[1]王 芳,胡 星,朱锦舟,等.自动化机器学习在剖宫产术后尿潴留预测模型中的应用[J].医学信息,2023,36(05):41-45.[doi:10.3969/j.issn.1006-1959.2023.05.007]
　WANG Fang,HU Xing,ZHU Jin-zhou,et al.Application of Automated Machine Learning in Prediction Model of Urinary Retention After Cesarean Section[J].Journal of Medical Information,2023,36(05):41-45.[doi:10.3969/j.issn.1006-1959.2023.05.007]
点击复制

自动化机器学习在剖宫产术后尿潴留预测模型中的应用()

医学信息[ISSN:1006-1959/CN:61-1278/R]

卷:
36卷
期数:
2023年05期
页码:
41-45
栏目:
临床信息学
出版日期:
2023-03-01

文章信息/Info

Title:
Application of Automated Machine Learning in Prediction Model of Urinary Retention After Cesarean Section
文章编号:
1006-1959（2023）05-0041-05
作者:
王 芳; 胡 星; 朱锦舟; 等
（1.苏州工业园区星湖医院手术室，江苏 苏州 215000；2.苏州工业园区星湖医院麻醉科，江苏 苏州 215000；3.苏州大学苏州医学院，江苏 苏州 215000）
Author(s):
WANG Fang; HU Xing; ZHU Jin-zhou; et al.
(1.Surgery Room of Suzhou Industrial Park Xinghu Hospital,Suzhou 215000,Jiangsu,China;2.Department of Anesthesiology,Suzhou Industrial Park Xinghu Hospital,Suzhou 215000,Jiangsu,China;3.Suzhou Medical College of Soochow University,Suzhou 215000,Jiangsu,China)
关键词:
剖宫产; 尿潴留; 自动化机器学习; 预测模型; SHAP可视化; LIME可视化
Keywords:
Cesarean section; Urinary retention; Automated machine learning; Prediction model; SHAP visualization; LIME visualization
分类号:
R473
DOI:
10.3969/j.issn.1006-1959.2023.05.007
文献标志码:
A
摘要:
目的 利用自动化机器学习方法，建立剖宫产术后尿潴留预测模型。方法 选取我院2018年1月-2022年1月手术室220例行剖宫产住院产妇，根据是否发生术后尿潴留结局分为尿潴留组（38例）和无尿潴留组（182例）。比较两组生育史及术中术后临床资料，利用H2O平台自动化机器学习框架，建立针对术后尿潴留结局的预测模型，通过绘制ROC曲线，计算曲线下面积（AUC）以评价模型的预测能力，并对模型特征进行可视化呈现。结果 两组疼痛评分、孕前BMI、产次、剖宫产史、胎儿体重、麻醉时间、手术时间、麻醉方式、尿管拔除时间及焦虑情况比较，差异有统计学意义（P＜0.05）；最佳模型为梯度提升机模型（GBM），Gini值0.987，R2为0.653，LogLoss为0.168；模型中重要变量包括疼痛评分、焦虑、麻醉时间、产次、麻醉方式、拔尿管时间及孕前BMI；变量SHAP特征图呈现了变量与模型整体预测的相关性，LIME反映在具体案例中变量的角色；GBM模型的ROC下面积为0.909（95%CI：0.880~0.939），准确度0.947，特异度为0.962，敏感度0.856。结论 基于GBM算法的剖宫产后尿潴留预测模型显示出良好的区分能力，可作为潜在的产后并发症风险初筛工具。
Abstract:
Objective To establish a prediction model of urinary retention after cesarean section by using automatic machine learning method.Methods From January 2018 to January 2022, 220 parturients who underwent cesarean section in the operating room of our hospital were selected and divided into urinary retention group (38 cases) and non-urinary retention group (182 cases) according to whether postoperative urinary retention occurred. The fertility history and intraoperative and postoperative clinical data of the two groups were compared. The H2O platform automated machine learning framework was used to establish a prediction model for postoperative urinary retention outcomes. By drawing the ROC curve, the area under the curve (AUC) was calculated to evaluate the predictive ability of the model, and the model features were visualized.Results There were significant differences in pain score, pre-pregnancy BMI, parity, history of cesarean section, fetal weight, anesthesia time, operation time, anesthesia method, catheter removal time and anxiety between the two groups (P<0.05). The best model was the gradient boosting machine (GBM), Gini value was 0.987, R2 was 0.653, LogLoss was 0.168; important variables in the model included pain score, anxiety, anesthesia time, parity, anesthesia method, catheter removal time and pre-pregnancy BMI. The variable SHAP feature map had shown the correlation between variables and the overall prediction of the model, and LIME reflects the role of variables in specific cases. The area under the ROC of the GBM model was 0.909 (95%CI:0.880-0.939), the accuracy was 0.947, the specificity was 0.962, and the sensitivity was 0.856.Conclusion The prediction model of urinary retention after cesarean section based on GBM algorithm shows good discrimination ability and can be used as a preliminary screening tool for potential risk of postpartum complications.

参考文献/References:

[1]Barba M,Frigerio M,Manodoro S,et al.Postpartum urinary retention: Absolute risk prediction model[J].Low Urin Tract Symptoms,2021,13(2):257-263. [2]Mulder FEM,Hakvoort RA,de Bruin JP,et al.Long-term micturition problems of asymptomatic postpartum urinary retention: a prospective case-control study[J].Int Urogynecol J,2018,29(4):481-488.[3]Handelman GS,Kok HK,Chandra RV,et al.eDoctor: machine learning and the future of medicine[J].J Intern Med,2018,284(6):603-619.[4]Erin LD,Sebastien P.H2O AutoML:Scalable Automatic Machine Learning[C]//7th ICML Workshop on Automated Machine Learning (AutoML).2020.[5]Mulder FE,Hakvoort RA,Schoffelmeer MA,et al.Postpartum urinary retention: a systematic review of adverse effects and management[J].Int Urogynecol J,2014,25(12):1605-1612.[6]Emons WH,Habiboviｃ M,Pedersen SS.Prevalence of anxiety in patients with an implantable cardioverter defibrillator: measurement equivalence of the HADS-A and the STAI-S[J].Qual Life Res,2019,28(11):3107-3116. [7]Thong ISK,Jensen MP,Miró J,et al.The validity of pain intensity measures: what do the NRS, VAS, VRS, and FPS-R measure?[J].Scand J Pain,2018,18(1):99-107.[8]Friedman JH.Greedy function approximation: a gradient boosting machine[J].Annals of Statistics,2001:1189-1232.[9]中华医学会计划生育学分会.剖宫产术后瘢痕子宫孕妇中期妊娠引产的专家共识[J].中华妇产科杂志,2019,54(6):381-386.[10]Humburg J,Troeger C,Holzgreve W,et al.Risk factors in prolonged postpartum urinary retention: an analysis of six cases[J].Arch Gynecol Obstet,2011,283(2):179-183.[11]Gupta A,Pampapati V,Khare C,et al.Postpartum urinary retention in women undergoing instrumental delivery: A cross-sectional analytical study[J].Acta Obstet Gynecol Scand,2021,100(1):41-47.[12]Ain QU,Shetty NKS.Postpartum urinary retention and its associated obstetric risk factors among women undergoing vaginal delivery in tertiary care hospital[J].J Gynecol Obstet Hum Reprod,2021,50(2):101837.[13]Mulder FE,Oude Rengerink K,van der Post JA,et al.Delivery-related risk factors for covert postpartum urinary retention after vaginal delivery[J].Int Urogynecol J,2016,27(1):55-60.[14]Mulder FE,Schoffelmeer MA,Hakvoort RA,et al.Risk factors for postpartum urinary retention: a systematic review and meta-analysis[J].BJOG,2012,119(12):1440-1446. [15]Mulder FE,Hakvoort RA,Schoffelmeer MA,et al.Postpartum urinary retention: a systematic review of adverse effects and management[J].Int Urogynecol J,2014,25(12):1605-1612.[16]Waldum ?魡H,Staff AC,Lukasse M,et al.Intrapartum pudendal nerve block analgesia and risk of postpartum urinary retention: a cohort study[J].Int Urogynecol J,2021,32(9):2383-2391.[17]王诗翔,陈亚静,李淑英,等.剖宫产术后尿潴留的风险评估及预测模型构建[J].全科护理,2021,19(34):4766-4770.[18]袁芳,罗友军,彭丽琴,等.剖宫产术后尿潴留影响因素分析[J].川北医学院学报,2017,32(4):514-517.[19]岳凤洁.二次剖宫产术后并发症发生情况的临床分析[J].医学信息,2018,31(8):114-115.

相似文献/References:

[1]刘汉芬.剖宫产术后护理干预对产妇康复效果的影响分析[J].医学信息,2018,31(01):169.[doi:10.3969/j.issn.1006-1959.2018.01.067]
　LIU Han-fen.Effect of Nursing Intervention on Maternal Rehabilitation after Cesarean Section[J].Journal of Medical Information,2018,31(05):169.[doi:10.3969/j.issn.1006-1959.2018.01.067]
[2]刘 礼.影响剖宫产术后瘢痕子宫足月分娩结局的相关因素分析[J].医学信息,2018,31(02):109.[doi:10.3969/j.issn.1006-1959.2018.02.039]
　LIU Li.Analysis of Related Factors Influencing the Outcome of Full Term Delivery of Scar Uterus after Cesarean Section[J].Journal of Medical Information,2018,31(05):109.[doi:10.3969/j.issn.1006-1959.2018.02.039]
[3]李 欢,刘 俊,夏军辉.妊娠期高血压疾病患者血浆Pro BNP水平及其临床意义[J].医学信息,2018,31(04):158.[doi:10.3969/j.issn.1006-1959.2018.04.059]
　LI Huan,LIU Jun,XIA Jun-hui.Plasma Pro BNP Levels in Patients with Hypertensive Disorder Complicating Pregnancy and their Clinical Significance[J].Journal of Medical Information,2018,31(05):158.[doi:10.3969/j.issn.1006-1959.2018.04.059]
[4]鲁秋彤.经阴道彩色多普勒超声在诊断剖宫产子宫疤痕妊娠中的 价值研究[J].医学信息,2018,31(05):160.[doi:10.3969/j.issn.1006-1959.2018.05.058]
　LU Qiu-tong.The Value of Transvaginal Color Doppler Ultrasound in the Diagnosis of Cesarean Scar Pregnancy[J].Journal of Medical Information,2018,31(05):160.[doi:10.3969/j.issn.1006-1959.2018.05.058]
[5]蒋红梅.子宫下段横形环状压迫缝合术治疗前置胎盘剖宫产术中 胎盘剥离面出血的临床研究[J].医学信息,2018,31(07):120.[doi:10.3969/j.issn.1006-1959.2018.07.040]
　JIANG Hong-mei.Clinical Study on the Treatment of Placental Exfoliation Bleeding in Caesarean Section of Placenta Previa by Transverse Annular Compression Suture of Lower Uterine Segment[J].Journal of Medical Information,2018,31(05):120.[doi:10.3969/j.issn.1006-1959.2018.07.040]
[6]冉会仁.小剂量舒芬太尼在剖宫产术麻醉中预防寒战及 牵拉痛的疗效评价[J].医学信息,2018,31(08):101.[doi:10.3969/j.issn.1006-1959.2018.08.032]
　RAN Hui-ren.Evaluation of the Efficacy of Low-dose Sufentanil in Preventing Chills and Pulling Pain during Cesarean Section Anesthesia[J].Journal of Medical Information,2018,31(05):101.[doi:10.3969/j.issn.1006-1959.2018.08.032]
[7]王玉霞.剖宫产术后子宫疤痕妊娠的临床诊治[J].医学信息,2018,31(10):48.[doi:10.3969/j.issn.1006-1959.2018.10.015]
　WANG Yu-xia.Clinical Diagnosis and Treatment of Uterine Scar Pregnancy after Cesarean Section[J].Journal of Medical Information,2018,31(05):48.[doi:10.3969/j.issn.1006-1959.2018.10.015]
[8]龙俊青.臀位外倒转术的应用研究[J].医学信息,2019,32(12):61.[doi:10.3969/j.issn.1006-1959.2019.12.019]
　LONG Jun-qing.Application Study of Breech Presentation[J].Journal of Medical Information,2019,32(05):61.[doi:10.3969/j.issn.1006-1959.2019.12.019]
[9]熊 丽,蒋明珉,何娅娟.以护士为主导的疼痛管理模式在剖宫产围手术期的应用研究[J].医学信息,2019,32(01):179.[doi:10.3969/j.issn.1006-1959.2019.01.058]
　XIONG Li,JIANG Ming-min,HE Ya-juan.Application Study of Nurse-led Pain Management Model in Perioperative Period of Cesarean Section[J].Journal of Medical Information,2019,32(05):179.[doi:10.3969/j.issn.1006-1959.2019.01.058]
[10]王 颖,石显江,陈斯乔,等.多巴胺预注射在剖宫产腰硬联合麻醉的临床应用[J].医学信息,2019,32(11):150.[doi:10.3969/j.issn.1006-1959.2019.11.044]
　WANG Ying,SHI Xian-jiang,CHEN Si-qiao,et al.Clinical Application of Dopamine Pre-injection in Cesarean Section Combined with Spinal Anesthesia[J].Journal of Medical Information,2019,32(05):150.[doi:10.3969/j.issn.1006-1959.2019.11.044]

更新日期/Last Update: 1900-01-01