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.
出处
《中华消化杂志》
CAS
CSCD
北大核心
1995年第6期313-316,共4页
Chinese Journal of Digestion
关键词
胆囊结石
预测
LOGISTIC
回归模型
胆囊
Cholesterol gallstone disease
Gallbladder
Prediction
Bile acids
Regressionanalysis.