Cancer in Poland –Geographic Information System Approach

• Main Authors: Przemyslaw Maciej JURA, 游希傑
• Other Authors: Der-Ming Liou
• Format: Others
• Language: en_US
• Online Access: http://ndltd.ncl.edu.tw/handle/86108342149401213873

碩士 === 國立陽明大學 === 公共衛生研究所 === 98 === Objectives Present manuscript intends to deliver as basic framework facilitating detailed insight into the burden of cancer in Poland, taking into account small region spatial variability. Currently, there aren’t any extensive geographical analyzes of cancer in Poland. Methods The analysis is concentrated on two approaches: frequentist – calculation of standardized rates and ratios to describe phenomenon; Bayesian – to estimate Relative Risk of cancer incidence and mortality in focus region. Bayesian statistics include both spatial and aspatial methods with empirical and hierarchical estimation: Poisson-Gamma model, Marshall local smoothing, full Poisson-Gamma model, Besag-York-Mollie model. Moreover, analysis includes probability mapping, and cluster analysis with Kulldorff’s scan statistics. Relative Risk estimation concentrates on three types of cancer: lung cancer, breast cancer, prostate cancer for period 2002-2006. Data has been acquired from National Cancer Registry. Framework has been established using MySQL database, CRAN R software, SaTScan™ and WinBUGS/OpenBUGS. Results Conducted analysis gave detailed picture of spatial relative risk distribution for the three most prevalent cancers in Poland. Created disease maps of incidence/mortality relative risk visualize regional variability to identify potentially existing geographical disparities. Conclusions Using automatized scripts written in R language it is possible to create a framework for efficient analysis of count data. Variety of available methods, and high computational power facilitate effective ways to maintain periodical monitoring of the phenomenon of interest. Visualized relative risk for listed cancer types is a valuable beginning for further ecological analysis, to find regions with unusually heightened risk. Disease mapping has a great potential for public health decision makers to rationally allocate resources and plan interventions.