社区非特异性下腰痛流行病学调查及预测模型的构建  

The epidemiology survey and construction of predictive model for NLBP in community

作　　者：王凯[1] 闵少雄[2] 徐新毅[2] 邱素均[2] 安胜利[1] 

机构地区：[1]南方医科大学公共卫生与热带医学学院生物统计学系,广州市510515 [2]南方医科大学附属珠江医院,广州市510282

出　　处：《实用医学杂志》2015年第2期306-308,共3页The Journal of Practical Medicine

基　　金：广东省科技计划项目(编号:2011B080701018)

摘　　要：目的：构建人群非特异性下腰痛（NLBP）预测模型。方法：采用便利抽样方法调查广州市某两个社区常住人口，将所有研究对象随机分为两部分，其中70％用于建模，30％用于考核模型。采用Logistic回归构建模型。结果：共回收1953份有效问卷。其中NLBP459例（23．50％，95％CI：21．12％～25．38％）；NLBP的影响因素分别为性别、教育程度、锻炼频率、劳动强度、振动接触、BMI、年龄、心理状况；模型预测准确率为81．44％（95％CI：78．12％～84．76％）。结论：所构建模型可用于临床上NLBP的辅助诊断及预防工作。Objective To construct the predictive model of NLBP in community. Methods Investigating a two community resident population. All subjects were randomly divided into two parts, a development （70%） cohort and a validation （30%） cohort. The development cohort was used to develop a predictive model. The validation cohort was used to prospectively evaluate the diagnostic yield of the model. Logistic regression analysis was used to construct the model. Results 2 300 questionnaires were issued, and 1 953 valid questionnaires were sifted,effective rateis 84.91%. There are 459 NLBP patients and the prevalence rate of NLBP was 23.50%（95%CI： 21.12% - 25.38% ） ; Affecting-factors of NLBP were gender, education, exercise, the labor intensity, vibration, BMI, age and psychological status; the accuracy rate of predictive model was 81.44%（95%C1：78.12% - 84.76%）. Conclusion The prediction model of nonspecific low back pain constructed by logistic regression analysis can be used for clinical auxiliary diagnosis of NLBP, and it is also suitable for preventing NLBPof community people.

关 键 词：非特异性下腰痛 危险因素 LOGISTIC回归 预测模型 

分 类 号：R681.53[医药卫生—骨科学]