初治肺结核治愈后复发危险因素cox比例风险模型研究探究资料.docVIP

PAGE PAGE 8 初治肺结核治愈后复发危险因素COX比例风险模型研究 广东省东莞市慢性病防治院（523008） 邓　斌　何庆伟　余锦香　周志刚　罗勇强　易来龙　张锡萍 邓思东 摘要 目的 探讨初治肺结核病人经首次抗痨治疗临床治愈后导致复发的危险因素，并同时建立复发事件的预测模型。方法 本研究采用COX比例风险模型对经首次抗痨治疗临床治愈后的1268例初治肺结核病例进行为期3—6年的历史前瞻性队列随访分析，研究导致病情复发的危险因素及临床意义。结果 对欲引入的14个研究因素进行COX比例风险模型分析显示共有9个研究因素（协变量）是初治肺结核病人首次抗痨治疗临床治愈后的复发危险因素，它们是年龄（X2）、吸烟（X4）、营养状况（X5）、规则治疗（X9）、排菌期（X10）、合并糖尿病（X11）、停药时残留空洞（X12），免疫治疗（X13）及耐药情况（X14），所建预后指数方程为W=3.7101+1.1416X2+1.3461X4-1.1436X5-2.1132X9+0.5136X10+ 0.3101X11+0.3821X12-1.9718X13+0.6921X14。解释变异度量V2值为49.1%，显示所建模型预测能力良好。结论 COX比例风险模型在初治肺结核首次抗痨治疗临床治愈后复发危险因素研究中是一种有效的工具，用所建COX模型对初治肺结核病人首次抗痨治疗的复发事件进行预测效果良好，COX模型值得进一步研究推广。 关键词 结核，肺；复发危险因素；COX比例风险模型；历史前瞻性研究 Study of Prognostic Replase Risk Factors after First anti-tuberculosis on Pulmonary Tuberculosis by COX Model. Abstract Object to study the prognostic replase risk factors after first anti-tuberculosis on pulmonary tuberculosis and to establish the forecast model for the replase event at the same time.Methods 1268 patients with pulmonary tuberculosis after first anti-tuberculosis were followed by historial prospective study for 3-6 years,and the prognostic relapse risk factors were studied by COX proportional hazards model.Result our study showns that there are 9 independent variables(covariables) which could effect the replase rate of patients with pulmonary tuberculosis after first anti-tuberculosis.they are age(X2) 、smoking(X4) 、nutrient condition(X5) 、regular cure(X9) 、discharge period of bacteria(X10),combine with diabetes(X11) 、residual fibrocavitary(X12) 、immune cure(X13) 、resistance drug(X14) ,our forecast model for the replase event is W=3.7101+1.1416X2+1.3461X4 -1.1436X5-2.1132X9+0.5136X10+0.3101X11+0.3821X12-1.9718X13+0.6921X14. explained variation V2 is 49.1%,it showns that the forecase ability of the COX model is best for the replase event.Conclusion this study suggests that the COX proportional hazards model is more effect in fitting the survival data and is much more sensible in analysis of prognostic r