社区糖尿病不同筛查方法效果及可操作性比较

黎衍云; 王思源; 杨沁平; 严青华; 吴菲; 程旻娜; 施燕

doi:10.11847/zgggws1135267

社区糖尿病不同筛查方法效果及可操作性比较

Effectiveness and operability of different screening methods for diabetes in community populations: a comparative analysis

摘要

摘要:
目的评价不同社区糖尿病筛查方法的筛查效果和实用意义，为社区糖尿病筛查流程的确立提供依据。
方法于2017年10 — 12月采用多阶段整群随机抽样方法在上海市抽取23 940名 ≥ 35岁中老年社区居民进行问卷调查、体格检查和血糖测定，通过比较早发现流程、风险评分法、空腹静脉血糖（FPG）等筛查方法的灵敏度、特异度、约登指数、受试者工作特征曲线下面积（AUROC）以及筛查过程中需要进行FPG和口服葡萄糖耐量试验（OGTT）检测比例评价不同筛查方法的筛查效果及可操作性。
结果最终纳入分析的17981名既往未确诊糖尿病者中，新诊断糖尿病前期者4129例，糖尿病前期患病率为22.96 %；新诊断出糖尿病者2213例，新诊断糖尿病患病率为12.31 %。早发现流程筛查未诊断糖尿病人群和糖尿病前期人群的灵敏度、特异度、约登指数、AUROC分别为87.3 %、60.2 %、0.479、0.858和61.0 %、58.9 %、0.199、0.620，风险评分法分别为96.7 %、15.2 %、0.119、0.680和93.3 %、15.8 %、0.091、0.602，FPG法分别为82.5 %、87.5 %、0.700、0.922和 47.8 %、86.0 %、0.338、0.708。早发现流程FPG和OGTT 2h静脉血糖（2h PG）需检测人数比例为45.7 %和42.6 %，风险评分法均为86.3 %，FPG法分别为100.0 %和21.1 %。
结论早发现流程可在获得较高糖代谢异常人群检出率的同时，最大程度减少静脉血糖检测人数，简化现场实施流程，可作为大规模社区人群糖尿病及糖尿病前期筛查的有效工具。

Abstract:
Objective To evaluate the efficiency and practical significance of different diabetes screening methods in community population and to provide evidences for the establishment of community diabetes screening process.
Methods Using multistage cluster random sampling, we recruited 23 940 permanent residents aged ≥ 35 years and without acute or serious chronic diseases/type I diabetes in urban communities and rural villages of 7 districts of Shanghai for a questionnaire survey, physical examination and blood glucose measurement during October – December, 2017. The effectiveness and operability of an early-detecting process, risk score assessment and fast plasma glucose (FPG) screening for type 2 diabetes (T2DM) were evaluated by comparing the three methods′ sensitivity, specificity, Youden index , area under the receiver operating characteristic curve (AUROC), the proportion of abnormal FPG and oral glucose tolerance test (OGTT).
Results Of the 17 981 previously undiagnosed diabetics identified among the residents and included in the analysis, 4 129 prediabetics and 2 213 diabetes were detected, with the prevalence rates of 22.96% and 12.31%, respectively. For the three methods (early-detecting process vs. risk score assessment vs. FPG screening) in detections of diabetes/prediabetic in previously undiagnosed diabetic population, the sensitivities (%) were 87.3/61.0 vs. 96.7/93.3 vs. 82.5/47.8; the specificities (%) were 60.2/58.9 vs. 15.2/15.8 vs. 87.5/86.0; the Youden indexes were 0.479/0.199 vs. 0.119/0.091 vs. 0.700/0.338, and the values of AUROC were 0.858/0.620 vs. 0.680/0.602 vs. 0.922/0.708; also for the three methods, the proportions (%) of the examinees requiring FPG/2-hour plasma glucose detection of OGTT were 45.7/42.6 vs. 86.3/86.3 vs. 100/21.1, respectively.
Conclusion In diabetes screening among community residents, early-detecting process could achieve a higher detection rate of abnormal glucose metabolism with a minimum number of venous blood glucose detection and simplified on-site implementation procedures and the process could be adopted effectively in large-scale prediabetics and diabetes screening in community populations.

HTML全文

参考文献(19)

施引文献

资源附件(0)