Abstract:
|
Analyses of epidemics are complicated by several factors, including the fact that the true dispersal mechanism of disease agents and the precise infection times of patients are often unobserved. Instead, we often observe the infection state of each unit at discrete time intervals. For example, consider a recent study of the Chagas disease vector Triatoma infestans in Arequipa, Peru. The data are limited to observed insect presence at each household at three time points over several years. In addition, we observe the number of insects at each household, although with measurement error. To address these challenges, we propose a novel susceptible-infected-observed-removed model that uses informative priors, incorporates the counts of vectors and complex dispersal dynamics observed in Arequipa. The fully Bayesian method is used to augment the data, estimate the dispersal parameters, and determine posterior infestation risk probabilities of households for future treatment. We investigate the properties of the model with simulation studies. Finally, the proposed methods are illustrated with an analysis of the Chagas disease vector data.
|
ASA Meetings Department
732 North Washington Street, Alexandria, VA 22314
(703) 684-1221 • meetings@amstat.org
Copyright © American Statistical Association.