Monitoring and prediction of an epidemic outbreak using syndromic observations

Abstract

The paper presents a method for syndromic surveillance of an epidemic outbreak due to an emerging disease, formulated in the context of stochastic nonlinear filtering. The dynamics of the epidemic is modeled using a stochastic compartmental epidemiological model with inhomogeneous mixing. The syndromic (typically non-medical) observations of the number of infected people (e.g. visits to pharmacies, sale of certain products, absenteeism from work/study, etc.) are assumed available for monitoring and prediction of the epidemic. The state of the epidemic, including the number of infected people and the unknown parameters of the model, are estimated via a particle filter. The numerical results indicate that the proposed framework can provide useful early prediction of the epidemic peak if the uncertainty in prior knowledge of model parameters is not excessive.

Forecasting influenza outbreak dynamics in Melbourne from Internet search query surveillance data

“…Dawson et al ., building on an earlier study that used synthetic surveillance data,6 combined a particle filter with an SEIR infection model for the purposes of early outbreak detection, using a dynamic Bayesian network to assimilate disease surveillance observations provided by an agent‐based model 31. The emphasis on this study was the use of this method for early outbreak detection (e.g.…”

“…Stochastic noise is included in the flows between compartments and in the model parameters, as per Skvortsov and Ristic 6. The entire population is assumed to be susceptible at the start of the calendar year, and an initial exposure occurs with daily probability p seed .…”

“…A variety of recursive Bayesian estimation methods (‘filters’) have been used for such forecasting purposes,1, 2 often in combination with Internet search query surveillance data and mechanistic models of infection 3, 4, 5, 6, 7, 8…”

Combining syndromic surveillance and ILI data using particle filter for epidemic state estimation

“…One of the tools commonly used for epidemic state estimation and prediction is a recursive Bayesian state estimation technique, and many examples of research using this technique can be found in the literature [e.g., Dukic et al (2012), Jegat et al (2008), Ong et al (2010), Vidal Rodeiro andLawson (2006), Shaman and Karspeck (2012), Skvortsov and Ristic (2012)]. Bayesian state estimation assumes some knowledge on the underlying dynamics of a system (system model), and recursively updates the degree of belief in system states by using sequentially available observation data.…”

“…Since in most cases the underlying model is not fully knowni.e., epidemic parameters in the model are typically unknown, these methods often estimate epidemic parameters as well as state variables. For example, Dukic et al (2012), Ong et al (2010), Skvortsov and Ristic (2012) use epidemic equations as a system model, and formulate a Bayesian filtering problem to estimate epidemic parameters and state variables. In the method developed in Dukic et al (2012), emphasis is placed on learning of the epidemic parameters.…”