基于决策树模型建立宫颈机能不全患者术后不良妊娠结局的风险预警模型  

Risk warning model of postoperative adverse pregnancy outcome in patients with cervical incompetence based on decision tree algorithm

作　　者：弋晶晶 李兴亭 蒲春容 陈蕾 YI Jingjing;LI Xingting;PU Chunrong;CHEN Lei(Department of Obstetrics,Deyang People’s Hospital,Deyang,Sichuan 618000,China)

机构地区：[1]德阳市人民医院产科,四川德阳618000

出　　处：《重庆医学》2025年第3期668-672,677,共6页Chongqing Medical Journal

摘　　要：目的探讨宫颈机能不全(CI)患者术后不良妊娠结局的危险因素,并基于决策树模型建立CI患者术后不良妊娠结局的风险预警模型。方法回顾性分析2022年2月至2023年4月该院收治的159例CI患者的临床资料,筛选不良妊娠结局的危险因素并构建术后不良妊娠结局的决策树模型,内部验证方法为5折交叉验证。结果不良妊娠结局发生率为22.64%。宫颈环扎时间、羊膜囊膨出、多次环扎、术前宫颈长度、羊水沉积物均为不良妊娠结局发生的影响因素(P<0.05)。羊水沉积物是CI患者术后发生不良妊娠结局最为重要的影响因素,术前宫颈长度对CI患者术后发生不良妊娠结局的影响较小。logistic回归模型的曲线下面积(AUC)值稍高于决策树模型。5折交叉验证的模型正确率为78.3%。结论临床治疗期间可结合上述两种模型从不同层面发现CI患者术后不良妊娠结局的影响因素,为临床医务人员对CI患者的病情评估及干预方案制订提供参考。Objective To explore the risk factors of adverse pregnancy outcomes in the patients with cervical incompetence(CI),and to establish a risk warning model of adverse pregnancy outcomes in CI patients based on decision tree model.Methods The clinical data of 159 patients with CI admitted and treated in this hospital from February 2022 to April 2023 were retrospectively analyzed,and the risk factors for adverse pregnancy outcomes were screened and the decision tree model for postoperative adverse pregnancy outcomes was constructed.The internal verification method was 5-fold cross-validation.Results The incidence rate of adverse pregnancy outcome was 22.64%.The pregnant weeks of cervical cerclage,amniotic cystocele,multiple cervical cerclage,preoperative cervical length and amniotic fluid sediment were all influential factors for adverse pregnancy outcome occurrence(P<0.05).The amniotic fluid sediment was the most important factor affecting the postoperative adverse pregnancy outcome in CI patients,and the preoperative cervical length had little influence on the postoperative adverse pregnancy outcomes in CI patients.The area under the curve(AUC)value of logistic regression model was slightly higher than that of the decision tree model.The accuracy rate of the 5-fold cross-validation model was 78.3%.Conclusion During clinical treatment,the above two models can be combined to find the influencing factors of postoperative adverse pregnancy outcomes in CI patients from different aspects,and provide references for clinical medical staff to evaluate the disease condition of CI patients and formulate the intervention plans.

关 键 词：宫颈机能不全 宫颈环扎手术 不良妊娠结局 影响因素 

分 类 号：R714.2[医药卫生—妇产科学]