摘要
目的:调查2019-2021年徐州市肺结核流行病学特征,并构建影响因素的Logistic回归预测模型.方法:选取2019年1月1日至2021年12月30日期间徐州市登记的788例肺结核病患者作为研究对象.观察徐州市2019-2021年肺结核流行病学特征,并采用单因素和多因素Logistic回归分析分析肺结核发病的影响因素.构建Log P回归预测模型,分析其预测效能.结果:流行病学统计结果显示:徐州市2019-2021年肺结核患者以男性、年龄≥ 60岁、居住地为农村、婚姻情况为已婚的患者居多,其中病程、文化程度的分类群体患病人数较为接近.单因素分析结果显示,肺结核发病与肺结核接触史、饮酒史、结核病健康教育、性别、家庭人均月收入、年龄、职业、吸烟史、卡介苗接种史、居住地有关(P＜0.05).多因素Logistic回归分析结果显示:居住地为农村、吸烟史、性别为男、饮酒史、家庭人均月收入＜3000元、年龄≥60岁、职业为农民或工人、肺结核接触史是肺结核发病的危险因素,而有卡介苗接种史则是其保护因素(P＜0.05).对应的LogP模型,经ROC分析可知:该预测模型有较好的对肺结核发病的预测效能,其ROC-AUC(0.95CI)为0.881(0.817～0.928),准确度可达86.5％.结论:徐州市应重点关注男性、老年、居住地为农村、有吸烟史、有饮酒史、家庭人均月收入低、职业为农民或工人、肺结核接触史的肺结核发病群体,且应积极开展接种卡介苗工作.此外,通过影响因素构建的预测模型对肺结核发病有较好的预测效能.
Abstract
Objective:To investigate the epidemiological characteristics of pulmonary tuberculosis in Xuzhou from 2019 to 2021,and construct a logistic regression prediction model for influencing factors.Methods:788 cases of pulmonary tuberculosis registered in Xuzhou from January 1,2019 to December 30,2021 were selected as the study subjects.Observed the epidemiological characteristics of pulmonary tuberculosis in Xuzhou from 2019 to 2021,and used univariate and multivariate logistic regression analysis to analyze the influencing factors of pulmonary tuberculosis incidence.Constructed a Log P regression prediction model and analyzed its predictive performance.Result:Epidemiological statistics showed that from 2019 to 2021,the majority of pulmonary tuberculosis patients in Xuzhou were male,aged ≥ 60 years old,residing in rural areas,and married,the number of patients in the disease course and education level classification group was relatively close.Univariate analysis showed that the incidence of pulmonary tuberculosis was related to contact history of pulmonary tuberculosis,drinking history,tuberculosis health education,gender,average monthly income of per person in family,age,occupation,smoking history,BCG vaccination history,and residence(P＜0.05).The results of multivariate logistic regression analysis showed that residing in rural areas,smoking history,gender as male,drinking history,average monthly income of per person in family＜3000 yuan,age≥60 years old,occupation as a farmer or worker,and contact history of pulmonary tuberculosis were risk factors for pulmonary tuberculosis,while BCG vaccination history was a protective factor(P＜0.05).The corresponding Log P model,after ROC analysis,it showed that this prediction model had good predictive performance for the incidence of pulmonary tuberculosis,with an ROC-AUC(0.95C1)of 0.881(0.817-0.928)and an accuracy of 86.5％.Conclusion:Xuzhou should focus on population affected by pulmonary tuberculosis of male,elderly,residing in rural areas,smoking history,drinking history,low average monthly income of per person in family,farmers or workers,and contact history of pulmonary tuberculosis,and should actively carry out the vaccination of BCG vaccine.In addition,the prediction model constructed through influencing factors has good predictive performance on the incidence of pulmonary tuberculosis.
NETL
NSTL
万方数据