摘要
目的通过建立决策树分类数据模型寻找影响行宫腔内人工授精(IUI)患者临床妊娠率的相关指标。方法回顾性分析2018年1月至2020年12月于邢台市第三医院生殖医学科行IUI的210例患者的399个治疗周期的临床资料。对缺失值检测并补充完整后,按照临床妊娠结局分为妊娠组和非妊娠组。比较两组定量项目结果差异;选择决策树模型,获得预测变量的重要性和决策树判断方程,再验证模型的诊断效能。结果妊娠组处理后精子密度、处理前精子活率、回收率、处理后前向运动精子总数和女方体重指数结果高于非妊娠组,差异具有统计学意义(P<0.05)。建立CHAID决策树模型的总精确度为91.228%,曲线下面积为0.957。主要影响指标预测变量的重要性由强到弱分别为回收率(0.417)、扳机日卵泡数(0.213)和首次精子处理操作者(0.153)。获得的判断方程为当扳机日卵泡数>1且首次精子处理操作者为操作者1或操作者3时,临床妊娠成功率为38.636%;当扳机日卵泡数≤1且回收率介于40.82%~50.40%,临床妊娠成功率为26.316%。结论回收率、扳机日卵泡数和首次精子处理操作者是影响临床妊娠成功与否的主要因素,决策树判断方程式在提高临床妊娠率方面具有一定作用。
Objective To explore the relevant indicators to affect the clinical pregnancy rate of intrauterine insemination(IUI)patients by establishing a decision tree analysis model.Methods The clinical data of 210 patients with 399 treatment cycles who underwent IUI in Department of Reproductive Medicine,the Third Hospital of Xingtai City from January 2018 to December 2020 were retrospectively analyzed.After missing values were detected and supplemented,the patients were divided into pregnant group and non-pregnant group according to clinical pregnancy outcome.The results of quantitative items were compared between the two groups.The decision tree analysis model was selected to obtain the importance of predictor variables and the decision tree judgment equation,and then the diagnostic efficiency of the model was verified.Results The sperm density after treatment,sperm motility rate before treatment,recovery rate,total number of forward motile sperm after treatment and female body mass index in the pregnant group were higher than those in the non-pregnant group,and the differences were statistically significant(P<0.05).The total accuracy of the CHAID decisiontree analysis model was 91.228%,and the area under the curve was 0.957.The importance of the main predictor variables was recovery rate(0.417),number of follicles on trigger day(0.213)and operator of the first sperm treatment(0.153).When the number of follicles on trigger day was>1 and the operator of the first sperm treatment was operator 1 or operator 3,the success rate of clinical pregnancy was 38.636%.When the number of follicles on trigger day was less than 1 and the recovery rate was between 40.82%and 50.40%,the success rate of clinical pregnancy was 26.316%.Conclusions The recovery rate,the number of follicles on trigger day and the operator of the first sperm treatment are the main factors affecting the success of clinical pregnancy.The decision tree judgment equation plays a certain role in improving the clinical pregnancy rate.
作者
尹志辉
董月稳
孙巧敏
李致文
刘建晓
贾亚男
董欣
洪燕
高佳丽
修晨
马国燕
YIN Zhihui;DONG Yuewen;SUN Qiaomin;LI Zhiwen;LIU Jianxiao;JIA Yanan;DONG Xin;HONG Yan;GAO Jiali;XIU Chen;MA Guoyan(Department of Clinical Laboratory,the Third Hospital of Xingtai City,Xingtai 054000,Hebei,China;Department of Reproductive Medicine,the Third Hospital of Xingtai City,Xingtai 054000,Hebei,China;Department of Obstetrics and Gynecology,the Third Hospital of Xingtai City,Xingtai 054000,Hebei,China)
出处
《中国性科学》
2022年第8期81-85,共5页
Chinese Journal of Human Sexuality
基金
河北省邢台市科技计划项目(2019ZC323)。
