论文部分内容阅读
目的构建喷泉水嗜肺军团菌污染贝叶斯网络预警模型,为预防和控制喷泉水军团菌污染提供科学依据。方法选择2015年深圳市公共场所正运行的70座喷泉作为研究对象,采用问卷调查、现场监测及实验室检测等方法收集相关数据,构建贝叶斯网络预警模型。选择2016年深圳市公共场所正运行的30座喷泉作为研究对象,收集相关数据对模型预警结果进行验证。结果深圳市喷泉水嗜肺军团菌检出率为42.00%(42/100)。从构建的贝叶斯网络模型结构看,对喷泉水中嗜肺军团菌最具有影响的因素有定期清洗、浊度、游离性余氯和溶解性总固体。该模型ROC曲线最佳诊断临界点为0.475,利用该临界点进行嗜肺军团菌阳性诊断ROC曲线下面积为0.941(95%CI:0.893~0.998),诊断灵敏度为92.90%,特异度为90.90%,预测准确率为93.33%(28/30)。结论本研究构建的贝叶斯网络模型预测准确率较高,可以满足喷泉水嗜肺军团菌污染预警要求,对公共场所喷泉水嗜肺军团菌污染的判定具有一定的参考价值。
Objective To construct Bayesian network warning model of Legionella pneumophila infection in fountain water and provide a scientific basis for the prevention and control of Legionella pneumophila contamination. Methods Seventy fountains in public places in Shenzhen were selected as research objects in 2015, and relevant data were collected by questionnaires, on-site monitoring and laboratory tests to build Bayesian network early-warning model. Select 30 fountains which are running in public places in Shenzhen in 2016 as the research object and collect relevant data to verify the model early warning results. Results The detection rate of Legionella pneumophila in water fountain in Shenzhen was 42.00% (42/100). From the construction of the Bayesian network model structure, the most influential factors of Legionella pneumophila in fountain water are regular cleaning, turbidity, free residual chlorine and total dissolved solids. The ROC curve of this model had the best diagnostic threshold of 0.475. The area under the ROC curve of positive detection of Legionella pneumophila using this critical point was 0.941 (95% CI: 0.893-0.998). The diagnostic sensitivity was 92.90% and the specificity was 90.90% , The prediction accuracy rate was 93.33% (28/30). Conclusion The Bayesian network model constructed in this study has a high prediction accuracy, which can meet the requirements of early warning of Legionella pneumophila infection in water fountains and has certain reference value for judging the Legionella pneumophila contamination in fountains in public places.