Mapping the Spatial Distribution of a Disease-Transmitting Insect in the Presence of Surveillance Error and Missing Data

• Activity Number: 348
• Type: Contributed
• Date/Time: Tuesday, July 31, 2012 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Andrew Hong*+
• Keywords: chagas disease ; surveillance

Abstract

Chagas disease is the most prevalent parasitic disease in the South America. Working with the Peruvian Ministry of Health (PMoH), we propose a Bayesian hierarchical model for the spatial distribution of the disease-transmitting insect vector. The primary focus of this study is addressing the difficulty in detection of the vector and how this difficulty translates into statistical uncertainty and impacts inference. In this work, we demonstrate through simulation the benefit of using surveillance information in terms of prediction accuracy. Further, simulated evidence demonstrates that failing to model surveillance uncertainty systematically under represents the spread and prevalence of the vector. On the other hand, our approach is shown to greatly improve performance, in spite of having weakly or mispecified priors. The result of this work is the computationally efficient sampling of the posterior distribution of vector presence, which smooths observationally, as well as spatially, in regions containing over 13,000 observations. The products of this work are currently being used by the PMoH as the guiding analysis in the on-going vector control campaign.