Spatio-temporal Modelling of Dengue Risk: Towards an Early Warning System for Brazil
Speaker: Professor Trevor Bailey, College of Engineering, Mathematics and Physical Sciences, University of Exeter
Date: Friday 2nd October 2015
Time: 3pm – 4pm
Location: Room 1.01, Seminar Room, Agriculture Building
Abstract:
The transmission of many infectious diseases can be affected by weather and climate variability, particularly those spread by arthropod vectors such as malaria and dengue. Previous epidemiological studies have demonstrated statistically significant associations between the incidence of such infectious diseases and climate variability, and have highlighted the potential for developing climate-based early warning systems for associated demics.
Dengue fever is now one of the most important emerging climate sensitive tropical diseases worldwide and Brazil experiences a higher morbidity from this disease than any other country. This talk describes results from a Brazil-UK collaborative project that modelled the spatio-temporal variation in dengue fever risk in Brazil using climate and non-climate information, with a view to
developing a national early warning system for dengue epidemics.
A negative binomial generalised linear mixed model is developed which makes allowances for unobserved confounding factors through a combination of structured and unstructured spatio-temporal random effects. The resulting spatio-temporal hierarchical model is implemented via a Bayesian framework using Markov Chain Monte Carlo (MCMC). Using the model, posterior predictive distributions for disease risk can be derived at each spatial location for a given month or season. This allows probabilistic forecasts to be issued and forecast uncertainty to be quantified.
Series: Statistics
Social Media Links