Drawing Inference About Multi-Serotype Infectious Diseases: A State-Space Model to Estimate the Duration of Cross-Protection Between Serotypes of Dengue Virus

• Activity Number: 139
• Type: Contributed
• Date/Time: Monday, July 30, 2012 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Nicholas G Reich*+
• Companies: University of Massachusetts
• Keywords: infectious disease ; public health ; biostatistics ; state-space models

Abstract

Many infectious diseases have multiple genetically-distinct pathogenic forms. Understanding the interactions between these different strains of the same disease is crucial to developing accurate models of disease dynamics at the population and individual level. To draw inference about these interactions, novel statistical methods that can incorporate and test different mechanistic models of the underlying disease process are required. As an example of such a model, we present an analysis of a 38-year time series with 12,197 dengue virus infections from an urban hospital in Bangkok, Thailand. Using a discrete-time state-space model, we estimated the level of cross-protection conferred by infection with one serotype of dengue against infection from another. Based on likelihood ratio tests, models including short-term cross-protection performed better than models without it. Assuming the duration of cross-protection follows an exponential distribution, we estimated an average duration of 23 months (95% CI: 11 to 52).