【摘要】 目的　建立解释和预测抗TB化疗结局的Logistic回归模型。　方法　采用病例对照研究方法及成组资料的非条件Logistic回归分析模型P =1/ {1+exp[-(a +∑bjXj) ] },对 2 77例涂阳TB患者的近期化疗结局与 3 9个有关因素之间的关系 ,利用向前逐步选择法进行分析 ,按α =0 0 5的水平 ,拟合回归预测模型 ,并以回代检验方法考察模型的预测效能。　结果　最终似合的方程为P =1/ {1+exp[-( 3 4773 -0 14 99X11-0 7983X2 3 -0 15 0 1X3 7) ] },回代检验的符合率为 77 3 7%。　结论　利用成组资料的非条件Logistic回归分析建立抗TB化疗结局的解释和预测模型是可行的更多还原

【Abstract】 Objective To construct Logistic Regression Model for analyzing and forecasting short-term result in chemotherapy of patients Methods A case-control stuty and Non-conditional Logistic Regression Analysis Model P=1/{1+exp[-(a+∑b jX j)]}have been used for analysis of 277 pulmonary tuberculosis patients treated with DOTS, 39 relevant influential factors were analyzed(α=0 05) Logistic Regression Model has been checked by returning original material to regression equation Results Through analysis of Non-conditional Logistic Regression,3 factors are eventually induced into the fitted regression equation,i e ,P=1/{1+exp[-(3 4773-0 1499X 11 -0 7983X 23 -0 1501X 37 )]},correct rate of which accounts for about 77 3% Conclusion Non-conditional Logistic Regression Model is feasible in analyzing and forecasting short-term result of chemotherapy of TB patients更多还原