研究目的 過往探討篩檢行為及其影響因素之研究僅探討個人因素與個人篩檢行為間之關係,鮮少探討地區因素對個人篩檢行為之影響,因此本研究欲應用多層次分析探討個人因素、地區因素與個人篩檢行為間之關係。 材料方法 本研究以基隆市民國95年戶籍資料為基準,連結民國88年至民國95年之基隆整合式篩檢(Keelung Community-based Integrated Screening,簡稱KCIS)資料,與截至民國96年5月之低收入戶資料。以Andersen之健康行為模式為研究架構基礎,探討個人之傾向因素(年齡、性別、婚姻狀況、教育程度)、個人之使能因素(低收入身份與KCIS經驗)以及地區因素(低收入人數比例、高教育程度比例、老年人口比例、行政區以及篩檢涵蓋率)與個人篩檢行為間之關係。分析民國95年設籍於基隆市之20歲以上民眾,排除不適當樣本後,共298,796人納入分析。本研究以SAS 9.1版統計軟體進行描述性統計與雙變項分析,以HLM 6.0版統計軟體進行多變項分析。 研究結果 本研究結果顯示影響民眾接受95年KCIS服務可能性之個人傾向因素為年齡、性別、婚姻狀況及教育程度。地區因素經標準化後,民眾接受95年KCIS服務可能性之地區層次影響因素為低收入人數比例、高教育程度比例、行政區及篩檢涵蓋率。地區因素之地區社會經濟狀況對個人因素與個人篩檢行為之交互作用顯示,低收入人數比例影響個人因素之性別與個人篩檢行為之關係、個人因素之教育程度與個人篩檢行為之關係、個人因素之婚姻關係與個人篩檢行為之關係、個人因素之KCIS經驗與個人篩檢行為之關係、個人因素之低收入身份與個人篩檢行為之關係;高教育程度比例影響個人因素之年齡與個人篩檢行為之關係、個人因素之性別與個人篩檢行為之關係、個人因素之教育程度與個人篩檢行為之關係。 結論與建議 本研究證實地區因素對個人篩檢行為有其影響力,且地區因素亦會影響個人因素與個人篩檢行為間之關係。依據研究結果提出以下建議:1. 本研究使用次級資料分析,使用之變項有限,未來研究可利用調查資料或健保就醫記錄,以補足本研究之不足。2. KCIS以深入社區設站的方式提供民眾健檢服務,相關衛生單位應進一步了解未接受KCIS服務之民眾是否接受其他篩檢服務,或者針對這群民眾訂定特定的介入策略與不同的服務提供方式。
Objectives: Previous studies demonstrated the relationship between individual factors and screening behaviors, but they seldom investigated the relationship between area factors and screening behaviors. The aim of this research was to investigate the associations between factors at area and individual level and screening behaviors by using multilevel analysis with a focus on area factors. Methods: The study applies mainly the census-registered database combining 1999 to 2006 Keelung Community-base Integrated Screening (KCIS) database and the low-income database up to May 2007. After deleting inappropriate subjects, there were 298,796 inhabitants over 20 years of age registered in Keelung City by 2006. The relationship of 2006 KCIS variables, individual variables (age, gender, martial status, education, and low-income) and area variables (low-income percentage, high-education percentage, aging percentage, KCIS coverage rate and district) are examined using multilevel logistic analysis through Health Behavior Model proposed by Andersen RM. Results: The participation rate was 26.95% for inhabitants registered in Keelung City. Older, women, married or divorced or widowed, higher education and having KCIS experiences were significantly associated with the likelihood of participating 2006 KCIS. For area variables, living in a lower low-income percentage area, living in a higher high-education percentage area, district, and lower KCIS coverage rate area were associated with the likelihood of participating 2006 KCIS. The cross-level interactions were found in the study. The low-income percentage and high-education percentage had impacts on the relationship of individual factors and screening behaviors. Conclusion: The study demonstrates the associations between factors at area and individual level and screening behaviors, especially the cross-level interactions. People with lower socio-economic status are less likely to participated 2006 KCIS in Keelung. The bureau of public health should investigate if they had received health services of other sources and propose specific strategies for them. The study is limited by secondary data due to the availability of variables and the type/content of variables. Data on health care utilization or primary survey can be adopted for future studies.