to study a twodimensional SDE model (e.g., (1) or (2)). And the global threshold dynamics of a highdimensional SDE model, e.g., (1) or (2), is desirable in the future study.
References
 1.
Hethcote, H.W.: The mathematics of infectious diseases. SIAM Rev. 42(4), 599–653 (2000)
 2.
Brauer, F., CastilloChavez, C.: Mathematical Models in Population Biology and Epidemiology, 2nd edn. Springer, New York (2012)
 3.
Cai, Y., Wang, W.M.: Fishhook bifurcation branch in a spatial heterogeneous epidemic model with crossdiffusion. Nonlinear Anal., Real World Appl. 30, 99–125 (2016)
 4.
Cai, Y., Kang, Y., Banerjee, M., Wang, W.M.: Complex dynamics of a hostparasite model with both horizontal and vertical transmissions in a spatial heterogeneous environment. Nonlinear Anal., Real World Appl. 40, 444–465 (2018)
 5.
Anderson, H., Britton, T.: Stochastic epidemics in dynamic populations: quasistationarity and extinction. J. Math. Biol. 41(6), 559–580 (2000)
 6.
Truscott, J.E., Gilligan, C.A.: Response of a deterministic epidemiological system to a stochastically varying environment. Proc. Natl. Acad. Sci. 100(15), 9067–9072 (2003)
 7.
Britton, T.: Stochastic epidemic models: a survey. Math. Biosci. 225(1), 24–35 (2010)
 8.
Gray, A., Greenhalgh, D., Hu, L., Mao, X., Pan, J.: A stochastic differential equation SIS epidemic model. SIAM J. Appl. Math. 71(3), 876–902 (2011)
 9.
Cai, Y., Kang, Y., Banerjee, M., Wang, W.: A stochastic SIRS epidemic model with infectious force under intervention strategies. J. Differ. Equ. 259(12), 7463–7502 (2015)
 10.
Xu, C.: Global threshold dynamics of a stochastic differential equation SIS model. J. Math. Anal. Appl. 447(2), 736–757 (2017)
 11.
Wang, W.M., Cai, Y., Li, J., Gui, Z.: Periodic behavior in a FIV model with seasonality as well as environment fluctuations. J. Franklin Inst. 354(16), 7410–7428 (2017)
 12.
Liu, Y., Li, W., Feng, J.: Graphtheoretical method to the existence of stationary distribution of stochastic coupled systems. J. Dyn. Differ. Equ. 30(2), 667–685 (2018)
 13.
Liu, Y., Li, W., Feng, J.: The stability of stochastic coupled systems with timevarying coupling and general topology structure. IEEE Trans. Neural Netw. Learn. Syst. 29(9), 4189–4200 (2018)
 14.
Allen, E.: Environmental variability and meanreverting processes. Discrete Contin. Dyn. Syst., Ser. B 21(7), 2073–2089 (2016)
 15.
Liu, Q., Jiang, D., Hayat, T., Ahmad, B.: Analysis of a delayed vaccinated SIR epidemic model with temporary immunity and Levy jumps. Nonlinear Anal. Hybrid Syst. 27, 29–43 (2018)
 16.
Liu, Q., Jiang, D., Shi, N.: Threshold behavior in a stochastic SIQR epidemic model with standard incidence and regime switching. Appl. Math. Comput. 316, 310–325 (2017)
 17.
Cai, Y., Jiao, J., Gui, Z., Liu, Y., Wang, W.M.: Environmental variability in a stochastic epidemic model. Appl. Math. Comput. 329, 210–226 (2018)
 18.
Zhang, S., Meng, X., Wang, X.: Application of stochastic inequalities to global analysis of a nonlinear stochastic SIRS epidemic model with saturated treatment function. Adv. Differ. Equ. 2018(1), 50 (2018)
 19.
Wang, W.M., Cai, Y., Ding, Z., Gui, Z.: A stochastic differential equation SIS epidemic model incorporating Ornstein–Uhlenbeck process. Physica A 509, 921–936 (2018)
 20.
Wang, Y., Jiang, D., Alsaedi, A., Hayat, T.: Modeling a stochastic HIV model with logistic target cell growth and nonlinear immune response function. Physica A 501, 276–292 (2018)
 21.
Cui, J., Tao, X., Zhu, H.: An SIS infection model incorporating media coverage. Rocky Mt. J. Math. 38(5), 1323–1334 (2008)
 22.
Cai, Y., Kang, Y., Banerjee, M., Wang, W.M.: A stochastic epidemic model incorporating media coverage. Commun. Math. Sci. 14(4), 893–910 (2016)
 23.
Khasminskii, R.: Stochastic Stability of Differential Equations, vol. 66, 2nd edn. Springer, Berlin (2012)
 24.
Rudnicki, R.: Longtime behaviour of a stochastic prey–predator model. Stoch. Process. Appl. 108(1), 93–107 (2003)
 25.
Guo, W., Cai, Y., Zhang, Q., Wang, W.M.: Stochastic persistence and stationary distribution in an SIS epidemic model with media coverage. Physica A 492, 2220–2236 (2018)
 26.
Wang, W.D.: Epidemic models with nonlinear infection forces. Math. Biosci. Eng. 3(1), 267–279 (2006)
 27.
Bakhtin, Y., Hurth, T.: Invariant densities for dynamical systems with random switching. Nonlinearity 25(10), 2937–2952 (2012)
 28.
Bakhtin, Y., Hurth, T., Lawley, S., Mattingly, J.: Smooth invariant densities for random switching on the torus. Nonlinearity 31(4), 1331–1350 (2018)
 29.
Faggionato, A.: Nonequilibrium thermodynamics of piecewise deterministic Markov processes. J. Stat. Phys. 137(2), 259–304 (2009)
 30.
Feller, W.: Diffusion processes in one dimension. Trans. Am. Math. Soc. 77(1), 1–31 (1954)
 31.
Karatzas, I., Shreve, S.E.: Brownian Motion and Stochastic Calculus, 2nd edn. Springer, New York (1997)
 32.
Ito, H.M.: Ergodicity of randomly perturbed Lorenz model. J. Stat. Phys. 35(1), 151–158 (1984)
 33.
Lasota, A., Mackey, M.C.: Chaos, Fractals, and Noise: Stochastic Aspects of Dynamics. Springer, New York (1994)
 34.
Higham, D.J.: An algorithmic introduction to numerical simulation of stochastic differential equations. SIAM Rev. 43(3), 525–546 (2001)
Acknowledgements
The authors would like to thank the anonymous referees for very helpful suggestions and comments which led to the improvement of our original manuscript.
Funding
This research was supported by the National Natural Science Foundation of China (Grant No. 61672013, 11601179 and 11461073), the Natural Science Foundation of the Jiangsu Higher Education Institutions of China (16KJB110003), and Huaian Key Laboratory for Infectious Diseases Control and Prevention (HAP201704).
Author information
Affiliations
Contributions
All authors contributed equally to the writing of this paper. All authors read and approved the final manuscript.
Corresponding author
Ethics declarations
Competing interests
The authors declare that they have no competing interests.
Additional information
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Rights and permissions
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.
About this article
Cite this article
Yang, B., Cai, Y., Wang, K. et al. Global threshold dynamics of a stochastic epidemic model incorporating media coverage. Adv Differ Equ 2018, 462 (2018). https://doi.org/10.1186/s136620181925z
Received:
Accepted:
Published:
MSC
 92D30
 60H10
 93E15
Keywords
 Basic reproduction number
 Feller’s test
 Invariant density
 Fokker–Planck equations