糖尿病是造成許多臨床疾病及死亡的因素，對於社會亦造成嚴重的經濟負擔。過去糖尿病的防制，多著重於疾病及併發症的診斷與治療，若能將防制工作提前到初級預防階段，早期發現「前期糖尿病（prediabetes）」個案，儘早給予健康促進之措施，則可以延緩糖尿病之發生，且在良好的血糖控制下，亦可預防血管併發症。因此，建立有效的篩檢工具以篩檢出「前期糖尿病」個案，是必要且刻不容緩的工作。 本研究以橫斷面研究設計，評估美國糖尿病學會於2004年所訂定的「糖尿病危險因子篩選工具」用於篩檢前期糖尿病病患之信效度，並推估前期糖尿病盛行率。以民國92年12月底設籍於台中市地區且年滿四十歲以上的民眾作為研究對象，分層隨機抽取4,278人，給予填寫結構性問卷，內容包括「糖尿病危險因子篩選工具」、基本人口學資料、個人健康行為與飲食習慣、個人疾病史及家族病史，以測試再測試方法評估篩選工具信度，以敏感度、特異度及操作特徵曲線(Receiver Operator Characteristic Curve, ROC曲線)下的面積評估「糖尿病危險因子篩選工具」用於篩檢「前期糖尿病」個案之效度。評估作為篩檢「前期糖尿病」工具之可行性，提供糖尿病防制篩檢的參考。 本研究結果顯示美國糖尿病學會2004年版之「糖尿病危險因子篩選工具」用以篩檢台中市40歲以上人口「前期糖尿病」個案時，ROC曲線下的面積在女性為64.9％；男性58.4％，以糖尿病危險篩選工具」與其他顯著危險因子繪製之ROC曲線結果顯示女性最佳切點為危險分數加總總分大於等於42分，此時敏感度為67.52％；特異度為64.4％；男性最佳切點為危險分數加總總分大於等於46分，此時敏感度為61.75％，特異度為54.05％。糖尿病篩選工具中各項危險因子再測試信度之Kappa值範圍介於0.46至0.94之間。研究結果顯示糖尿病危險因子篩選工具僅具有中等程度之信效度，故不建議將此工具實際施行使用，本研究推估台中市40歲以上人口前期糖尿病盛行率為27.29％（95％信賴區間24.61-29.97％）。 本研究所使用的糖尿病危險因子篩選工具，雖不適用於「前期糖尿病」之實際篩檢工作，卻可用於健康教育與宣導活動，來提高大眾對糖尿病的警覺性。國內缺乏糖尿病人群篩檢方法的相關調查研究，且本國糖尿病危險因子之定義也尚未完全釐清，所以如何發展一套適用於本國且有效的簡易篩檢糖尿病或前期糖尿病工具，值得未來研究繼續探討。; Diabetes has known no cure, and is a major cause of morbidity and mortality. In a addition to the impact of health, diabetes also plays an important role on the burden of national medical expenditure. During the past years, many efforts were put on the diagnosis and treatment of DM and its complications. It will be more effective if our efforts can aim at primary preventions such as the early detection and health promotion of the prediabetes. Many studies showed that it was possible to prevent or delay the occurrence of diabetes among prediabetes individuals by diet control or exercise, regardless of age, sex, and race. From public health viewpoint, it is worthwhile to identify Prediabetes people in community through screening program. The objectives of the current study are to evaluate the reliability and validity of “Diabetes risk factor screening tool” proposed by American Diabetes Association in 2004 and to estimate the prevalence of prediabetes among residents aged 40 and over in Taichung city. A total of 1,384 residents aged 40 and over in Taichung city were recruited. One thousand and seventy-three individuals completed a structured questionnaire, including sex, age, personal health behavior, disease history, family disease history and “Diabetes risk factor screening tool”. The receiver operator characteristic (ROC) curve was used to determine best cut point of the sensitivity and specificity of the “Diabetes risk factor screening tool”. The results show that the ROC curve’s area under curve (AUC) for females and males were 64.9% and 58.4% respectively. The sensitivity and specificity for the best cut-off point were 67.52% and 64.4% for females and were 61.75% and 54.5% for males. The kappa values of test-retest reliability for all risk factors ranged from 0.46 to 0.94. Our study results show that the sensitivity and specificity don’t reach the minimum requirement of standard (70%) for screening in community. It is not suggest using this screen tool to community screening prediabetes. The estimate of prediabetes prevalence in population over 40 years of age in Taichung City is 27.29 %.(95% confidence interval: 24.61-29.97%). Although the “Diabetes risk factor screening tool” is not good enough to screen prediabetes individuals in community. However it can be used for a healthy education to raise the public's awareness to prediabetes.
