Date@ F2005/5/20(fri)15:00|
Venue@ FDw601

Invited Speaker: Prof. H. Nishiura(Universitat Tubingen)

Title:
Statistical Approaches to the Transmission Dynamics - Bridging infectious
disease epidemiology and mathematical modeling

Abstract:
The key epidemiologic determinants or parameters play one of the most
important parts for mathematical studies of infectious diseases. Previously
developed quantitative approaches, such as given in Bailey (1957) or
Anderson & May (1991), sometimes make us suspect of invalid methodologies,
technical flaws and huge uncertainties. Although most of them attribute to
the unrealistic assumptions or incomplete models, some can be elaborated
through the recent statistical knowledge. Here I show two examples using
life-data on bioterrorism associated diseases through historical analyses
(i.e., plague and smallpox). One is novel and simple approach which uses
transmission network (who infected whom) for statistical estimation of
transmission potential, namely basic reproduction number R0, of pneumonic
plague. Different bootstrap algorithms enabled us to reconstruct epidemic
trees, while two steps of likelihood-based approaches give us the
distributions (estimates) of R0 as well as serial interval. Another example
is an application of statistical modeling to estimate the duration of
vaccine-induced immunity against smallpox infection. Bivariate Poisson count
data model and binomial sampling process are demonstrated to be very useful
tools with age-stratified notification data. Durations of the direct and
indirect effects of smallpox immunization are reasonably estimated, and
obtained results provide us with significant public health implications.
Another implications for modeling approaches, such as in simulation studies,
following obtained estimates will be discussed.