胆总管结石相关因素及指标的Logistic回归判别分析

目的:探讨胆总管结石的预测方法。方法:采用回顾性调查的方式,对胆总管结石的16种相关因素进行Logistic回归分析,建立相应的判别模型。结果:Logistic回归将性别、黄疸史、胰腺炎史、B超或CT所示的胆总管内径的最宽处、谷草转氨酶、谷丙转氨酶、γ-谷胺酰转移酶、胆囊结石史或胆囊切除史纳入判别模型后判别效果最好;回代检验样本结果显示的判别模型的灵敏度为89.4%,特异性为80.0%。结论:利用临床资料进行Logistic回归判别分析,可以提高胆总管结石预测的准确性,有助于指导临床。

Logistic regression discriminant analysis of relative factors for choledocholithiasis

Objective To investigate a detective methods of choledocholithiasis.Methods Using retrospective study methods, a logistic regression descriminant analysis of 16 factors related to choledocholithiasis was made and a relative discriminant model was constructed.Results Logistic regression analyses that had sex, history of jaundice or jaundice, the widest inner diameter of choledochus, AST,ALT and the history of cholecystectomy, cholecystolithiasis and pancreatitis included into discriminant model, gave the best predictive result. A test sample showed the discriminant model had a sensitivity of 89.4%, and a specificity of 80.0%.Conclusions Discriminant analysis of logistic regression with clinical data is helpful for diagnosis and treatment of choledocholithiasis.It can also increase the accuracy of the predicion of cholecystolithiasis and serve as a clinical guide.