耐多药肺结核的影响因素及风险预测

Analysis on the influence factors and risk prediction of multidrug-resistant tuberculosis

目的探索耐多药肺结核(MDR-TB)的发病危险因素,并构建风险预测模型。方法整群抽取2014年6月至2015年12月武汉市结核病防治所登记耐多药肺结核病例为病例组,未患有结核病的健康人群为对照组。自行设计调查表,面对面调查收集相关资料。用logistic回归法筛选危险因素,构建风险预测模型;采用H-L检验评价拟合优度,ROC曲线下面积来评估其预测效能。结果多因素logistic回归分析显示,性别为男性、婚姻状况(未在婚)、无固定职业、经常去公共场所为MDR-TB发生的独立危险因素(P<0.05)。H-L检验结果为:χ~2=6.508,P=0.482。ROC曲线下面积为0.810(95%CI:0.757~0.862)。结论性别为男性、婚姻状况(未在婚)、无固定职业、经常搭乘公共交通工具(去公共场所)与MDR-TB的发生相关。Logistic回归模型预测MDR-TB发生风险精度较高。

Objective To explore the risk factors of multidrug-resistant tuberculosis,and construct the risk prediction model.Methods Cluster sampling was used to select 125 cases with MDR pulmonary tuberculosis registered in Wuhan Institute for Tuberculosis Control as case group,and 125 healthy people were included as control group.Self-designed questionnaire and face to face investigation were conducted.Logistic regression was used to analyze the influence factors and the risk factor prediction model was constructed.The H-L test was used to evaluate goodness of fit,and the area under the ROC curve was to elucidate its predictive effectiveness.Results Logistic regression analysis revealed that male,unmarried or divorced,unemployed y and those always went to public places were the independent risk factors of MDR-TB.The result of H-L test was χ~2=6.508,P=0.482.The area under ROC curve was 0.810（95% CI：0.757-0.862）.Conclusions Male,unmarried or divorced,unemployed and those always went to public places were related to MDR-TB.The use of Logistic regression model to predict the risk of MDR-TB was of high precision.