背景：找出高死亡率地理聚集(熱點)與相關社經因素是降低死亡率地理不平等的重要分析。但是，使用不同地理單元的分析，常常會有不一致的結論。
目的：探討不同地理單元找出死亡率熱點與社經因素相關是否有差異。
方法：首先以人工比對之方式將全國95-99年死因登記檔之地址進行比對與定位經緯度，共計692,372筆。接著以鄉鎮市區與村里兩種地理單元計算標準化死因別死亡率，分別使用95%信賴區間與空間自相關(Moran’s I與LISA)檢定，找出在兩種地理單元的死亡率熱區。最後探討不同地理單元死亡率與社經指標相關係數之差異。
結果：本研究選擇都市化程度較低的台東縣，以及都市化程度高一點的台南市做為熱區與相關性分析的範例，因台南市共37個區，其地理高度與人口發布差異大，因此區內異質性高，因此以這兩個縣市作為呈現範例。台東縣村里全死因空間自相關值0.07，台南市空間自相關值-0.03，顯示台東縣的全死因聚集程度高於台南市。在空間自相關的熱區分析上，台東縣以海端鄉加拿村、崁頂村、利稻村，以及延平鄉的永康村、鸞山村共5個村里是死因別死亡率高的熱區。台東縣海端鄉是台東縣面積最大的鄉鎮，鄉內人口90%為布農族原住民，肝硬化的死亡率上，可發現海端鄉加拿村肝硬化死亡率是全部村里中最高。台南市的部分以七股區七股里、西寮里、三股里，龍崎區牛埔里、楠坑里，後壁區菁寮里與左鎮區二寮里共7個村里是死因別死亡率高的熱區。不同地理單元所造成的影響，在台東惡性腫瘤死亡率上，從鄉鎮單元來看，海端鄉、延平鄉、鹿野鄉皆為冷區，只有一個熱區在東河鄉，但在村里單元海端鄉霧鹿村、海端鄉崁頂村、鹿野鄉永安村是熱區，且東河鄉沒有一個熱區村里，整個村里單元下的熱區共有8個。台東的心臟疾病，在鄉鎮市單元會有2個熱區是池上鄉與大武鄉，但在村里單元變成海端鄉海端村才是熱區，池上鄉沒有村里為熱區，而大武鄉內的尚武村是鄉內唯一村里單元的熱區。台東的腦血管在鄉鎮市單元下看沒有任何熱區，在村里單元則有5個集中在海端鄉。台東的肝硬化死亡率的分布在鄉鎮市單元的熱區為海端鄉與延平鄉，從村里單元除利稻村以外，海端鄉與延平鄉的村里皆為熱區。在台南的熱區分析結果，全死因在區單元僅2個熱區－楠西區與北門區，村里單元則為46個，差異非常大，且村里單元下的楠西區沒有任何村里熱區。腦血管在區單元2個熱區，村里為24個，原本南化區為區單元下的熱區之一，但在村里單元僅剩東和里為熱區村里，其餘熱區村里分散各區。在事故傷害則相反，區單元下3個熱區，南化區並非其中之一，村里單元下的熱區有35個，南化區的4個村里為熱區。糖尿病區單元下1個熱區，村里單元有22個。慢性下呼吸道區單元下3個熱區，村里則有34個，且區單元下白河區並非熱區，但村里單元下的白河區有3個村里聚集情形。以上顯示不同地理單元造成對熱區的結論不一致情形。另以死因別死亡率進行相關性分析，結果死因別為腦血管疾病、惡性腫瘤與糖尿病，會在鄉鎮市區與村里有相關性方向相反的問題。台南市的事故傷害與國高中職比例關係，在區顯示-0.379的負相關，村里單元為0.016的低度正相關，可能為在區單元時僅龍崎、南化區為熱區，其餘原台南市區的安平區、東區等事故傷害死亡率低，因此為負相關，但在村里單元時，熱區數目多，除了龍崎嶇楠坑里外，其他村里平均事故傷害死亡率也比較高，而造成低度正相關。顯示MAUP問題會影響死因別死亡率與社會經濟指標的相關分析。
結論：本研究使用村里單元進行死亡率熱點分析，的確比鄉鎮市區單元更能精準找出高健康風險地區。本研究也發現某些死因不同地理單元相關性出現較明顯差異，未來還應該進一步透過田野調查方式進一步了解其中機轉。

INTRODUCTION: Understanding the differences of mortality hotspot in different geographic units and its spatial correlation with socioeconomic indicators are prominent to explore the high risk area that needs more resource attention. However, using different scale of geographic units often leads to inconsistent conclusions.
METHODS: Firstly, the mortality data was obtained from Taiwan mortality data, 2006 to 2011, then address converted into latitude and longitude through the statistical area access service. Then, the standardized mortality rate was calculated by the two geographic units in township and village. The 95% confidence interval and spatial autocorrelation (Moran’s I and LISA) were used to identify the mortality hotspots in the two geographic units. Finally, the differences between the mortality and socioeconomic indicators in Pearson's correlation coefficients in different geographic units are discussed.
RESULTS: The auto-correlation of the total cause of death in Taitung County was 0.07, and Tainan City was -0.03. Taitung County has significant mortality hot spots located in Haiduen Township (Jiana Village, Yuding Village, Lidao Village), and Yanping Village ( Yongkang Village and Lushan Village). Regarding to Tainan, it has significant mortality hot spots located in Nanxi District, Nanhua District, Zuozhen District and Beimen District. The correlation analysis between socioeconomic and cause of death cerebrovascular disease, malignant tumor and diabetes revealed the opposite direction in the township and village scales.
CONCLUSION: We found inconsistent hotspots in different scale of geographic units. This study also found that some causes of death were related to the correlation of socioeconomic indicators in different geographic units. In the future, we should explore the mechanism through field investigation.

Barnes, T. L., Colabianchi, N., Hibbert, J. D., Porter, D. E., Lawson, A. B., & Liese, A. D. (2016).Scale effects in food environment research: Implications from assessing socioeconomic dimensions of supermarket accessibility in an eight-county region of South Carolina. Applied Geography,68, 20-27. doi:10.1016/j.apgeog.
Casper ML, B. E., Williams G. I., Halverson J. A, Braham V. E, & Greenlund K. J. (2003). Atlas of stroke mortality: racial, ethnic, and geographic disparities in the United States. Atlanta, GA: Department of Health and Human Services, Centers for Disease Control and Prevention.
Caudeville, J., Bonnard, R., Boudet, C., Denys, S., Govaert, G., & Cicolella, A. (2012).Development of a spatial stochastic multimedia exposure model to assess population exposure at a regional scale. Science of the Total Environment, 432, 297-308. doi:10.1016/j.scitotenv.
Dobson, J. E., Brlght, E. A., Coleman, P. R., Durfee, R. C., & Worley , B. A. (2000).LandScan: A global population database for estimating populations at risk. American Society for Photogramm and Remote Sensing, 66(7), 849-857.
Duncan, D. T., Kawachi, I., Subramanian, S. V., Aldstadt, J., Melly, S. J., & Williams, D. R. (2014). Examination of how neighborhood definition influences measurements of youths' access to tobacco retailers: a methodological note on spatial misclassification. American Journal of Epidemiology, 179(3), 373-381. doi:10.1093/aje/kwt251.
Dunn, J. R., Burgess, B., & Ross, N. A. (2005).Income distribution, public services expenditures, and all cause mortality in US states. Journal of Epidemiology and Community Health, 59(9), 768-774. doi:10.1136/jech.
Fontanella, C. A., Saman, D. M., Campo, J. V., Hiance-Steelesmith, D. L., Bridge, J. A., Sweeney, H. A., & Root, E. D. (2018).Mapping suicide mortality in Ohio: A spatial epidemiological analysis of suicide clusters and area level correlates. Preventive Medicine, 106, 177-184. doi:10.1016/j.ypmed.
Francis P. Boscoe , Linda. W. P. (2013).Choosing geographic units for choropleth rate maps, with an emphasis on public health applications. Geography and Geographic Information Science, 30(3), 237-248.
Garb, J. L., Cromley, R. G., & Wait, R. B. (2007).Estimating populations at risk for disaster preparedness and response. Journal of Homeland Security and Emergency Management, 4(1), 1-17.
Greer, S., Schieb, L. J., Ritchey, M., George, M., & Casper, M. (2016). County health factors associated with avoidable deaths from cardiovascular disease in the United States, 2006-2010. Public Health Reports, 131(3), 438-448. doi:10.1177/003335491613100310.
Hay, S. I., Noor, A. M., Nelson, A., & Tatem, A. J. (2005). The accuracy of human population maps for public health application. Tropical Medicine and International Health, 10(10), 1073-1086. doi:10.1111/j.1365-3156.2005.01487.x.
Hsieh, Y. H. (2017). Ascertaining the impact of catastrophic events on dengue outbreak: The 2014 gas explosions in Kaohsiung, Taiwan. PLoS One, 12(5), e0177422. doi:10.1371/journal.pone.0177422.
Kim, D. (2016). The associations between US state and local social spending, income inequality, and individual all-cause and cause-specific mortality: the national longitudinal mortality study. Preventive Medicine, 84, 62-68. doi:10.1016/j.ypmed.
Kim, H., & Yao, X. (2010).Pycnophylactic interpolation revisited: integration with the dasymetric-mapping method. International Journal of Remote Sensing, 31(21), 5657-5671.
Kuo, C. T., & Chen, D. R. (2017). Double disadvantage: income inequality, spatial polarization and mortality rates in Taiwan. Journal of Public Health. doi:10.1093/pubmed/fdx179.
Kwan, M. P. (2009).From place-based to people-based exposure measures. Social Science & Medicine, 69(9), 1311-1313. doi:10.1016/j.socscimed.
Kwan, M. P.(2012). The uncertain geographic context problem. Annals of the American Association of Geographers, 102(5), 958-968.
Lee, C. C., Chien, S. H., Hung, S. K., Yang, W. Z., & Su, Y. C. (2012).Effect of individual and neighborhood socioeconomic status on oral cancer survival. Oral Oncology, 48(3), 253-261. doi:10.1016/j.oraloncology.2011.10.002.
LIoyd, C. D. (2014). The modifiable areal unit problem (1 st Eds.), Exploring Spatial Scale in Geography (pp. 29-38). Hoboken, New Jersey: John Wiley & Sons.
Lu, T. H., Li, S. T., & Yin, W. H. (2017). Graphical representations of mortality data with confidence intervals. Circulation: Cardiovascular Quality and Outcomes, 10(3). doi:10.1161/CIRCOUTCOMES.116.002763.
Marceau, D. J., & Hay, G. J. (1999).Remote sensing contributions to the scale issue. Canadian Journal of Remote Sensing, 25(4), 357-366.
McDavid H. K., Ling, Q., Song, R., & Hall, H. I. (2008).County-level socioeconomic status and survival after HIV diagnosis, United States. Annals of Epidemiology, 18(12), 919-927. doi:10.1016/j.annepidem.
Mellor, J. M., & Milyo, J. D. (2004).Individual health status and racial minority concentration in US states and counties. American Public Health Association, 94(6), 1043-1048.
Openshaw, S. (1984). The Modifiable Areal Unit Problem, Concepts and Techniques in Modern Geography 38. Geobooks, Norwich, England.
Parenteau, M. P., & Sawada, M. C. (2011).The modifiable areal unit problem (MAUP) in the relationship between exposure to NO2 and respiratory health. International Journal of Health Geographics, 10, 58. doi:10.1186/1476-072X-10-58.
Pedigo, A., Seaver, W., & Odoi, A. (2011). Identifying unique neighborhood characteristics to guide health planning for stroke and heart attack: fuzzy cluster and discriminant analyses approaches. PLoS One, 6(7), e22693. doi:10.1371/journal.pone.0022693.
Phillips, J. A. (2013). Factors associated with temporal and spatial patterns in suicide rates across U.S. states, 1976-2000. Demography, 50(2), 591-614. doi:10.1007/s13524-012-0176-y.
Ross, N. A., Tremblay, S. S., & Graham, K. (2004). Neighbourhood influences on health in Montreal, Canada. Social Science & Medicine, 59(7), 1485-1494. doi:10.1016/j.socscimed.2004.01.016.
Roth, G. A., Dwyer-Lindgren, L., Bertozzi-Villa, A., Stubbs, R. W., Morozoff, C., Naghavi, M.,&Murray, C. J. L. (2017).Trends and Patterns of Geographic Variation in Cardiovascular Mortality Among US Counties, 1980-2014. The Journal of the American Medical Association, 317(19), 1976-1992. doi:10.1001/jama.2017.4150.
Saib, M. S., Caudeville, J., Carre, F., Ganry, O., Trugeon, A., & Cicolella, A. (2014).Spatial relationship quantification between environmental, socioeconomic and health data at different geographic levels. International Journal of Environmental Research and Public Health, 11(4), 3765-3786. doi:10.3390/ijerph110403765.
Scott, L., Mobley, L. R., & Il'yasova, D.(2017). Geospatial analysis of inflammatory breast cancer and associated community characteristics in the United States. International Journal of Environmental Research and Public Health, 14(4). pii: E404. doi:10.3390/ijerph14040404.
Skinner, J., Weinstein, J. N., Sporer, S. M., & Wennberg, J. E. (2003).Racial, ethnic, and geographic disparities in rates of knee arthroplasty among Medicare patients. The New England Journal of Medicine, 349(14), 1350-1359. doi:10.1056/NEJMsa021569.
Tagashira, N., & Okabe, A. (2002). The modifiable areal unit problem in a regression model whose independent variable is a distance from a predetermined point. Geographical Analysis,34(1), 1-20.
Talbert-Slagle, K. M., Canavan, M. E., Rogan, E. M., Curry, L. A., & Bradley, E. H. (2016).State variation in HIV/AIDS health outcomes: the effect of spending on social services and public health. AIDS, 30(4), 657-663. doi:10.1097/QAD.0000000000000978.
Tian, N., Goovaerts, P., Zhan, F. B., & Wilson, J. G. (2010). Identification of racial disparities in breast cancer mortality: does scale matter? International Journal of Health Geographics, 9, 35. doi:10.1186/1476-072X-9-35.
Tom Koch , K. D. (2007).Medical mapping: The revolution in teaching—and using—maps for the analysis of medical issues. Journal of Geography, 103(2), 76-85.
WCHD. (2017). Health Inequities by Census Tract- 2017 Health Equity Assessment Summary Retrieved from http://www.williamscountyhealth.org/wp-content/uploads/2014/09/Health-Inequity-Report_FINAL.pdf
Weichselbaum, J., Petrini-Monteferri, F., Papathoma, M., Wagner, W., & Hackner, N. (2005).Sharpening census information in GIS to meet real-world conditions: the case for Earth Observation . Wessex Institute of Technology Transactions on Ecology and the Environment, 1, 143-152.
Woodcock, C. E., & Strahler, A. H. (1987).The factor of scale in remote sensing. Remote Sensing of Environment ,21(3), 311-332.
Yang, T. C., & Jensen, L. (2015).Exploring the inequality-mortality relationship in the US with Bayesian spatial modeling. Population Research and Policy Review, 34(3), 437-460. doi:10.1007/s11113-014-9350-9.
Yao, N., Foltz, S. M., Odisho, A. Y., & Wheeler, D. C. (2015).Geographic analysis of urologist density and prostate cancer mortality in the United States. PLoS One, 10(6), e0131578. doi:10.1371/journal.pone.0131578.

中文部分
內政部。社會經濟資料服務平台. 取自https://segis.moi.gov.tw/STAT/Web/Portal/STAT_PortalHome.aspx
吳焄雯 (2015)。 國土資訊系統社會經濟資料庫之推動與應用. 國土及公共治理季刊, 3(1), 101-107.
呂宗學 (2006)。不同科學典範 “看到” 不同東西？以1854年倫敦霍亂流行調查為例. 成大醫訊, 17(2), 26-30.
林季平與王永慈 (2012)。 社會發展指標資料庫及地理資訊系統. 國土資訊系統通訊, 84.
林慧淳 (2001)。地區剝奪與死亡率之相關：以台灣為例 (未出版之碩士論文)。國立台灣大學，台北市。
黃任薇 (2007)。 GIS網格解析度之研究.(碩士論文)。取自臺灣博碩士論文系統（系統編號 095NCKU5347031）。
劉麗雯，張利能 (2004)。 GIS在地方政府賑災組織資源管理上的運用：以九二一震災的社會服務資源管理體系建構為例. 東吳社會學報, 16, 1-31.
衛生福利部。105年衛生福利部死因統計. 取自 http://www.mohw.gov.tw/cp-16-33598-1.html
羅悦之 (2017)。台灣死因別死亡率之社會經濟不平等（1971-2012）：生態研究 (未出版之碩士論文)。國立台灣大學，台北市。