Research Collaborations: Laurie Ainsworth

Laurie Ainsworth: Mapping disease or mortality rates is a useful way of displaying the geographic variation in rates and in identifying regions of high and low rates for further follow-up in a surveillance context. Raw rates are not reliable for mapping as they tend to be highly variable and imprecise for small areas. Poisson models which accommodate spatial variation provide smoothed risk estimates. In some cases though, the data contain excess zeros. These may arise because of immunity or other protective factors. In the spatial analysis of pine weevil attacks in British Columbia, for example, the notion of resistant trees is currently hypothesized. This thesis develops methods for handling such excess zeros.