Event Type:

Mathematical Biology Seminar

Date/Time:

Thursday, November 15, 2012 - 05:00 to 06:00

Location:

STAG 107

Event Link:

Guest Speaker:

Institution:

Clemson University

Abstract:

The study of dengue dynamics at the population scale have significantly contributed to the

understanding of dengue transmission. Most studies have used point estimates of parameter

values derived from clinical and laboratory experiments: in particular, data on population-level

parameters such as transmission or susceptibility are extremely limited due to inability to fea-

sibly conduct experiments of infection in people and instead must be estimated from indirect

population-level data. We suggest a Bayesian approach which uses Monte Carlo Markov Chain

(MCMC) simulation to find estimates for the unknown parameters of a generic dengue mathe-

matical model we formulated based on previous dengue models. Prior knowledge is combined

with data on hospital visits to perform the statistical inference on the unknown parameters.

Our model allows for the inclusion of different hypotheses about dengue epidemiology and we

explore the consistency of clinical data with the epidemiological hypothesis by determining

goodness of fit of the model to the data for each combination of hypothesis. We use Akaike

Information Criterion on the results from the Bayesian MCMC on our dengue model and select

a model that most parsimoniously agrees with the data. A dengue vaccine is expected to be

available within 3-4 years and we explore the vaccine allocation policy to curb the dengue infec-

tion. Specifically, we consider the possibility of a transient period when instantaneous number

of infectious individuals can be higher than what the infectious number of individuals would

have been without the use of vaccination.