臺灣女性癌症的空間分析

Spatial Analysis of Female Cancers in Taiwan

胡立諄(Li-Chun Hu) ； 賴進貴(Jinn-Guey Lay)

《臺灣地理資訊學刊》 4期 (2006/04) Pp. 39-55

https://doi.org/10.29790/JTGIS.200604.0004

摘要

地理資訊系統在醫學地理研究的應用已經逐漸普及，然而目前的應用大都以地圖呈現分布型態為主。流行病學界常透過疾病地圖來探索可能致病因子，然而視覺判定往往不夠精確。近年來空間分析方法發展迅速，使研究者得以利用空間統計方法來探索資料的空間特性；而空間迴歸模式將影響的空間連續性納入考慮，能更確切掌握因子的相關性。本研究利用行政院衛生署2003年出版的癌症地圖集，使用LISA (Local Indicator of Spatial Association)指標來探討女性癌症的分布型態是否呈現聚集，並分析不同時期癌症發生與死亡聚集的空間變化。結果顯示乳癌發生率有明顯的空間聚集特徵，本研究進一步探討社會經濟因子與乳癌發生率的相關性，透過空間誤差模式分析，考慮空間自相關所造成的干擾，使參數估計結果較傳統迴歸模式精確，顯示此一方法值得參考使用。

關鍵字

空間分析 ； 空間自相關 ； 醫學地理

並列摘要

GIS has been widely applied in medical geography research. Most applications, nevertheless, are limited to presenting distribution patterns. In epidemiology, disease maps are often used to explore the causes of diseases. Visual interpretation of these maps, however, usually leads to concerns of imprecision. This research seeks to take advantage of recent developments in spatial statistics to improve researchers' exploration of the spatial characteristics and patterns revealed by data. The inclusion of spatial regression models particularly can enhance our understanding of the relationships between various potential casual factors of diseases.
The case in discussion is whether the distribution of females with breast cancers exhibits spatial clustering. The source is gleaned from the statistical data of cancer atlas published by Taiwan’s Department of Health in 2003, which included age-standardized incidence and mortality rates. Local Indicator of Spatial Association (LISA) provides a useful index for the analysis of distribution patterns. The result yields that the occurrence of breast cancers exhibits spatial clustering. Utilizing spatial error regression model, we further find out the importance of considering the factor of spatial dependency apart from socioeconomic factors in tracing the causes of breast cancers.

並列關鍵字

spatial analysis ； spatial autocorrelation ； medical geography

參考文獻

American Institute for Cancer Research (AICR)(1999).Stopping Cancer BeforeIit Starts.New York:Golden Books.

Anselin, L.(1995).Local Indicators of Spatial Association-LISA.Geographical Analysis.27(2),94-115.

Spatial Econometrics

Exploring Spatial Data with GeoDaTM: A Workbook

Brody, J. G.,Aschengrau, A.,McKelvey, W.,Rudel, R. A.,Swartz, C. H.,Kennedy, T.(2004).Breast Cancer Risk and Historical Exposure to Pesticides from Wide-Area Applications Assessed with GIS.Environmental Health Perspecttives.112(8),889-897.