用Logistic回归模型预测胆固醇结石易患人群  

PREDICTION OF CHOLESTEROL STONES USING LOGISTICREGRESSION MODEL:A PROSPECTIVE STUDY

作　　者：韩天权[1] 张圣道[1] 陈胜[1] 乙芳[1] 韩士荣[1] 陶志[1] 朱全梅[1] 

机构地区：[1]上海第二医科大学瑞金医院,上海第二医科大学卫生管理系,上海人民出版社医务科

出　　处：《中华消化杂志》1995年第6期313-316,共4页Chinese Journal of Digestion

摘　　要：采用Ｌｏｇｉｓｔｉｃ回归模型对胆固醇结石高危人群的筛选进行了初步的尝试。对胆囊的形态学和动力学以及血清生化各种指标进行优化组合。５７例胆固醇结石病人和１７２例健康对照者用作建立模型，对有高血压史但无胆石的６８例，用来作胆固醇结石前瞻性预测。根据模型判别得出，男组５２例中有成石危险的有３３例，女组１６例中有危险的计１０例。２４至３０个月后日超复查发现其中７例新发胆石，发病率为１０％，预测正确率达７１％（５／７）。结果显示，对胆固醇结石高危人群的预测是可行的。sing Logistic regression model,prediction of cholesterol stones was done.The regressionmodel was developed from 57 patients with cholesterol stone and l72 health volunteers ascontrols. The gallbladder morphology and kinetics were measured with ultrasonography, andbiochemical analysis at the same time.68 patients with hypertension but without gallstoneswere studied for their susceptibility to cholesterol gallstone. 33 out of 52 male and l0 out of16 female were predicted to have risk of stone calculated according to the rgeressionequations. The formation of gallstone was examined with ultrasound scanning 24 to 30months afterwards.7 cases had been found to have gallstone(incidence l0%)with accurateprediction rate 71%.The results suggest that prediction of population at the risk ofcholesterol gallstone is feasible.

关 键 词：胆囊结石 预测 LOGISTIC 回归模型 胆囊 

分 类 号：R657.420.2[医药卫生—外科学]