基于机器学习产后压力性尿失禁风险预测模型的构建及验证

doi:10.6040/j.issn.1671-7554.0.2024.1431

摘要/Abstract

摘要： 目的基于最新的产后压力性尿失禁(postpartum stress urinary incontinence, PPSUI)定义,采用5种机器学习算法构建风险预测模型,筛选最优模型并评估其临床应用价值。方法采用前瞻性研究设计,纳入1 208 例产妇,基于问卷与电子病历收集数据,并按8∶2比例随机划分训练集和测试集。采用单因素分析和随机森林算法筛选特征,构建基于Logistic回归、决策树、随机森林、支持向量机和极限梯度提升(extreme gradient boosting, XGBoost)的PPSUI 预测模型,并通过网格搜索优化超参数。模型训练与验证采用Bootstrap方法和十折交叉验证,以提升稳定性和泛化能力。最终,从分类性能、临床适用性及预测可靠性等方面综合评估各模型,并筛选最优预测模型。结果 XGBoost预测模型为最优预测模型,测试集AUC值为0.993(95%CI: 0.985~0.998, P<0.01)。决策曲线分析显示,该模型在中等阈值范围内净收益最高,校准曲线接近理想状态,预测可靠性较优,具备较高的临床应用价值。结论 XGBoost适用于PPSUI高危人群的早期筛查和风险评估,为精准医学和产后健康管理提供科学依据。

关键词: 压力性尿失禁, 产后, 风险预测, 机器学习, 预测模型

Abstract: Objective To develop risk prediction models for postpartum stress urinary incontinence(PPSUI)using five machine learning algorithms based on the latest PPSUI definition, identify the optimal model, and evaluate its clinical applicability. Methods This study adopted a prospective design and included 1,208 postpartum women. Data were collected from questionnaires and electronic medical records, and the dataset was randomly divided into a training set and a test set in an 8∶2 ratio. Feature selection was performed using univariate analysis and the random forest algorithm. Five PPSUI prediction models were developed based on Logistic regression, decision tree, random forest, support vector machine, and extreme gradient boosting(XGBoost), with hyperparameter optimization conducted through grid search. Model training and validation were performed using the bootstrap method and ten-fold cross-validation to enhance stability and generalizability. The models were comprehensively evaluated in terms of classification performance, clinical applicability, and predictive reliability to identify the optimal prediction model. Results XGBoost was identified as the optimal prediction model, achieving the highest AUC in the test set(AUC=0.993, 95%CI: 0.985-0.998, P<0.01). Decision curve analysis showed that XGBoost provided the highest net benefit within the intermediate threshold range, while the calibration curve was closest to the ideal state, indicating superior predictive reliability and significant clinical applicability. Conclusion XGBoost is suitable for early screening and risk assessment of high-risk PPSUI populations, providing a scientific basis for precision medicine and postpartum health management.

Key words: Stress urinary incontinence, Postpartum, Risk prediction, Machine learning, Predictive model

中图分类号:

R714.46

王丽云,高天勤,刘雨佳,陈青,陈柳,沙凯辉. 基于机器学习产后压力性尿失禁风险预测模型的构建及验证[J]. 山东大学学报 (医学版), 2025, 63(6): 55-66.

WANG Liyun, GAO Tianqin, LIU Yujia, CHEN Qing, CHEN Liu, SHA Kaihui. Development and validation of a postpartum stress urinary incontinence risk prediction model based on machine learning[J]. Journal of Shandong University (Health Sciences), 2025, 63(6): 55-66.

参考文献

[1] Doumouchtsis SK, de Tayrac R, Lee J, et al. An International Continence Society(ICS)/International Urogynecological Association(IUGA)joint report on the terminology for the assessment and management of obstetric pelvic floor disorders[J]. Int Urogynecol J, 2023, 34(1): 1-42.
[2] You C, Zhao YJ, Zhang C, et al. Pelvic floor parameters predict postpartum stress urinary incontinence: a prospective MRI study[J]. Insights Imaging, 2023, 14(1): 160. doi:10.1186/s13244-023-01488-5
[3] Coll CM, Franco EM, Qureal LA, et al. Hormonal influence in stress urinary incontinence during pregnancy and postpartum[J]. Reprod Sci, 2022, 29(8): 2190-2199.
[4] Dai SD, Chen HT, Luo TZ. Prevalence and factors of urinary incontinence among postpartum: systematic review and meta-analysis[J]. BMC Pregnancy Childbirth, 2023, 23(1): 761. doi:10.1186/s12884-023-06059-6
[5] Wang K, Xu XL, Jia GM, et al. Risk factors for postpartum stress urinary incontinence: a systematic review and meta-analysis[J]. Reprod Sci, 2020, 27(12): 2129-2145.
[6] Rashidi Fakari F, Hajian S, Darvish S, et al. Predictors of help-seeking behaviors in women with urinary incontinence: Based on Iranian womens lens[J]. PLoS One, 2023, 18(8): e0289785. doi:10.1371/journal.pone.0289785
[7] Rashidi Fakari F, Hajian S, Darvish S, et al. Explaining factors affecting help-seeking behaviors in women with urinary incontinence: a qualitative study[J]. BMC Health Serv Res, 2021, 21(1): 60. doi:10.1186/s12913-020-06047-y
[8] 李巧巧, 于霞, 罗映娟, 等. 女性压力性尿失禁疾病负担研究进展[J]. 中国计划生育和妇产科, 2022, 14(9): 10-12.
[9] Ngiam KY, Khor IW. Big data and machine learning algorithms for health-care delivery[J]. Lancet Oncol, 2019, 20(5): e262-e273.
[10] Zhang A, Xing L, Zou J, et al. Shifting machine learning for healthcare from development to deployment and from models to data[J]. Nat Biomed Eng, 2022, 6(12): 1330-1345.
[11] Miku s M, Cori c M, Matak L, et al. Validation of the UDI-6 and the ICIQ-UI SF-Croatian version[J]. Int Urogynecology J, 2020, 31(12): 2625-2630.
[12] Huang L, Zhang SW, Wu SL, et al. The Chinese version of ICIQ: a useful tool in clinical practice and research on urinary incontinence[J]. Neurourol Urodyn, 2008, 27(6): 522-524.
[13] Legendre B, Cerasuolo D, Dejardin O, et al. How to deal with missing data? Multiple imputation by chained equations: recommendations and explanations for clinical practice[J]. Nephrol Ther, 2023, 19(3): 171-179.
[14] Speiser JL. A random forest method with feature selection for developing medical prediction models with clustered and longitudinal data[J]. J Biomed Inform, 2021, 117: 103763. doi:10.1016/j.jbi.2021.103763
[15] AlQuaiz AM, Kazi A, AlYousefi N, et al. Urinary incontinence affects the quality of life and increases psychological distress and low self-esteem[J]. Healthcare, 2023, 11(12): 1772. doi:10.3390/healthcare11121772
[16] Suar G, Cevik F, Simal Yavuz N, et al. Urinary incontinence in the postpartum 1-year period: its prevalence and effect on psychosocial status of women[J]. Low Urin Tract Symptoms, 2023, 15(5): 191-199.
[17] Chisholm LP, Sebesta EM, Gleicher S, et al. The burdens of incontinence: quantifying incontinence product usage and costs in women[J]. Neurourol Urodyn, 2022, 41(7): 1601-1611.
[18] Benítez-Andrades JA, García-Ordás MT, álvarez-González M, et al. Detection of the most influential variables for preventing postpartum urinary incontinence using machine learning techniques[J]. Digit Health, 2022, 8: 20552076221111289. doi:10.1177/20552076221111289
[19] 史可为, 吴亚飞, 方亚. 临床预测模型的可解释性及应用进展[J]. 现代预防医学, 2023, 50(6): 1122-1127. SHI Kewei, WU Yafei, FANG Ya. Interpretability and application of clinical prediction models[J]. 2023, 50(6): 1122-1127.
[20] Xu CC, Guo Y, Chi XL, et al. Establishment and validation of a simple nomogram for predicting early postpartum stress urinary incontinence among women with vaginal delivery: a retrospective study[J]. BMC Womens Health, 2023, 23(1): 8. doi:10.1186/s12905-023-02160-2
[21] Liu W, Qian LX. Establishment and validation of a risk prediction model for postpartum stress urinary incontinence based on pelvic floor ultrasound and clinical data[J]. Int Urogynecol J, 2022, 33(12): 3491-3497.
[22] Eric Jelovsek J, Piccorelli A, Barber MD, et al. Prediction models for postpartum urinary and fecal incontinence in primiparous women[J]. Female Pelvic Med Reconstr Surg, 2013, 19(2): 110-118.
[23] Zhang DF, Bo KR, Montejo R, et al. Influence of pelvic floor muscle training alone or as part of a general physical activity program during pregnancy on urinary incontinence, episiotomy and third- or fourth-degree perineal tear: Systematic review and meta-analysis of randomized clinical trials[J]. Acta Obstet Gynecol Scand, 2024, 103(6): 1015-1027.
[24] Krishnan R, Rajpurkar P, Topol EJ. Self-supervised learning in medicine and healthcare[J]. Nat Biomed Eng, 2022, 6(12): 1346-1352.
[25] Chen RJ, Wang JJ, Williamson DFK, et al. Algorithmic fairness in artificial intelligence for medicine and healthcare[J]. Nat Biomed Eng, 2023, 7(6): 719-742.
[26] 国家卫生健康委员会发布《全国医院信息化建设标准与规范(试行)》:明确医院信息化建设内容和建设要求[J]. 医学信息学杂志, 2018(4): 94.
[27] 孙烨祥, 吕筠, 沈鹏, 等. 健康医疗大数据驱动下的疾病防控新模式[J]. 中华流行病学杂志, 2021, 42(8): 1325-1329. SUN Yexiang, LYU Jun, SHEN Peng, et al. A new model for disease control and prevention driven by big data in healthcare[J]. Chinese Journal of Epidemiology, 2021,42(8): 1325-1329.
[28] Sheng XM, Wang J, Shaw JM, et al. Pelvic floor support and symptoms 1 year postpartum in relation to excess body habitus before, during and after first vaginal birth[J]. Female Pelvic Med Reconstr Surg, 2022, 28(5): 273-279.
[29] DeLancey JOL, Masteling M, Pipitone F, et al. Pelvic floor injury during vaginal birth is life-altering and preventable: what can we do about it?[J]. Am J Obstet Gynecol, 2024, 230(3): 279-294.

相关文章 15

[1]	陈瑛翼,游倩,王意,张帆,李凤,季舒铭,徐浩源,饶志勇. 办公室职员肌肉质量减少预测模型的开发与验证[J]. 山东大学学报 (医学版), 2025, 63(4): 26-35.
[2]	王宝炫,焦杰,张厚君,刘奇,于冠英. 衰弱与肌少症评估在胃肠道肿瘤术后结局预测中的应用与展望[J]. 山东大学学报 (医学版), 2025, 63(4): 51-58.
[3]	李永,崔书君,杨飞,张凡,殷晓霞. 基于增强MRI的亚区域影像组学模型可预测乳腺癌患者新辅助化疗后的病理完全反应[J]. 山东大学学报 (医学版), 2025, 63(1): 81-89.
[4]	孙丽娜,白红艳,牛宗格,张福帅,曲仪庆. 基于SII构建及评价预测ARDS住院死亡率的在线临床风险模型[J]. 山东大学学报 (医学版), 2024, 62(7): 10-20.
[5]	郭振江,王宁,赵光远,杜立强,崔朝勃,刘防震. 基于机器学习建立术前预测近端胃癌食管切缘阳性模型[J]. 山东大学学报 (医学版), 2024, 62(7): 78-83.
[6]	王静,刘晓菲,曾荣,许长娟,张锦涛,董亮. 基于机器学习算法鉴定哮喘的坏死性凋亡相关生物标志物[J]. 山东大学学报 (医学版), 2024, 62(7): 21-32.
[7]	刁玉洁,林琳,李文瑄,王洲洋,江蓓,胡迎迎,刘广义. NPR预测ANCA相关血管炎不良肾脏预后及其协同多因素优化模型[J]. 山东大学学报 (医学版), 2024, 62(2): 60-68.
[8]	梁永媛,蔡培飞,郑桂喜. 基于多检验变量和机器学习算法的结肠癌诊断模型建立及价值评估[J]. 山东大学学报 (医学版), 2024, 62(2): 51-59.
[9]	王玉涛,孙岩. 单细胞转录组学与机器学习综合分析揭示腹主动脉瘤潜在生物标志物[J]. 山东大学学报 (医学版), 2024, 62(11): 40-53.
[10]	张景慧,王娟,赵玉洁,段淼,刘毅然,林敏娟,谯旭,李真,左秀丽. 基于机器学习的胃肠道疾病舌诊模型构建[J]. 山东大学学报 (医学版), 2024, 62(1): 38-47.
[11]	张红媛,顾永忠,侯哲. 产科47例子宫切除的临床分析[J]. 山东大学学报 (医学版), 2023, 61(7): 96-100.
[12]	钟璐,薛付忠. 基于贝叶斯网络不确定性推理的肺癌风险预测模型[J]. 山东大学学报 (医学版), 2023, 61(4): 86-94.
[13]	刘亚军,郎昭,郭安忆,刘文勇. 骨科冲击波治疗的智能化发展现状及趋势分析[J]. 山东大学学报 (医学版), 2023, 61(3): 7-13.
[14]	吴南,仉建国,朱源棚,陈癸霖,陈泽夫. 人工智能在脊柱畸形诊疗中的应用[J]. 山东大学学报 (医学版), 2023, 61(3): 14-20.
[15]	朱正阳,沈靖菲,陈思璇,叶梅萍,杨惠泉,周佳南,梁雪,张鑫,张冰. 磁敏感加权成像不同影像组学模型预测胶质瘤IDH基因突变[J]. 山东大学学报 (医学版), 2023, 61(12): 44-50.

多维度评价

Viewed

Full text

Abstract

Cited

Shared

Discussed