Abstract

A Bayesian spatiotemporal generalized linear regression with varying-coefficient model is proposed to examine geographic variation of the medical prescription use for ischemic stroke and to have a clearer understanding of association of the relevant risk factors such as comorbidities, medication and environmental society with mortality rate in ischemic stroke. The spatial heterogeneity of coefficients for important factors that may affect the mortality in ischemic stroke across 349 townships in Taiwan. By applying spatial-temporal models, we can understand the spatial variation in risk to ischemic stroke. It then turns out that, based on the findings, we can properly arrange medical resources and reduce the life-threatening damage caused by the uneven distribution of medical resources. The data was collected from a retrospective cohort study using 2004-2012 National health insurance research database.