Stochastic FluSiM for influenza transmission dynamics

Influenza A (H1N1) global outbreak in 2009 highlighted the need to better understand the dynamics of influenza transmission. To assist health authorities in decision making during influenza pandemics, models are used to project the spread of influenza and to evaluate the cost benefit of intervention such as quarantine and vaccination. The deterministic epidemiology model SIR (susceptible-infectiverecovered) is widely used by the United States Centers for Disease Control and Prevention (US CDC) and many researchers. In USM, an in-house SIR model codenamed FluSiM was developed to track the evolution of influenza such as H1N1 in a specified population and to evaluate benefit of interventions. FluSiM has been validated using the 2009 Influenza A (H1N1) data for Malaysia. A more realistic analysis and projection of the spread of influenza can be obtained by incorporating biological and epidemiological heterogeneity into the model. In this paper, the in-house deterministic FluSiM model is enhanced into a stochastic model by allowing vital disease transmission process parameters, to randomly change with time, given specified means and standard deviations. These stochastic disease transmission parameters that control the progression of influenza follow a normal distribution within the reported range. This stochastic FluSiM model allows us to investigate the uncertainty of disease evolution from its onset. Simulation results of the stochastic FluSiM model show the general bell-shaped epidemic curve, but occasionally exhibit disease evolution that is beyond the range of deterministic SIR.