{"articleRight":{"code":"F","displayResorceName":"Free","ipAllow":true,"access":true},"searchEn":false,"sessionUser":"","existTranslate":true,"PrevArticle":"10.13558/j.cnki.issn1672-3686.2026.004.002","article":{"id":"24C7A0668D634591BB28F16550B0DC2C","isNewRecord":false,"title":"Construction of a perioperative pain prediction model for patients undergoing flexible ureteroscopic lithotripsy under local anesthesia based on machine learning algorithms","journalTitle":"Clinical Education and General Practice","journalId":"1776813995475472384","publisherId":"f8d5e518b66b4e12af2177121e32f5d2","doi":"10.13558/j.cnki.issn1672-3686.2026.004.003","volumeCode":"24","issueCode":"4","elocationid":"294","language":"en_US","spage":"294","epage":"298","sort":"16723686000000240000000400000002940000000298","advancePub":"normal","fla":"QKYXLCYJY","issn":"1672-3686","pubYear":"2026","pubMonth":"04","purchaseDate":1767542400000,"pubDate":1778145029000,"pdfPath":"/1672-3686/D0186F85D3DC4DBEB11A8CFA12A6D881.pdf","pdfMarkPath":"/1672-3686/D0186F85D3DC4DBEB11A8CFA12A6D881-mark.pdf","createDate":1778145029000,"updateDate":1778145029000,"resourseId":"1eaee8279fbd48749e96d91f6029e79d","intro":"<sec> <p><strong>Objective</strong></p> <p>To investigate the influencing factors of perioperative pain in patients undergoing flexible ureteroscopy lithotripsy (FURL) under local anesthesia, and to construct and validate a predictive model for perioperative pain based on machine learning algorithms.</p> </sec><sec> <p><strong>Methods</strong></p> <p>A total of 224 patients with calculi who underwent FURL under local anesthesia were prospectively enrolled. According to the presence or absence of moderate to severe pain during the perioperative period, patients were divided into a significant pain group (<italic>n</italic>=61) and a non-significant pain group (<italic>n</italic>=163). Relevant clinical data were collected. The influencing factors of significant perioperative pain were analyzed. Through univariate analysis and combined feature selection using LASSO regression, predictive models for perioperative pain in patients undergoing FURL under local anesthesia were constructed based on machine learning algorithms including logistic regression, classification and regression tree (CRT) algorithm, and back propagation neural network (BPNN) algorithm. The predictive value of the models constructed by the three methods for perioperative pain was compared using the receiver operator characteristic (ROC) curve.</p> </sec><sec> <p><strong>Results</strong></p> <p>A total of 9 factors were screened out through univariate analysis and LASSO regression, including gender, complicated diabetes mellitus, American society of anesthesiologists (ASA) classification, mean stone diameter, urinary tract infection, ureteral stricture, indwelling ureteral stent, Hamilton depression scale (HAMD) score, and Hamilton anxiety scale (HAMA) score. Multivariate logistic regression analysis based on these factors showed that gender, ASA classification, mean stone diameter, urinary tract infection, indwelling ureteral stent, HAMD score, and HAMA score were independent influencing factors for perioperative pain (<italic>P</italic><0.05). The predictive model constructed using the CRT decision tree method indicated that gender, mean stone diameter, HAMD score, and HAMA score were categorical factors for perioperative pain. According to the standardized importance of independent variables in the BPNN model, the top five factors influencing perioperative pain were HAMD score, HAMA score, mean stone diameter, ASA classification, and complicated diabetes mellitus. The area under the curve (AUC) of the models constructed by the three machine learning algorithms were all higher than 0.80, with the BPNN model demonstrating the best predictive performance (AUC=0.99).</p> </sec><sec> <p><strong>Conclusion</strong></p> <p>Predictive models for perioperative pain in patients undergoing FURL under local anesthesia based on machine learning algorithms all exhibit good predictive efficacy, among which the BPNN model shows the optimal diagnostic performance.</p> </sec>","specialId":"","includeLanguage":"2","retraction":"未撤稿","price":0.00,"publishScope":"3","printDateStr":"2026-04-30","abstractKeyPoints":"","cResourceType":{"id":"1eaee8279fbd48749e96d91f6029e79d","isNewRecord":false,"resourceNameCh":"免费获取","resourceNameEn":"Free","path":"/images/free.png","code":"F"},"publisher":{"isNewRecord":true,"nameCh":"浙江大学出版社有限责任公司","nameEn":"Zhejiang University Press","coverPath":"1776849242325778432","abbreviated":"JZU","partial":false},"fundingList":[{"isNewRecord":true,"code":"2022097","sourceId":"衢州市第一批指导性科技攻关项目"},{"isNewRecord":true,"code":"2023128","sourceId":"龙游县科技计划项目"}],"mostReadType":"article","latestVersion":"1","baseId":"D0186F85D3DC4DBEB11A8CFA12A6D881","crosUpdateTime":1778145130000,"buyStatus":true,"favoriteStatus":false,"resourseCode_str":"F","status":1,"elocationNum":0,"articleSort":0,"resourceTypeStr":"Free","purchaseDateStr":"2026-01-05","acceptDateStr":"","createDateStr":"2026-05-07","onlineDateStr":"2026-05-07","isAuth":false},"isAuthor":false,"metaInfo":"Pages 294-298","isAuth":true,"journalType":"CH","NextArticle":"10.13558/j.cnki.issn1672-3686.2026.004.004","pdfPath":"https://academax.com/zju-parse/view/pdf/","headingList":[{"id":"2052315070360326144","isNewRecord":false,"nameZh":"论著","nameEn":"Original Articles"}],"islease":false,"lang":"en_US"}