| 许沁,杨晓敏.基于Logistic回归与决策树模型的妊娠期高血压疾病患者产后发生高血压的影响因素分析[J].上海护理,2026,26(7): |
| 基于Logistic回归与决策树模型的妊娠期高血压疾病患者产后发生高血压的影响因素分析 |
| Analysis of influencing factors for postpartum hypertension in patients with gestational hypertension based on logistic regression and decision tree model |
| 投稿时间：2025-04-23 修订日期：2026-07-15 |
| DOI： |
| 中文关键词: Logistic回归 决策树 妊娠期高血压 影响因素 高血压家族史 |
| 英文关键词:Logistic regression Decision tree Pregnancy induced hypertension Influencing factors Family history of hypertension |
| 基金项目: |
|
| 摘要点击次数: 88 |
| 全文下载次数: 0 |
| 中文摘要: |
| 【】 目的 探讨影响妊娠期高血压疾病患者产后发展为持续性高血压的关键因素,通过应用Logistic回归与决策树模型识别出可靠的预测指标,为妊娠期高血压疾病的预防和管理提供科学依据和有效策略。方法 回顾性分析选取2021年12月~2023年12月期间在我院生产的731例妊娠期高血压患者的临床资料,产后42天到院复诊检测血压,在此期间参与者未使用任何降血压药物,在对每位参与者进行至少三次血压测量,将三次测量收缩压≥140mmHg,舒张压≥90mmHg 的患者纳入高血压组,其余患者纳入血压正常组。利用SPSS 23.0进行数据分析,χ2检验比较不同特征妊娠期高血压疾病患者的差异,Logistic回归和决策树模型进行分析妊娠期高血压疾病患者产后发生高血压的影响因素,并采用受试者操作特征(receiver operating characteristic,ROC)曲线的曲线下面积(area under the curve,AUC)比较Logistic 回归模型与决策树模型的预测效果。结果 731例妊娠期高血压患者中,产后发生高血压的患者有95例,未发生高血压的患者有636例,发生率为13.00%。两组在婚姻状况、职业、文化程度、被动/主动吸烟情况、饮酒史、产次、孕次、妊娠胎数、妊娠期糖尿病史等资料比较无统计学意义(P＞0.05),在年龄、体质指数 (Body Mass Index,BMI)、食盐摄入量、妊娠期高血压疾病严重程度、高血压家族史等资料差异有统计学意义(P＜0.05)。多因素Logistic回归分析显示,年龄≥35岁、BMI ≥28kg/m2、食盐摄入量偏多、较为严重的妊娠期高血压疾病、有高血压家族史均为妊娠期高血压疾病患者产后发生高血压的影响因素(P<0.05)。决策树模型分析显示,年龄、妊娠期高血压疾病严重程度、高血压家族史是妊娠期高血压疾病患者产后发生高血压影响因素。ROC比较显示,Logistic 回归模型与决策树模型的AUC差异无统计学意义(0.692与0.636,Z=-1.280,P＞0.05)。结论 基于Logistic 回归模型与决策树模型在识别妊娠期高血压疾病患者产后发生高血压影响因素方面具有较好的一致性。针对这些风险因素,建议制定个性化的干预措施,如生活方式的调整和定期监测,以降低产后发生高血压的风险。 |
| 英文摘要: |
| 【】 Objective Exploring the key factors affecting the postpartum development of patients with gestational hypertension to persistent hypertension, and identifying reliable predictive indicators through the application of logistic regression and decision tree models, providing scientific basis and effective strategies for the prevention and management of gestational hypertension.Methods A retrospective analysis was conducted on the clinical data of 731 patients with pregnancy-induced hypertension our hospital from December 2021 to December 2023. Blood pressure was measured during a follow-up visit 42 days postpartum. None of the participants used any antihypertensive medication during this period.? Patients with three measurements of systolic blood pressure ≥140mmHg and diastolic blood pressure ≥90mmHg were classified into the hypertension group, while the rest were included in the normal blood pressure group. Data were analyzed using SPSS 23.0. The chi-square (χ2) test was employed to compare differences among patients with pregnancy-induced hypertension of different characteristics. Logistic regression and decision tree models were used to analyze the influencing factors for the development of postpartum hypertension in patients with pregnancy-induced hypertension. The predictive performances of the logistic regression model and the decision tree model were compared using the area under the receiver operating characteristic (ROC) curve (AUC). Results Among 731 patients with gestational hypertension, 95 had postpartum hypertension and 636 had no hypertension, with an incidence rate of 13.00%. There was no statistical significance between the two groups in terms of marital status, occupation, education level, passive/active smoking, drinking history, birth times, pregnancy times, number of pregnancies, and history of diabetes in pregnancy (P>0.05), but there was statistical significance in terms of age, body mass index (BMI), salt intake, severity of hypertensive disorder in pregnancy, and family history of hypertension (P<0.05). Multivariate logistic regression analysis showed that age ≥ 35 years old, BMI ≥ 28kg/m2, excessive salt intake, more severe gestational hypertension, and a family history of hypertension were all influencing factors for postpartum hypertension in patients with gestational hypertension (P<0.05). Decision tree model analysis shows that age, severity of gestational hypertension, and family history of hypertension are influencing factors for postpartum hypertension in patients with gestational hypertension. The ROC comparison showed that there was no statistically significant difference in AUC between the logistic regression model and the decision tree model (0.692 and 0.636, Z=-1.280, P>0.05). Conclusion The logistic regression model and decision tree model have good consistency in identifying the influencing factors of postpartum hypertension in patients with gestational hypertension. For these risk factors, it is recommended to develop personalized intervention measures, such as lifestyle adjustments and regular monitoring, to reduce the risk of postpartum hypertension. |
| 查看全文 查看/发表评论 下载PDF阅读器 |
| 关闭 |
|
|
|