Individual-Based Computational Model Used to Explain 2009 Pandemic H1n1 in Rural Campus Community

Abstract

In the beginning of fall semester 2009, over 2,000 students contacted the student health service at Washington State University to report symptoms of influenza. The epidemic in Pullman, WA made national news, and many speculated on the severity and extent of the disease spread. Analysis of data from the influenza A(H1N1)pdm09 epidemic in Pullman, WA offers an opportunity to gain insights into characteristics of this rural campus community outbreak. In this study, an individual-based stochastic epidemic simulation model was used with the data to estimate infection parameters and make projections of the number of symptomatic individuals that would result given a variety of plausible scenarios. The parameters that were estimated include the number of individuals initially infected and the basic reproductive ratio (R0). The model was then used to predict the magnitude of infection with vaccination, isolation and quarantine. The results show that the best single intervention strategy is vaccination, and the reduction in infection is greatest when vaccination, isolation and quarantine are used simultaneously.

Modelling the epidemic spread of an H1N1 influenza outbreak in a rural university town

“…Unlike the boarding school dataset (number of data points, n = 14) [18] or the island dataset (n = 50) [5], in our study we have a much larger dataset (n = 104). While other studies used simulation or other modelling approaches to predict future epidemic dynamics [8][9][10][11]14], in this work we used compartmental modelling to determine what mechanisms most likely played a role in the spread of this influenza epidemic in university town setting. Several characteristics distinguish this type of population from others, in that individuals are young adults aged between 18 and 26 years who are highly connected by the university, and they interact according to similar routines by way of adherence to the academic schedule.…”

“…In the case of 2009 pandemic H1N1, modelling approaches were used to predict epidemic dynamics [8][9][10][11], to assess the efficacy of the CDC planned vaccination scheme for the US epidemic [12], to examine the effect of control measures [13], to assess the potential severity and transmissibility of the pandemic [14], and to determine the contagious period of this viral strain [15] (for review, see [16]). …”

Estimating epidemic parameters: Application to H1N1 pandemic data

“…In particular, from the central limit theorem derived in [16], it follows that the Monte-Carlo approxi- Recall that m = μn for our SIR model discussed in Section 4.1, as well as θ = (μ, λ, γ ), with ODE (5) taking the form of an original Kermack-McKendrick ODE for susceptible (x), infectious (y), and removed (z) (see [11]). Therefore c θ (t) = (x(t), y(t), z(t)) is the solution oḟ (13) x(0) = 1 and y(0) = μ < 1, z(0) = 0. This is the basic classical model for epidemic disease spread within a fixed size population, with initial concentration (1 + μ).…”

“…The adjusted data for the most part fall within the approximate 95% confidence envelope for the stochastic model (10) which is the area between top and bottom curves. These confidence bounds are obtained by Monte-Carlo simulations of multiple stochastic trajectories of (10) (five of which are shown in gray for illustration) which fluctuate around the deterministic SIR trajectory (13) represented by a smooth middle curve (red). The model fitting is truncated on day 101 so the final epidemic phase (after day 90) is not well-captured.…”

“…Swine flu epidemic data. H1N1 daily infective counts on the Pullman campus of WSU for the period of 101 days in Fall 2009, as in[13].…”