目的:研究多因素判别分析评分在子宫内膜异位症诊断中的应用价值,提高其临床诊断率。方法:对2004年1月～2009年2月在该院行妇科手术的268例患者进行回顾性分析,用SAS软件包中的DISCRIM、FREQ过程处理临床数据。结果:子宫内膜异位症的Bayes判别评分及临床术前诊断的灵敏度分别为89.5%、78.9%;特异度为88.2%、82.4%;阳性预测值96.2%、93.0%;阴性预测值为71.5%、45.2%。Bayes判别诊断效果优于临床诊断,两者比较差异有统计学意义(χ2=4.05,P<0.05)。结论:多因素判别评分对提高子宫内膜异位症的术前诊断率有重要的应用价值。 Objective: To study the value of multifactorial discriminant analysis in the diagnosis of endometriosis and to improve its clinical diagnosis rate. Methods: A retrospective analysis was performed on 268 patients undergoing gynecological surgery from January 2004 to February 2009 in our hospital. The clinical data were processed with the DISCRIM and FREQ procedures in the SAS software package. Results: The Bayes discriminant score of endometriosis and the sensitivity of clinical preoperative diagnosis were 89.5%, 78.9% respectively; the specificity was 88.2%, 82.4%; the positive predictive value was 96.2%, 93.0%; the negative predictive value was 71.5 %, 45.2%. Bayes discriminant diagnosis was better than clinical diagnosis, the difference was statistically significant (χ2 = 4.05, P <0.05). Conclusion: Multivariate discriminant scores have an important value in improving the preoperative diagnosis rate of endometriosis.