Monday, August 2, 2010 - 1:50 PM

COS 12-2: Using Bayesian techniques to estimate the force of infection from serologic data

Alicia M. Ellis1, Thomas Scott2, Aaron King3, Amy Morrison2, Sharon Minnick4, Claudio Rocha5, Brett Forshey5, Steven T. Stoddard6, Kevin Russell5, James Olson5, Patrick Blair5, Douglas Watts5, Moises Sihuincha7, and Tadeusz Kochel5. (1) Research & Policy in Infectious Disease Dynamics, (2) University of California - Davis, (3) University of Michigan, (4) UC-Davis, (5) Naval Medical Research Center, (6) University of California, Davis, (7) Loreto Regional Reference Laboratory

Background/Question/Methods

Although the force of infection (i.e., the rate at which susceptible individuals become infected) is a useful measure for comparing transmission dynamics across space and time, it is difficult to estimate.  This is because not all disease cases are reported or recorded, some are asymptomatic and not detected, and it is difficult to determine the number of susceptible individuals that were exposed.  To obtain better estimates of the force of infection for a given area, better data and/or better statistical techniques are needed to handle uncertainties in the data that are available. Here, we present: (1) data from a longitudinal cohort study of dengue transmission and entomology in Iquitos, Peru from 1999-2005, (2)  Bayesian analysis of serostatus data to obtain more accurate estimates of the force of infection, and (3) preliminary analyses of the associations between the force of infection, weather, and entomological measures.  Results/Conclusions

The 3 dengue serotypes had surprisingly similar dynamics despite the fact that dengue-3 was newly introduced into the population and epidemic during this time period.  The force of infection was associated with minimum and average temperature, but not maximum temperature, presumably because of their influence on the extrinsic incubation period in the mosquito vector.  We are currently analyzing the long-term entomological data that includes estimates of the proportion of positive containers, number of pupae, and number of adult females in each house sampled across the city.  We will test the hypothesis that increases in disease transmission are related to increases in the mosquito population, and explore the spatial dynamics of the force of infection.

This study is broadly important for disease ecology because it provides a novel statistical technique to obtain more accurate estimates of the force of infection from serostatus data.  More specifically, this study will provide crucial information for improving monitoring and control strategies for dengue by identifying the factors that covary with the force of infection.