首页 期刊 论文 学科刊群 资讯 加盟 关于
EN
image
Article(id=1241329570801045540, tenantId=1146029695717560320, journalId=1227665162245664772, issueId=1241329570129956900, articleNumber=null, orderNo=null, doi=10.20043/j.cnki.MPM.202503139, pmid=null, cstr=null, oa=null, hot=null, price=null, onlineType=0, articleFormat=0, articleType=null, articleTypeStr=null, receivedDate=1741708800000, receivedDateStr=2025-03-12, revisedDate=null, revisedDateStr=null, acceptedDate=null, acceptedDateStr=null, onlineDate=1773885632467, onlineDateStr=2026-03-19, pubDate=1752076800000, pubDateStr=2025-07-10, doiRegisterDate=null, doiRegisterDateStr=null, onlineIssueDate=1773885632467, onlineIssueDateStr=2026-03-19, onlineJustAcceptDate=null, onlineJustAcceptDateStr=null, onlineFirstDate=null, onlineFirstDateStr=null, sourceXml=null, magXml=null, createTime=1773885632467, creator=13701087609, updateTime=1773885632467, updator=13701087609, issue=Issue{id=1241329570129956900, tenantId=1146029695717560320, journalId=1227665162245664772, year='2025', volume='52', issue='13', pageStart='2305', pageEnd='2496', issueExtLink='null', onlineDate='null', pubDate='null', beforeIssueId=null, nextIssueId=null, price=null, status=1, issueComplete=1, articleOrder=1, issueType=-1, specialIssue=null, createTime=1773885632307, creator=13701087609, updateTime=1773885763730, updator=13701087609, preIssue=null, nextIssue=null, ext={EN=IssueExt(id=1241330121425080472, tenantId=1146029695717560320, journalId=1227665162245664772, issueId=1241329570129956900, language=EN, specialIssueTitle=, coverIllustrator=null, specialIssueEditor=, specialIssueAbout=), CN=IssueExt(id=1241330121425080473, tenantId=1146029695717560320, journalId=1227665162245664772, issueId=1241329570129956900, language=CN, specialIssueTitle=, coverIllustrator=null, specialIssueEditor=, specialIssueAbout=)}, issueFiles=null}, startPage=2310, endPage=2316, ext={EN=ArticleExt(id=1241329571124006951, articleId=1241329570801045540, tenantId=1146029695717560320, journalId=1227665162245664772, language=EN, title=Prediction model for tuberculosis recurrence in newly treated patients based on machine learning algorithms, columnId=1240413921954295836, journalTitle=Modern Preventive Medicine, columnName=Epidemiology and Statistical Methods, runingTitle=null, highlight=null, articleAbstract=
Objective

To systematically compare the performance of seven machine learning algorithms in constructing prediction models for tuberculosis (TB) recurrence among newly treated patients in Kashgar, Xinjiang, providing data support for optimizing recurrence intervention strategies in high-burden areas.

Methods

We analyzed 69 476 successfully treated new TB patients from 2016 to 2022 in Kashgar, with follow-up through 2023. Independent predictors were selected through multivariate logistic regression. Seven models (logistic regression, decision tree, random forest, multilayer perceptron, XGBoost, LightGBM, and elastic net) were developed and validated. The optimal model was interpreted using SHapley Additive exPlanations (SHAP).

Results

Among 69 476 cases, 9 444 (13.59%) experienced recurrence by 2023. Fourteen independent predictors were identified. The seven models showed AUC values ranging from 0.705 to 0.762 in the training set, with the decision tree model performing best (AUC=0.762, 95%CI: 0.758-0.766) and demonstrating good calibration. SHAP analysis revealed sputum culture results at diagnosis, local TB burden, and treatment modality as the top three predictive factors.

Conclusion

The decision tree model based on routine surveillance data shows high predictive performance for TB recurrence, with interpretable features that can facilitate early identification of high-risk individuals in clinical practice.

, correspAuthors=null, authorNote=null, correspAuthorsNote=null, copyrightStatement=null, copyrightOwner=null, extLink=null, articleAbsUrl=null, sourceXml=null, magXml=null, pdfUrl=null, pdf=null, pdfFileSize=null, pdfExtLink=null, richHtmlUrl=null, mobilePdfUrl=null, reviewReport=null, pdfFirstPage=null, abstractGraph=null, abstractGraphContent=null, abstractVideo=null, citation=null, cebUrl=null, magXmlContent=null, mapNumber=null, authorCompany=null, fund=null, authors=null, authorsList=KAMILI Mai-ri-ha-ba, YIMAMU Mai-wu-la-jiang, Yan-jie WANG, Yu-wei WANG, ABULIMITI A-li-mi-re, KAMILI Mai-di-nu-er, Yang XIANG), CN=ArticleExt(id=1241329573468622902, articleId=1241329570801045540, tenantId=1146029695717560320, journalId=1227665162245664772, language=CN, title=基于机器学习算法的初治结核病患者复发预测模型研究, columnId=1228016567632462653, journalTitle=现代预防医学, columnName=流行病与统计方法, runingTitle=null, highlight=null, articleAbstract=
目的

基于新疆喀什地区结核病常规监测数据，系统比较7种机器学习算法构建的初治结核病患者复发预测模型的效能，为高负担地区优化复发干预策略提供数据支持。

方法

以喀什地区2016—2022年成功治疗的初治结核病患者为研究对象，分析其截至2023年末的复发情况。通过多因素logistic回归筛选独立预测因子，并基于logistic回归、决策树、随机森林、多层感知器、极限梯度提升树、轻量级梯度提升机算法和弹性网络构建并验证复发预测模型，对最优模型通过采用沙普利加性解释（SHapley additive exPlanations, SHAP）法解析变量贡献度。

结果

纳入的69 476例研究对象截至2023年末复发9 444例（13.59%）。通过logistic回归确定出14个独立预测因子，7种预测模型训练集AUC值范围为0.705～0.762，其中决策树模型表现最优（AUC=0.762, 95%CI: 0.758～0.766），且校准度良好。基于SHAP值的变量重要性排序结果显示，初诊痰培养结果、现住地结核病负担及就诊方式是前三位重要预测因子。

结论

基于常规监测数据构建的决策树模型对初治结核复发具有较高预测效能，其可解释性特征有助于临床优先识别高危个体。

, correspAuthors=null, authorNote=null, correspAuthorsNote=
向阳，E-mail:
, copyrightStatement=本刊刊出的所有文章不代表中华预防医学会和本刊编委会的观点，除非特别声明。, copyrightOwner=中华预防医学会和四川大学华西公共卫生学院, extLink=null, articleAbsUrl=null, sourceXml=pBMSVgErGfBNiHupA+ihCg==, magXml=5hNl7j2O9MvT2ex+P9t6tw==, pdfUrl=null, pdf=XgzY+GYGbFaSGhagX3hz0g==, pdfFileSize=1144585, pdfExtLink=null, richHtmlUrl=null, mobilePdfUrl=null, reviewReport=null, pdfFirstPage=null, abstractGraph=Wyu7wOls7n9Qsmq85FePjw==, abstractGraphContent=null, abstractVideo=null, citation=null, cebUrl=null, magXmlContent=VxcdZtgSOOU/5FZXbHzr4A==, mapNumber=null, authorCompany=null, fund=null, authors=

买日哈巴·卡米力（1998—），女，硕士在读，研究方向：传染病流行病学

, authorsList=买日哈巴·卡米力, 买吾拉江·依马木, 王艳杰, 王雨薇, 阿丽米热·阿不力米提, 麦迪努尔·卡米力, 向阳)}, authors=[Author(id=1241329751189672072, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, orderNo=0, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=null, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1241329751290335370, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, authorId=1241329751189672072, language=EN, stringName=KAMILI Mai-ri-ha-ba, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=School of Public Health, Xinjiang Medical University, Urumqi, Xinjiang 830017, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1241329751390998667, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, authorId=1241329751189672072, language=CN, stringName=买日哈巴·卡米力, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1.新疆医科大学公共卫生学院，新疆 乌鲁木齐 830017, bio={"content":"

买日哈巴·卡米力（1998—），女，硕士在读，研究方向：传染病流行病学

"}, bioImg=null, bioContent=

买日哈巴·卡米力（1998—），女，硕士在读，研究方向：传染病流行病学

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1241329750900265088, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, xref=1., ext=[AuthorCompanyExt(id=1241329750908653697, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, companyId=1241329750900265088, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=School of Public Health, Xinjiang Medical University, Urumqi, Xinjiang 830017, China), AuthorCompanyExt(id=1241329750912848002, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, companyId=1241329750900265088, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1.新疆医科大学公共卫生学院，新疆 乌鲁木齐 830017)])]), Author(id=1241329751474884749, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, orderNo=1, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=null, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1241329751533605007, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, authorId=1241329751474884749, language=EN, stringName=YIMAMU Mai-wu-la-jiang, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=2, address=null, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1241329751592325264, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, authorId=1241329751474884749, language=CN, stringName=买吾拉江·依马木, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=2, address=2.喀什地区疾病预防控制中心, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1241329751026094211, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, xref=2., ext=[AuthorCompanyExt(id=1241329751034482820, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, companyId=1241329751026094211, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2.喀什地区疾病预防控制中心)])]), Author(id=1241329751676211346, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, orderNo=2, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=null, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1241329751776874644, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, authorId=1241329751676211346, language=EN, stringName=Yan-jie WANG, firstName=Yan-jie, middleName=null, lastName=WANG, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=School of Public Health, Xinjiang Medical University, Urumqi, Xinjiang 830017, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1241329751915286677, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, authorId=1241329751676211346, language=CN, stringName=王艳杰, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1.新疆医科大学公共卫生学院，新疆 乌鲁木齐 830017, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1241329750900265088, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, xref=1., ext=[AuthorCompanyExt(id=1241329750908653697, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, companyId=1241329750900265088, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=School of Public Health, Xinjiang Medical University, Urumqi, Xinjiang 830017, China), AuthorCompanyExt(id=1241329750912848002, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, companyId=1241329750900265088, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1.新疆医科大学公共卫生学院，新疆 乌鲁木齐 830017)])]), Author(id=1241329752020144281, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, orderNo=3, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=null, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1241329752141779099, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, authorId=1241329752020144281, language=EN, stringName=Yu-wei WANG, firstName=Yu-wei, middleName=null, lastName=WANG, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=School of Public Health, Xinjiang Medical University, Urumqi, Xinjiang 830017, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1241329752250831004, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, authorId=1241329752020144281, language=CN, stringName=王雨薇, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1.新疆医科大学公共卫生学院，新疆 乌鲁木齐 830017, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1241329750900265088, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, xref=1., ext=[AuthorCompanyExt(id=1241329750908653697, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, companyId=1241329750900265088, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=School of Public Health, Xinjiang Medical University, Urumqi, Xinjiang 830017, China), AuthorCompanyExt(id=1241329750912848002, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, companyId=1241329750900265088, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1.新疆医科大学公共卫生学院，新疆 乌鲁木齐 830017)])]), Author(id=1241329752334717087, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, orderNo=4, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=null, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1241329752431186081, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, authorId=1241329752334717087, language=EN, stringName=ABULIMITI A-li-mi-re, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=School of Public Health, Xinjiang Medical University, Urumqi, Xinjiang 830017, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1241329752531849378, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, authorId=1241329752334717087, language=CN, stringName=阿丽米热·阿不力米提, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1.新疆医科大学公共卫生学院，新疆 乌鲁木齐 830017, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1241329750900265088, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, xref=1., ext=[AuthorCompanyExt(id=1241329750908653697, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, companyId=1241329750900265088, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=School of Public Health, Xinjiang Medical University, Urumqi, Xinjiang 830017, China), AuthorCompanyExt(id=1241329750912848002, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, companyId=1241329750900265088, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1.新疆医科大学公共卫生学院，新疆 乌鲁木齐 830017)])]), Author(id=1241329752628318372, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, orderNo=5, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=null, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1241329752716398758, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, authorId=1241329752628318372, language=EN, stringName=KAMILI Mai-di-nu-er, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=3, address=null, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1241329752804479143, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, authorId=1241329752628318372, language=CN, stringName=麦迪努尔·卡米力, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=3, address=3.新疆医科大学临床医学部, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1241329751105785989, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, xref=3., ext=[AuthorCompanyExt(id=1241329751114174598, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, companyId=1241329751105785989, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=3.新疆医科大学临床医学部)])]), Author(id=1241329752871588011, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, orderNo=6, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=893664450@qq.com, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1241329752980639919, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, authorId=1241329752871588011, language=EN, stringName=Yang XIANG, firstName=Yang, middleName=null, lastName=XIANG, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=School of Public Health, Xinjiang Medical University, Urumqi, Xinjiang 830017, China, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null), CN=AuthorExt(id=1241329753072914608, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, authorId=1241329752871588011, language=CN, stringName=向阳, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=1, address=1.新疆医科大学公共卫生学院，新疆 乌鲁木齐 830017, bio=null, bioImg=null, bioContent=null, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1241329750900265088, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, xref=1., ext=[AuthorCompanyExt(id=1241329750908653697, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, companyId=1241329750900265088, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=School of Public Health, Xinjiang Medical University, Urumqi, Xinjiang 830017, China), AuthorCompanyExt(id=1241329750912848002, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, companyId=1241329750900265088, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1.新疆医科大学公共卫生学院，新疆 乌鲁木齐 830017)])])], keywords=[Keyword(id=1241329753265852599, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, language=EN, orderNo=1, keyword=Tuberculosis), Keyword(id=1241329753387487418, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, language=EN, orderNo=2, keyword=Recurrence), Keyword(id=1241329753471373503, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, language=EN, orderNo=3, keyword=Machine learning), Keyword(id=1241329753584619713, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, language=EN, orderNo=4, keyword=Treatment success), Keyword(id=1241329753689477315, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, language=CN, orderNo=1, keyword=结核病), Keyword(id=1241329753769169092, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, language=CN, orderNo=2, keyword=复发), Keyword(id=1241329753861443782, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, language=CN, orderNo=3, keyword=机器学习), Keyword(id=1241329753915969736, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, language=CN, orderNo=4, keyword=治疗成功)], refs=[Reference(id=1241329756109590769, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, doi=null, pmid=null, pmcid=null, year=2024, volume=39, issue=4, pageStart=403, pageEnd=411, url=null, language=null, rfNumber=［1］, rfOrder=0, authorNames=李涛, 赵雁林, 贾忠伟, journalName=疾病监测, refType=null, unstructuredReference=李涛， 赵雁林， 贾忠伟． 结核病复发及其影响因素研究进展［J］.疾病监测，2024，39（4）：403-411., articleTitle=结核病复发及其影响因素研究进展, refAbstract=null), Reference(id=1241329756256391414, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, doi=null, pmid=null, pmcid=null, year=2024, volume=39, issue=4, pageStart=403, pageEnd=411, url=null, language=null, rfNumber=［1］, rfOrder=1, authorNames=Li T, Zhao YL, Jia ZW, journalName=Disease Surveillance, refType=null, unstructuredReference=Li T， Zhao YL, Jia ZW. Progress in research of tuberculosis recurrence and influencing factors［J］. Disease Surveillance. 2024, 39（4）: 403-411.（In Chinese）, articleTitle=Progress in research of tuberculosis recurrence and influencing factors, refAbstract=null), Reference(id=1241329756369637628, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, doi=null, pmid=null, pmcid=null, year=2021, volume=18, issue=20, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=［2］, rfOrder=2, authorNames=Teferi MY, El-Khatib Z, Boltena MT, journalName=International Journal of Environmental Research and Public Health, refType=null, unstructuredReference=Teferi MY. El-Khatib Z, Boltena MT .et al. Tuberculosis treatment outcome and predictors in Africa: a systematic review and Meta-Analysis ［J］. International Journal of Environmental Research and Public Health, 2021，18（20）: 10678., articleTitle=Tuberculosis treatment outcome and predictors in Africa: a systematic review and Meta-Analysis, refAbstract=null), Reference(id=1241329756512243971, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, doi=null, pmid=null, pmcid=null, year=2022, volume=44, issue=7, pageStart=690, pageEnd=697, url=null, language=null, rfNumber=［3］, rfOrder=3, authorNames=钟倩红, 马晓慧, 钟永辉, journalName=中国防痨杂志, refType=null, unstructuredReference=钟倩红．马晓慧，钟永辉，等.2014-2018年佛山市肺结核患者复发情况及影响因素分析［J］. 中国防痨杂志，2022,44（7）:690-697., articleTitle=2014-2018年佛山市肺结核患者复发情况及影响因素分析, refAbstract=null), Reference(id=1241329756642267400, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, doi=null, pmid=null, pmcid=null, year=2022, volume=44, issue=7, pageStart=690, pageEnd=697, url=null, language=null, rfNumber=［3］, rfOrder=4, authorNames=Zhong QH, Ma XH, Zhong YH, journalName=Chinese Journal of Antituberculosis, refType=null, unstructuredReference=Zhong QH, Ma XH, Zhong YH, et al. Analysis of recurrence rate of pulmonary tuberculosis patients in Foshan and the influencing factors from 2014 to 2018 ［J］. Chinese Journal of Antituberculosis,2022, 44（7）: 690-697.（In Chinese）, articleTitle=Analysis of recurrence rate of pulmonary tuberculosis patients in Foshan and the influencing factors from 2014 to 2018, refAbstract=null), Reference(id=1241329756751319308, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, doi=null, pmid=null, pmcid=null, year=2015, volume=10, issue=5, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=［4］, rfOrder=5, authorNames=Hung CL, Chien JY, Ou CY, journalName=PLOS One, refType=null, unstructuredReference=Hung CL, Chien JY, Ou CY. Associated factors for tuberculosis recurrence in Taiwan: a nationwide nested case-control study from 1998 to 2010［J］. PLOS One. 2015, 10（5）: e0124822., articleTitle=Associated factors for tuberculosis recurrence in Taiwan: a nationwide nested case-control study from 1998 to 2010, refAbstract=null), Reference(id=1241329756856176916, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, doi=null, pmid=null, pmcid=null, year=2022, volume=84, issue=5, pageStart=648, pageEnd=657, url=null, language=null, rfNumber=［5］, rfOrder=6, authorNames=Luo Y, Xue Y, Song HJ, journalName=Journal of Infection, refType=null, unstructuredReference=Luo Y, Xue Y, Song HJ, et al. Machine learning based on routine lahoratory indicators promoting the discrimination between active tuberculosis and latent tuberculosis infection［J］. Journal of Infection,2022, 84（5）: 648-657., articleTitle=Machine learning based on routine lahoratory indicators promoting the discrimination between active tuberculosis and latent tuberculosis infection, refAbstract=null), Reference(id=1241329756944257305, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, doi=null, pmid=null, pmcid=null, year=2024, volume=10, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=［6］, rfOrder=7, authorNames=Falcao I, Cardoso DL, Coutinho DSA, journalName=PeerJ Computer Science, refType=null, unstructuredReference=Falcao I, Cardoso DL, Coutinho DSA, et al. Model for predicting drug resistance based on the clinical profile of tuberculosis patients using machine learning techniques ［J］. PeerJ Computer Science, 2024, 10:e2246., articleTitle=Model for predicting drug resistance based on the clinical profile of tuberculosis patients using machine learning techniques, refAbstract=null), Reference(id=1241329757095252253, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, doi=null, pmid=null, pmcid=null, year=2024, volume=29, issue=1, pageStart=618, pageEnd=null, url=null, language=null, rfNumber=［7］, rfOrder=8, authorNames=Kong HB, Li Y, Shen Y, journalName=European Journal of Medical Research, refType=null, unstructuredReference=Kong HB, Li Y, Shen Y， et al. Predicting the risk of pulmonary embolism in patients with tuberculosis using machine learning algorithms ［J］. European Journal of Medical Research, 2024, 29（1）:618., articleTitle=Predicting the risk of pulmonary embolism in patients with tuberculosis using machine learning algorithms, refAbstract=null), Reference(id=1241329757191721254, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, doi=null, pmid=null, pmcid=null, year=2024, volume=14, issue=1, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=［8］, rfOrder=9, authorNames=Anh NK, Phat NK, Thu NQ, journalName=Scientific Reports, refType=null, unstructuredReference=Anh NK. Phat NK, Thu NQ, et al. Discovery of urinary biosignatures for tuberculosis and nontuberculous mycobacteria classification using metabolomics and machine learning ［J］. Scientific Reports, 2024, 14（1）: 15312., articleTitle=Discovery of urinary biosignatures for tuberculosis and nontuberculous mycobacteria classification using metabolomics and machine learning, refAbstract=null), Reference(id=1241329757267218732, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, doi=null, pmid=null, pmcid=null, year=2024, volume=14, issue=1, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=［9］, rfOrder=10, authorNames=Chen JF, Jiang YL, Li ZH, journalName=Scientific Reports, refType=null, unstructuredReference=Chen JF, Jiang YL, Li ZH, et al. Predictive machine learning models for anticipating loss to follow-up in tuberculosis patients throughout anti-TB treatment journey［J］. Scientific Reports, 2024, 14（1）: 24685., articleTitle=Predictive machine learning models for anticipating loss to follow-up in tuberculosis patients throughout anti-TB treatment journey, refAbstract=null), Reference(id=1241329757376270642, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, doi=null, pmid=null, pmcid=null, year=2019, volume=30, issue=4, pageStart=355, pageEnd=359, url=null, language=null, rfNumber=［10］, rfOrder=11, authorNames=彭红, 虞浩, 姜洁, journalName=江苏预防医学, refType=null, unstructuredReference=彭红， 虞浩， 姜洁， 等．江苏省初治结核病复发流行病学特征及影响因素［J］.江苏预防医学，2019，30（4）：355-359., articleTitle=江苏省初治结核病复发流行病学特征及影响因素, refAbstract=null), Reference(id=1241329757485322550, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, doi=null, pmid=null, pmcid=null, year=2019, volume=30, issue=4, pageStart=355, pageEnd=359, url=null, language=null, rfNumber=［10］, rfOrder=12, authorNames=Peng H, Yu H, Jiang J, journalName=Jiangsu Journal of Preventive Medicine, refType=null, unstructuredReference=Peng H, Yu H, Jiang J, et al. Epidemiological characteristics and associated factors with tuberculosis recurrence in Jiangsu province［J］.Jiangsu Journal of Preventive Medicine, 2019, 30 （4）: 355-359.（In Chinese）, articleTitle=Epidemiological characteristics and associated factors with tuberculosis recurrence in Jiangsu province, refAbstract=null), Reference(id=1241329757594374465, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, doi=null, pmid=null, pmcid=null, year=2022, volume=28, issue=5, pageStart=684, pageEnd=689, url=null, language=null, rfNumber=［11］, rfOrder=13, authorNames=Ruan QL, Yang QL, Sun F, journalName=Clinical Microbiology and Infection, refType=null, unstructuredReference=Ruan QL, Yang QL, Sun F, et al. Recurrent pulmonary tuberculosis after treatment success: a population-based retrospective study in China［J］. Clinical Microbiology and Infection, 2022, 28（5）: 684-689., articleTitle=Recurrent pulmonary tuberculosis after treatment success: a population-based retrospective study in China, refAbstract=null), Reference(id=1241329757682454855, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, doi=null, pmid=null, pmcid=null, year=2023, volume=127, issue=null, pageStart=1, pageEnd=10, url=null, language=null, rfNumber=［12］, rfOrder=14, authorNames=Eksombatchai D, Jeong D, Mok J, journalName=International Journal of Infectious Diseases, refType=null, unstructuredReference=Eksombatchai D, Jeong D, Mok J, et al. Sex differences in the impact of diabetes mellitus on tuberculosis recurrence: a retrospective National cohort study［J］. International Journal of Infectious Diseases,2023, 127: 1-10., articleTitle=Sex differences in the impact of diabetes mellitus on tuberculosis recurrence: a retrospective National cohort study, refAbstract=null), Reference(id=1241329757799895374, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, doi=null, pmid=null, pmcid=null, year=2022, volume=34, issue=3, pageStart=268, pageEnd=271, url=null, language=null, rfNumber=［13］, rfOrder=15, authorNames=周旭, 王巧智, 万燕萍, journalName=预防医学, refType=null, unstructuredReference=周旭， 王巧智， 万燕萍， 等．衡阳市成功治疗的肺结核患者5年复发影响因素分析［J］.预防医学，2022,34（3）:268-271., articleTitle=衡阳市成功治疗的肺结核患者5年复发影响因素分析, refAbstract=null), Reference(id=1241329757929918809, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, doi=null, pmid=null, pmcid=null, year=2022, volume=34, issue=3, pageStart=268, pageEnd=271, url=null, language=null, rfNumber=［13］, rfOrder=16, authorNames=Zhou X, Wang QZ, Wan YP, journalName=Journal of Preventive Medicine, refType=null, unstructuredReference=Zhou X, Wang QZ, Wan YP, et al. Factors affecting the five-year recurrence among successfully-treated pulmonary tuberculosis patients in Hengyang City ［J］. Journal of Preventive Medicine, 2022,34（3）: 268-271.（In Chinese）, articleTitle=Factors affecting the five-year recurrence among successfully-treated pulmonary tuberculosis patients in Hengyang City, refAbstract=null), Reference(id=1241329758043165023, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, doi=null, pmid=null, pmcid=null, year=2022, volume=2022, issue=1, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=［14］, rfOrder=17, authorNames=Mujtaba MA, Richardson M, Shahzad H, journalName=Journal of Tropical Medicine, refType=null, unstructuredReference=Mujtaba MA, Richardson M, Shahzad H, et al. Demographic and clinical determinants of tuberculosis and TB recurrence: a Double-Edged retrospective study from Pakistan ［J］. Journal of Tropical Medicine, 2022, 2022（1）: 4408306., articleTitle=Demographic and clinical determinants of tuberculosis and TB recurrence: a Double-Edged retrospective study from Pakistan, refAbstract=null), Reference(id=1241329758189965671, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, doi=null, pmid=null, pmcid=null, year=2024, volume=11, issue=1, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=［15］, rfOrder=18, authorNames=Vega V, Cabrera-Sanchez J, Rodríguez S, journalName=BMJ Open Respir Res, refType=null, unstructuredReference=Vega V, Cabrera-Sanchez J, Rodríguez S, et al. Risk factors for pulmonary tuberculosis recurrence, relapse and reinfection: a systematic review and meta-analysis ［J］. BMJ Open Respir Res,2024, 11（1）: e002281., articleTitle=Risk factors for pulmonary tuberculosis recurrence, relapse and reinfection: a systematic review and meta-analysis, refAbstract=null), Reference(id=1241329758441623925, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, doi=null, pmid=null, pmcid=null, year=2022, volume=116, issue=null, pageStart=14, pageEnd=20, url=null, language=null, rfNumber=［16］, rfOrder=19, authorNames=Qiu BB, Wu ZC, Tao BL, journalName=International Journal of Infectious Diseases, refType=null, unstructuredReference=Qiu BB, Wu ZC, Tao BL, et al. Risk factors for types of recurrent tuberculosis （reactivation versus reinfection）: A global systematic review and meta-analysis ［J］. International Journal of Infectious Diseases, 2022, 116: 14-20., articleTitle=Risk factors for types of recurrent tuberculosis （reactivation versus reinfection）: A global systematic review and meta-analysis, refAbstract=null)], funds=[Fund(id=1241329755744686308, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, awardId=81860589, language=CN, fundingSource=国家自然科学基金(81860589), fundOrder=null, country=null), Fund(id=1241329755841155303, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, awardId=2022 D01C203, language=CN, fundingSource=自治区自然科学基金(2022 D01C203), fundOrder=null, country=null), Fund(id=1241329755925041385, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, awardId=null, language=CN, fundingSource=新疆维吾尔自治区”十四五“高等学校特色学科-公共卫生与预防医学, fundOrder=null, country=null)], companyList=[AuthorCompany(id=1241329750900265088, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, xref=1., ext=[AuthorCompanyExt(id=1241329750908653697, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, companyId=1241329750900265088, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=School of Public Health, Xinjiang Medical University, Urumqi, Xinjiang 830017, China), AuthorCompanyExt(id=1241329750912848002, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, companyId=1241329750900265088, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=1.新疆医科大学公共卫生学院，新疆 乌鲁木齐 830017)]), AuthorCompany(id=1241329751026094211, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, xref=2., ext=[AuthorCompanyExt(id=1241329751034482820, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, companyId=1241329751026094211, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=2.喀什地区疾病预防控制中心)]), AuthorCompany(id=1241329751105785989, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, xref=3., ext=[AuthorCompanyExt(id=1241329751114174598, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, companyId=1241329751105785989, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=3.新疆医科大学临床医学部)])], figs=[ArticleFig(id=1241329754167627980, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, language=EN, label=Figure 1, caption=ROC curves of the recurrence prediction model for training (A) and validation set (B), figureFileSmall=m12YaBjqoDtBRcnMhe3X5Q==, figureFileBig=Wyu7wOls7n9Qsmq85FePjw==, tableContent=null), ArticleFig(id=1241329754259902669, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, language=CN, label=图1, caption=复发预测模型训练集和验证集的ROC图

注：A为训练集；B为验证集。

, figureFileSmall=m12YaBjqoDtBRcnMhe3X5Q==, figureFileBig=Wyu7wOls7n9Qsmq85FePjw==, tableContent=null), ArticleFig(id=1241329754406703311, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, language=EN, label=Figure 2, caption=Calibration curves of recurrence prediction model for training and validation set, figureFileSmall=6kifiYUuwywrJZWr49V/WA==, figureFileBig=xnmq1ENDjMys1bghLWyiyw==, tableContent=null), ArticleFig(id=1241329754528338129, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, language=CN, label=图2, caption=复发预测模型训练集和验证集的校准曲线图

注：A为训练集；B为验证集。

, figureFileSmall=6kifiYUuwywrJZWr49V/WA==, figureFileBig=xnmq1ENDjMys1bghLWyiyw==, tableContent=null), ArticleFig(id=1241329754654167250, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, language=EN, label=Figure 3, caption=Feature importance ranking, SHAP swarm plot and SHAP value feature dependence plot, figureFileSmall=AfcIr9LmT/KWphDT+avjkA==, figureFileBig=+9bouoJOKvHxv7bipDdM5w==, tableContent=null), ArticleFig(id=1241329754754830548, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, language=CN, label=图3, caption=特征重要性排序、SHAP蜂群图及SHAP值的特征依赖图

注：A这特征重要性排序；B为SHAP蜂群图；C为SHAP值的特征依赖图。

, figureFileSmall=AfcIr9LmT/KWphDT+avjkA==, figureFileBig=+9bouoJOKvHxv7bipDdM5w==, tableContent=null), ArticleFig(id=1241329754901631190, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, language=EN, label=Table 1, caption=

Comparison of the training set and the validation set

, figureFileSmall=null, figureFileBig=null, tableContent=
特征训练集验证集χ2值P值
性别0.3170.574
女性24 12610 292
男性24 50610 552
年龄段（岁）2.8730.412
0 ～ <304 5151 914
30 ～ <455 0492 141
45 ～ <6010 7984 532
≥6028 27012 257
职业3.7080.295
学生和教师1 263530
居家人员4 3601 781
农民42 01818 118
其他991415
就诊方式0.5370.464
被动方式25 76711 107
主动方式22 8659 737
是否单纯结核性胸膜炎0.0030.958
否47 99520 572
是637272
治疗结局0.0460.830
完成疗程37 63016 113
治愈11 0024 731
是否使用FDC0.0390.844
否5 8892 513
是42 74318 331
初诊痰检结果0.1670.683
阴性43 34518 556
阳性5 2872 288
初治两个月末痰检结果0.0540.817
阴性48 24620 675
阳性386169
初诊痰培养结果0.4100.815
阴性12 7785 516
阳性6 5412 818
无结果29 31312 510
诊疗延误时长（d）0.0070.934
<3023 73710 181
≥3024 89510 663
治疗方案4.2420.120
2HRZE/4HR47 03420 149
2HRZE/7-10HRE1 156531
其他442164
管理单位类型0.5230.469
医院20 9599 045
疾病预防控制中心27 67311 799
现住地结核病负担0.7700.857
低4 7011 999
中8 3463 533
高17 1217 340
严重18 4647 972
), ArticleFig(id=1241329755061014744, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, language=CN, label=表1, caption=

训练集和验证集比较

, figureFileSmall=null, figureFileBig=null, tableContent=
特征训练集验证集χ2值P值
性别0.3170.574
女性24 12610 292
男性24 50610 552
年龄段（岁）2.8730.412
0 ～ <304 5151 914
30 ～ <455 0492 141
45 ～ <6010 7984 532
≥6028 27012 257
职业3.7080.295
学生和教师1 263530
居家人员4 3601 781
农民42 01818 118
其他991415
就诊方式0.5370.464
被动方式25 76711 107
主动方式22 8659 737
是否单纯结核性胸膜炎0.0030.958
否47 99520 572
是637272
治疗结局0.0460.830
完成疗程37 63016 113
治愈11 0024 731
是否使用FDC0.0390.844
否5 8892 513
是42 74318 331
初诊痰检结果0.1670.683
阴性43 34518 556
阳性5 2872 288
初治两个月末痰检结果0.0540.817
阴性48 24620 675
阳性386169
初诊痰培养结果0.4100.815
阴性12 7785 516
阳性6 5412 818
无结果29 31312 510
诊疗延误时长（d）0.0070.934
<3023 73710 181
≥3024 89510 663
治疗方案4.2420.120
2HRZE/4HR47 03420 149
2HRZE/7-10HRE1 156531
其他442164
管理单位类型0.5230.469
医院20 9599 045
疾病预防控制中心27 67311 799
现住地结核病负担0.7700.857
低4 7011 999
中8 3463 533
高17 1217 340
严重18 4647 972
), ArticleFig(id=1241329755161678042, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, language=EN, label=Table 2, caption=

Study variable coding table

, figureFileSmall=null, figureFileBig=null, tableContent=
变量赋值
复发状态未复发=0，复发=1
性别女=0，男=1
年龄（岁）0 ～ <30 =1，30 ～<45 =2，45 ～<60 =3，≥60 =4
职业学生和教师=1，居家人员=2，农民=3,其他=4
就诊方式主动方式=0，被动方式=1
是否单纯结核性胸膜炎否=0，是=1
治疗结局治愈=0，完成疗程=1
是否使用FDC否=0，是=1
初诊痰检结果阴性=0，阳性=1
初治两个月末痰检结果阴性=0，阳性=1
初诊痰培养结果阴性=0，阳性=1，无结果=2
诊疗延误时长（d）<30 =0，≥30 =1
治疗方案2HRZE/4HR =1，2HRZE/7-10HRE=2，其他=3
管理单位类型医院=0，疾病预防控制中心=1
现住地结核病负担低=1，中=2，高=3，严重=4
), ArticleFig(id=1241329755266535644, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, language=CN, label=表2, caption=

研究变量赋值表

, figureFileSmall=null, figureFileBig=null, tableContent=
变量赋值
复发状态未复发=0，复发=1
性别女=0，男=1
年龄（岁）0 ～ <30 =1，30 ～<45 =2，45 ～<60 =3，≥60 =4
职业学生和教师=1，居家人员=2，农民=3,其他=4
就诊方式主动方式=0，被动方式=1
是否单纯结核性胸膜炎否=0，是=1
治疗结局治愈=0，完成疗程=1
是否使用FDC否=0，是=1
初诊痰检结果阴性=0，阳性=1
初治两个月末痰检结果阴性=0，阳性=1
初诊痰培养结果阴性=0，阳性=1，无结果=2
诊疗延误时长（d）<30 =0，≥30 =1
治疗方案2HRZE/4HR =1，2HRZE/7-10HRE=2，其他=3
管理单位类型医院=0，疾病预防控制中心=1
现住地结核病负担低=1，中=2，高=3，严重=4
), ArticleFig(id=1241329755346227422, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, language=EN, label=Table 3, caption=

Univariate and multivariate logistic regression analysis of factors influencing recurrence

, figureFileSmall=null, figureFileBig=null, tableContent=
特征复发未复发单因素logistic回归分析多因素logistic回归分析
OR（95%CI）P值aOR（95%CI）P值
性别
女性3 14220 98411
男性3 46821 0381.101 (1.045 ～ 1.160)<0.0011.175 (1.113 ～ 1.241)<0.001
年龄段（岁）
0 ～ <303064 20911
30 ～ <454004 6491.183 (1.014 ～ 1.381)0.0331.092 (0.928 ～ 1.284)0.288
45 ～ <601 4629 3362.154 (1.894 ～ 2.449)<0.0011.864 (1.623 ～ 2.140)<0.001
≥604 44223 8282.564 (2.273 ～ 2.892)<0.0012.200 (1.930 ～ 2.508)<0.001
职业
学生和教师511 21211
居家人员5313 8293.296 (2.455 ～ 4.424)<0.0011.712 (1.251 ～ 2.343)0.001
农民5 95436 0643.923 (2.961 ～ 5.199)<0.0012.227 (1.651 ～ 3.004)<0.001
其他749171.918 (1.329 ～ 2.768)0.0011.255 (0.857 ～ 1.838)0.243
就诊方式
被动方式4 67821 08911
主动方式1 93220 9330.416 (0.393 ～ 0.440)<0.0010.515 (0.484 ～ 0.548)<0.001
是否单纯结核性胸膜炎
否6 59741 39811
是136240.131 (0.075 ～ 0.227)<0.0010.333 (0.182 ～ 0.611)<0.001
治疗结局
完成疗程5 49132 13911
治愈1 1199 8830.663 (0.619 ～ 0.709)<0.0010.543 (0.483 ～ 0.612)<0.001
是否使用FDC
否6515 23811
是5 95936 7841.303 (1.196 ～ 1.420)<0.0011.190 (1.072 ～ 1.321)0.001
初诊痰检结果
阴性5 67737 66811
阳性9334 3541.422 (1.318 ～ 1.534)<0.0012.523 (2.240 ～ 2.843)<0.001
初治两个月末痰检结果
阴性6 52341 72311
阳性872991.861 (1.464 ～ 2.366)<0.0012.027 (1.567 ～ 2.622)<0.001
初诊痰培养结果
阴性71112 06711
阳性6145 9001.844 (1.650 ～ 2.061)<0.0011.728 (1.520 ～ 1.965)<0.001
无结果5 25824 0553.710 (3.420 ～ 4.024)<0.0012.559 (2.347 ～ 2.790)<0.001
诊疗延误时长（d）
<302 69321 04411
≥303 91720 9781.459 (1.384 ～ 1.538)<0.0011.238 (1.158 ～ 1.324)<0.001
治疗方案
2HRZE/4HR6 55440 48011
2HRZE/7-10HRE341 1220.187 (0.133 ～ 0.264)<0.0010.392 (0.269 ～ 0.571)<0.001
其他224200.324 (0.211 ～ 0.497)<0.0010.572 (0.366 ～ 0.896)0.015
管理单位类型
医院2 70618 25311
疾病预防控制中心3 90423 7691.108 (1.051 ～ 1.168)<0.0011.144 (1.073 ～ 1.218)<0.001
现住地结核病负担
低5894 11211
中1 0527 2941.007 (0.904 ～ 1.122)0.9011.068 (0.944 ～ 1.210)0.296
高1 88915 2320.866 (0.784 ～ 0.956)0.0040.888 (0.796 ～ 0.992)0.035
严重3 08015 3841.398 (1.272 ～ 1.536)<0.0011.351 (1.207 ～ 1.511)<0.001
), ArticleFig(id=1241329755509805280, tenantId=1146029695717560320, journalId=1227665162245664772, articleId=1241329570801045540, language=CN, label=表3, caption=

复发影响因素的单因素、多因素logistic回归分析

, figureFileSmall=null, figureFileBig=null, tableContent=
特征复发未复发单因素logistic回归分析多因素logistic回归分析
OR（95%CI）P值aOR（95%CI）P值
性别
女性3 14220 98411
男性3 46821 0381.101 (1.045 ～ 1.160)<0.0011.175 (1.113 ～ 1.241)<0.001
年龄段（岁）
0 ～ <303064 20911
30 ～ <454004 6491.183 (1.014 ～ 1.381)0.0331.092 (0.928 ～ 1.284)0.288
45 ～ <601 4629 3362.154 (1.894 ～ 2.449)<0.0011.864 (1.623 ～ 2.140)<0.001
≥604 44223 8282.564 (2.273 ～ 2.892)<0.0012.200 (1.930 ～ 2.508)<0.001
职业
学生和教师511 21211
居家人员5313 8293.296 (2.455 ～ 4.424)<0.0011.712 (1.251 ～ 2.343)0.001
农民5 95436 0643.923 (2.961 ～ 5.199)<0.0012.227 (1.651 ～ 3.004)<0.001
其他749171.918 (1.329 ～ 2.768)0.0011.255 (0.857 ～ 1.838)0.243
就诊方式
被动方式4 67821 08911
主动方式1 93220 9330.416 (0.393 ～ 0.440)<0.0010.515 (0.484 ～ 0.548)<0.001
是否单纯结核性胸膜炎
否6 59741 39811
是136240.131 (0.075 ～ 0.227)<0.0010.333 (0.182 ～ 0.611)<0.001
治疗结局
完成疗程5 49132 13911
治愈1 1199 8830.663 (0.619 ～ 0.709)<0.0010.543 (0.483 ～ 0.612)<0.001
是否使用FDC
否6515 23811
是5 95936 7841.303 (1.196 ～ 1.420)<0.0011.190 (1.072 ～ 1.321)0.001
初诊痰检结果
阴性5 67737 66811
阳性9334 3541.422 (1.318 ～ 1.534)<0.0012.523 (2.240 ～ 2.843)<0.001
初治两个月末痰检结果
阴性6 52341 72311
阳性872991.861 (1.464 ～ 2.366)<0.0012.027 (1.567 ～ 2.622)<0.001
初诊痰培养结果
阴性71112 06711
阳性6145 9001.844 (1.650 ～ 2.061)<0.0011.728 (1.520 ～ 1.965)<0.001
无结果5 25824 0553.710 (3.420 ～ 4.024)<0.0012.559 (2.347 ～ 2.790)<0.001
诊疗延误时长（d）
<302 69321 04411
≥303 91720 9781.459 (1.384 ～ 1.538)<0.0011.238 (1.158 ～ 1.324)<0.001
治疗方案
2HRZE/4HR6 55440 48011
2HRZE/7-10HRE341 1220.187 (0.133 ～ 0.264)<0.0010.392 (0.269 ～ 0.571)<0.001
其他224200.324 (0.211 ～ 0.497)<0.0010.572 (0.366 ～ 0.896)0.015
管理单位类型
医院2 70618 25311
疾病预防控制中心3 90423 7691.108 (1.051 ～ 1.168)<0.0011.144 (1.073 ～ 1.218)<0.001
现住地结核病负担
低5894 11211
中1 0527 2941.007 (0.904 ～ 1.122)0.9011.068 (0.944 ～ 1.210)0.296
高1 88915 2320.866 (0.784 ～ 0.956)0.0040.888 (0.796 ～ 0.992)0.035
严重3 08015 3841.398 (1.272 ～ 1.536)<0.0011.351 (1.207 ～ 1.511)<0.001
)], attaches=null, journal=Journal(id=1227664546253402114, delFlag=0, nameCn=现代预防医学, nameEn=Modern Preventive Medicine, nameHistory1=null, nameHistory2=null, issn=1003-8507, eissn=null, cn=51-1365/R, coden=null, periodic=3, language=CN, oaType=null, ccby=null, superviseOffice=null, ownerOffice=null, pubOffice=null, editorOffice=null, officeType=null, aims=null, clcCode=null, officeProv=null, officeCity=null, officeAddr=null, officeZip=null, officeEmail=null, officePhone=null, editDirector=null, officeDirector=null, officeDirectorPhone=null, officeStaffNum=null, officeEmpNum=null, coverPicUrl=IeiuPXEZi6AA+k0VfvoiOQ==, journalPrice=null, startedYear=null, abbrevIsoEn=Modern Preventive Medicine, journalRemark=null, publicationField=null, createdTime=1770627636734, updatedTime=1770628902248, createdBy=18614031015, updatedBy=13701087609, firstLetterCn=M, firstLetterEn=M, subjectCode=Life Sciences, subjectName=null, subjectCodeEn=Life Sciences, subjectNameEn=null, picCn=IeiuPXEZi6AA+k0VfvoiOQ==, picEn=/9iTl8/ndms4tBz1fL28Pg==, jcr=null, cjcr=null, exts=[JournalExt(id=1227669854342280188, language=CN, name=现代预防医学, nameHistory1=null, nameHistory2=null, managedBy=, sponsoredBy=, publishedBy=, editorOffice=, officeProv=null, officeCity=null, officeAddr=, officeZip=, editDirector=, officeDirector=null, officePhone=null, coverPicUrl=null, journalRemark=, submitArticleUrl=null, websiteUrl=, createdTime=1770628902278, updatedTime=1770628902278, createdBy=13701087609, updatedBy=13701087609, submissionGuidelinesUrl=, submissionAuthorUrl=http://xdyfyxzz.paperopen.com/#/regist, submissionEditorUrl=http://xdyfyxzz.paperopen.com/#/Login, submissionReviewUrl=http://xdyfyxzz.paperopen.com/#/Login, submissionCeEditorUrl=, submissionAeEditorUrl=, option={"copyright":""}), JournalExt(id=1227669854396806141, language=EN, name=Modern Preventive Medicine, nameHistory1=null, nameHistory2=null, managedBy=, sponsoredBy=, publishedBy=, editorOffice=, officeProv=null, officeCity=null, officeAddr=, officeZip=, editDirector=, officeDirector=null, officePhone=null, coverPicUrl=null, journalRemark=, submitArticleUrl=null, websiteUrl=, createdTime=1770628902291, updatedTime=1770628902291, createdBy=13701087609, updatedBy=13701087609, submissionGuidelinesUrl=, submissionAuthorUrl=http://xdyfyxzz.paperopen.com/#/regist, submissionEditorUrl=http://xdyfyxzz.paperopen.com/#/Login, submissionReviewUrl=http://xdyfyxzz.paperopen.com/#/Login, submissionCeEditorUrl=, submissionAeEditorUrl=, option={"copyright":""})], databaseList=null, tenantJournalId=1227665162245664772, websiteList=[Website(id=1227687234141352800, webName=null, webTitle=null, webDomain=null, webCopyrigh=null, webIpcNo=null, seoTitle=null, seoKeywords=null, seoDescription=null, tenantJournalId=null, journalId=1227665162245664772, journalNameCn=null, journalNameEn=null, grayFlag=null, tenantId=1146029695717560320, platformId=null, journalGroupId=null, journalGroupNameCn=null, journalGroupNameEn=null, type=1, domain=https://castjournals.cast.org.cn/joweb/xdyfyx/CN, language=CN, createTime=1770633045945, createBy=18614031015, updateTime=1770633090526, updateBy=18614031015, name=现代预防医学-中文, tplId=1146099689490845704, title=现代预防医学, delFlag=0, indexPage=/home, props=[WebsiteProps(id=1227687735088051072, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1227687234141352800, code=articleTextType, value=kx, createTime=1770633165380, updateTime=1770633165380, creator=18614031015, updator=18614031015), WebsiteProps(id=1227687735071273853, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1227687234141352800, code=banner, value=null, createTime=1770633165376, updateTime=1770633165376, creator=18614031015, updator=18614031015), WebsiteProps(id=1227687735113216899, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1227687234141352800, code=grayFlag, value=0, createTime=1770633165386, updateTime=1770633165386, creator=18614031015, updator=18614031015), WebsiteProps(id=1227687735062885244, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1227687234141352800, code=logo, value=https://castjournals.cast.org.cn/joweb/xdyfyx/CN/file/pic?fileId=/XB5plC0xuykmQnycvtyrw==, createTime=1770633165374, updateTime=1770633165374, creator=18614031015, updator=18614031015), WebsiteProps(id=1227687735125799813, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1227687234141352800, code=minRunFlag, value=0, createTime=1770633165389, updateTime=1770633165389, creator=18614031015, updator=18614031015), WebsiteProps(id=1227687735083856767, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1227687234141352800, code=picServerUrl, value=https://castjournals.cast.org.cn/joweb/xdyfyx/CN/file/pic, createTime=1770633165379, updateTime=1770633165379, creator=18614031015, updator=18614031015), WebsiteProps(id=1227687735121605508, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1227687234141352800, code=silenceFlag, value=0, createTime=1770633165388, updateTime=1770633165388, creator=18614031015, updator=18614031015), WebsiteProps(id=1227687735079662462, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1227687234141352800, code=staticResourcePath, value=https://castjournals.cast.org.cn/joweb/cast_kjdb_cn_619/, createTime=1770633165378, updateTime=1770633165378, creator=18614031015, updator=18614031015), WebsiteProps(id=1227687735096439681, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1227687234141352800, code=themeColor, value=null, createTime=1770633165382, updateTime=1770633165382, creator=18614031015, updator=18614031015), WebsiteProps(id=1227687735104828290, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1227687234141352800, code=themeStyle, value=null, createTime=1770633165384, updateTime=1770633165384, creator=18614031015, updator=18614031015)]), Website(id=1227687234338485094, webName=null, webTitle=null, webDomain=null, webCopyrigh=null, webIpcNo=null, seoTitle=null, seoKeywords=null, seoDescription=null, tenantJournalId=null, journalId=1227665162245664772, journalNameCn=null, journalNameEn=null, grayFlag=null, tenantId=1146029695717560320, platformId=null, journalGroupId=null, journalGroupNameCn=null, journalGroupNameEn=null, type=1, domain=https://castjournals.cast.org.cn/joweb/xdyfyx/EN, language=EN, createTime=1770633045992, createBy=18614031015, updateTime=1770633115374, updateBy=18614031015, name=现代预防医学-英文, tplId=1146101810881728533, title=Modern Preventive Medicine, delFlag=0, indexPage=/home, props=[WebsiteProps(id=1227687709129507332, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1227687234338485094, code=articleTextType, value=kx, createTime=1770633159191, updateTime=1770633159191, creator=18614031015, updator=18614031015), WebsiteProps(id=1227687709108535809, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1227687234338485094, code=banner, value=null, createTime=1770633159186, updateTime=1770633159186, creator=18614031015, updator=18614031015), WebsiteProps(id=1227687709167256071, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1227687234338485094, code=grayFlag, value=0, createTime=1770633159200, updateTime=1770633159200, creator=18614031015, updator=18614031015), WebsiteProps(id=1227687709095952896, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1227687234338485094, code=logo, value=https://castjournals.cast.org.cn/joweb/xdyfyx/EN/file/pic?fileId=/XB5plC0xuykmQnycvtyrw==, createTime=1770633159183, updateTime=1770633159183, creator=18614031015, updator=18614031015), WebsiteProps(id=1227687709179838985, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1227687234338485094, code=minRunFlag, value=0, createTime=1770633159203, updateTime=1770633159203, creator=18614031015, updator=18614031015), WebsiteProps(id=1227687709121118723, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1227687234338485094, code=picServerUrl, value=https://castjournals.cast.org.cn/joweb/xdyfyx/EN/file/pic, createTime=1770633159189, updateTime=1770633159189, creator=18614031015, updator=18614031015), WebsiteProps(id=1227687709171450376, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1227687234338485094, code=silenceFlag, value=0, createTime=1770633159201, updateTime=1770633159201, creator=18614031015, updator=18614031015), WebsiteProps(id=1227687709112730114, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1227687234338485094, code=staticResourcePath, value=https://castjournals.cast.org.cn/joweb/cast_kjdb_en_623/, createTime=1770633159187, updateTime=1770633159187, creator=18614031015, updator=18614031015), WebsiteProps(id=1227687709133701637, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1227687234338485094, code=themeColor, value=null, createTime=1770633159192, updateTime=1770633159192, creator=18614031015, updator=18614031015), WebsiteProps(id=1227687709154673158, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1227687234338485094, code=themeStyle, value=null, createTime=1770633159197, updateTime=1770633159197, creator=18614031015, updator=18614031015)])], journalTitle=现代预防医学, weixinUrl=null, journalUrl=http://xdyfyxzz.paperopen.com/, iacademicId=null, status=1, seqNo=null, journalTitleEn=Modern Preventive Medicine, journalPhotoCn=IeiuPXEZi6AA+k0VfvoiOQ==, journalPhotoEn=/9iTl8/ndms4tBz1fL28Pg==, journalFirstLetter=M, journalRecommend=null, journalNew=null, journalCollection=null, jcrJf=null, cjcrJf=null, jcrJfStr=null, cjcrJfStr=null, submissionFirstDecision=null, sciSubjectClassification=null, casSubjectClassification=null, citeScore=null, totalCitationFrequency=null, icpCode=null, psCode=null, advertisingLicenseCode=null, copyrightInformation=null, country=null, option=, provinceCode=null, provinceName=null, collectFlag=false), detailUrlCn=https://castjournals.cast.org.cn/joweb/xdyfyx/CN/10.20043/j.cnki.MPM.202503139, detailUrlEn=https://castjournals.cast.org.cn/joweb/xdyfyx/EN/10.20043/j.cnki.MPM.202503139, pdfUrlCn=https://castjournals.cast.org.cn/joweb/xdyfyx/CN/PDF/10.20043/j.cnki.MPM.202503139, pdfUrlEn=https://castjournals.cast.org.cn/joweb/xdyfyx/EN/PDF/10.20043/j.cnki.MPM.202503139, aliStartDate=null, aliEndDate=null, collectionFlag=false, citedCount=null, citedUrl=null, reference=null)
收藏切换
基于机器学习算法的初治结核病患者复发预测模型研究
收藏切换
PDF下载
买日哈巴·卡米力 1 , 买吾拉江·依马木 2 , 王艳杰 1 , 王雨薇 1 , 阿丽米热·阿不力米提 1 , 麦迪努尔·卡米力 3 , 向阳 1
现代预防医学 | 流行病与统计方法 2025,52(13): 2310-2316
收起
收藏切换
现代预防医学 | 流行病与统计方法 2025, 52(13): 2310-2316
基于机器学习算法的初治结核病患者复发预测模型研究
全屏
买日哈巴·卡米力1, 买吾拉江·依马木2, 王艳杰1, 王雨薇1, 阿丽米热·阿不力米提1, 麦迪努尔·卡米力3, 向阳1
作者信息
  • 1.新疆医科大学公共卫生学院，新疆 乌鲁木齐 830017
  • 2.喀什地区疾病预防控制中心
  • 3.新疆医科大学临床医学部

    • 买日哈巴·卡米力（1998—），女，硕士在读，研究方向：传染病流行病学

通讯作者:

向阳，E-mail:
Prediction model for tuberculosis recurrence in newly treated patients based on machine learning algorithms
KAMILI Mai-ri-ha-ba1, YIMAMU Mai-wu-la-jiang2, Yan-jie WANG1, Yu-wei WANG1, ABULIMITI A-li-mi-re1, KAMILI Mai-di-nu-er3, Yang XIANG1
Affiliations
  • School of Public Health, Xinjiang Medical University, Urumqi, Xinjiang 830017, China
出版时间: 2025-07-10 doi: 10.20043/j.cnki.MPM.202503139
文章导航
收藏切换
摘要
收起
目的

基于新疆喀什地区结核病常规监测数据，系统比较7种机器学习算法构建的初治结核病患者复发预测模型的效能，为高负担地区优化复发干预策略提供数据支持。

方法

以喀什地区2016—2022年成功治疗的初治结核病患者为研究对象，分析其截至2023年末的复发情况。通过多因素logistic回归筛选独立预测因子，并基于logistic回归、决策树、随机森林、多层感知器、极限梯度提升树、轻量级梯度提升机算法和弹性网络构建并验证复发预测模型，对最优模型通过采用沙普利加性解释（SHapley additive exPlanations, SHAP）法解析变量贡献度。

结果

纳入的69 476例研究对象截至2023年末复发9 444例（13.59%）。通过logistic回归确定出14个独立预测因子，7种预测模型训练集AUC值范围为0.705～0.762，其中决策树模型表现最优（AUC=0.762, 95%CI: 0.758～0.766），且校准度良好。基于SHAP值的变量重要性排序结果显示，初诊痰培养结果、现住地结核病负担及就诊方式是前三位重要预测因子。

结论

基于常规监测数据构建的决策树模型对初治结核复发具有较高预测效能，其可解释性特征有助于临床优先识别高危个体。

结核病  /  复发  /  机器学习  /  治疗成功
Objective

To systematically compare the performance of seven machine learning algorithms in constructing prediction models for tuberculosis (TB) recurrence among newly treated patients in Kashgar, Xinjiang, providing data support for optimizing recurrence intervention strategies in high-burden areas.

Methods

We analyzed 69 476 successfully treated new TB patients from 2016 to 2022 in Kashgar, with follow-up through 2023. Independent predictors were selected through multivariate logistic regression. Seven models (logistic regression, decision tree, random forest, multilayer perceptron, XGBoost, LightGBM, and elastic net) were developed and validated. The optimal model was interpreted using SHapley Additive exPlanations (SHAP).

Results

Among 69 476 cases, 9 444 (13.59%) experienced recurrence by 2023. Fourteen independent predictors were identified. The seven models showed AUC values ranging from 0.705 to 0.762 in the training set, with the decision tree model performing best (AUC=0.762, 95%CI: 0.758-0.766) and demonstrating good calibration. SHAP analysis revealed sputum culture results at diagnosis, local TB burden, and treatment modality as the top three predictive factors.

Conclusion

The decision tree model based on routine surveillance data shows high predictive performance for TB recurrence, with interpretable features that can facilitate early identification of high-risk individuals in clinical practice.

Tuberculosis  /  Recurrence  /  Machine learning  /  Treatment success
买日哈巴·卡米力, 买吾拉江·依马木, 王艳杰, 王雨薇, 阿丽米热·阿不力米提, 麦迪努尔·卡米力, 向阳. 基于机器学习算法的初治结核病患者复发预测模型研究. 现代预防医学, 2025 , 52 (13) : 2310 -2316 . DOI: 10.20043/j.cnki.MPM.202503139
KAMILI Mai-ri-ha-ba, YIMAMU Mai-wu-la-jiang, Yan-jie WANG, Yu-wei WANG, ABULIMITI A-li-mi-re, KAMILI Mai-di-nu-er, Yang XIANG. Prediction model for tuberculosis recurrence in newly treated patients based on machine learning algorithms[J]. Modern Preventive Medicine, 2025 , 52 (13) : 2310 -2316 . DOI: 10.20043/j.cnki.MPM.202503139
正文
收起
结核病复发是阻碍全球终止结核病进程的重要挑战，高负担地区尤为显著。多项研究报告显示，初治结核病患者中2%～10%在治疗成功后约五年内出现复发，在新疆喀什地区这类高负担地区该比例可高达约15%[1]。复发患者耐药率增高、病情变复杂，易造成疾病的迁延难愈和进一步扩散[2]。精准预测复发、优化干预策略刻不容缓。在大数据与人工智能技术快速发展的当下，机器学习算法为结核病复发预测提供了有效路径。本研究基于喀什地区近七年结核病监测数据，系统比较七种常用机器学习算法预测复发效能，通过可解释性分析揭示关键风险因子，旨在构建适配资源受限场景的复发预测工具，为优化高危人群靶向干预提供依据。
1　资料与方法
收起
1.1　资料来源
本研究数据从《结核病管理信息系统》系统中导出，是经匿名化处理的常规监测数据，不涉及伦理审查和知情同意。
1.2　研究对象
喀什地区2016—2022年治疗成功的初治结核病患者（排除信息错误或缺失、现住地非喀什地区的患者）。
1.3　相关定义
“复发”是指曾接受过抗结核治疗，且在疗程结束后被判定为“成功治疗”（包括“治愈”或“完成疗程”），但后期又被重新登记为结核病[3]。本研究为了排除治疗不成功患者被误判的可能性，将“成功治疗”后六个月内就出现再次复发的患者视为治疗不成功患者[4]。“诊疗延误时长”是指从患者出现结核病相关症状或体征开始，到最终获得适当治疗之间所经历的时间间隔[1]。现住地结核病负担以2016—2022年喀什地区各县市结核病年均发病率划分：<200/10万为“低”；200/10万～300/10万为“中”；300/10万～400/10万为“较高”；>400/10万为“严重”。治疗方案中“2HRZE”表示强化期药物组合，包括异烟肼、利福平、吡嗪酰胺和乙胺丁醇，每日一次，连续使用2个月，“4HR”表示巩固期药物组合，包括异烟肼和利福平，每日一次，连续使用4个月，“7-10HRE”表示巩固期药物组合，包括异烟肼、利福平和乙胺丁醇，每日一次，连续使用7～10个月。
1.4　预测模型
1.4.1　预测模型的构建
将研究对象按7∶3随机分为训练集和验证集。在训练集中，以“是否复发”为因变量，纳入病案信息中性别、年龄等14个因素为自变量，采用单因素、多因素logistic回归筛选复发的独立影响因素，以此作为预测因子，构建基于logistic回归、决策树（decision tree）、随机森林（random forest）、多层感知器（multilayer perceptron, MLP）、极限梯度提升树（eXtreme gradient boosting, XGboost）、轻量级梯度提升机算法（light gradient boosting machine, light GBM）、弹性网络（elastic net, Enet）7种机器学习算法的复发预测模型。随机过采样实例法（random oversampling examples, ROSE）用于使训练集中复发患者与未复发患者比例接近1：1平衡，以提高模型预测效能，十折交叉验证法(cross-validation，CV)和网格搜索法用于寻找模型的最优超参数。
1.4.2　预测模型的验证与评价
以训练集上预测效果最佳的超参数作为模型的初始化超参数对验证集进行验证。受试者工作特征（receiver operating characteristic，ROC）曲线下面积（area under the curve，AUC）、校准曲线用于评价模型的区分度、准确性。
1.4.3　SHAP模型解释
对区分度最优模型采用SHAP分析法进行模型解释，绘制SHAP特征重要性排序图、SHAP蜂群图及SHAP值特征依赖图，深入理解模型的决策过程和各特征对预测结果的具体贡献。
1.5　统计学方法
分别采用Excel 2016和R 4.4.1进行数据整理与分析。通过计算自变量间的容差和方差膨胀因子判断变量共线性。定性资料采用频数和率进行统计描述，采用χ2检验分析组间差异性。使用sample函数对数据集进行随机划分。Rose包中的ROSE函数用于数据不平衡处理。采用logistic回归对各变量进行单因素分析，对其P值小于0.05的变量纳入多因素logistic回归分析中，以P值小于0.05的变量作为预测因子纳入到模型构建过程。Tidymodels包用于模型构建和模型评估，fastshap包用于SHAP模型解释，iml包用于H统计量交互作用强度分析。双侧检验，检验水准α=0.05。
2　结果
收起
2.1　复发概况
2016—2022年治疗成功的69 476例初治结核病患者中，截至2023年末复发9 444例（13.59%）。男性复发率为14.09%，高于女性的13.09%；农民复发率为14.20%，高于非农民的10.74%；≥60岁者复发率为18.61%，高于其他年龄段者的11.93%。
2.2　预测因子筛选
以研究对象性别、年龄段、职业、就诊方式、是否单纯结核性胸膜炎、治疗结局、初诊痰检结果、初诊痰培养结果、初治两个月末痰检结果、诊疗延误时长、治疗方案、固定剂量复合制剂（fixed-dose combination, FDC）使用情况、管理单位类型、现住地结核病负担等14个因素作为研究变量，各变量间容差均小于0.2、方差膨胀因子均大于5，不存在显著共线性。将研究对象按7：3随机划分为训练集（48 632例）和验证集（20 844例），两组患者各研究变量间差异均无统计学意义（P值均>0.05），见表1。变量赋值情况见表2。在训练集，对单因素logistic回归分析有统计学意义的变量进一步纳入多因素logistic回归分析，多因素分析结果显示，14个研究变量均为复发的独立影响因素。见表3。
2.3　模型构建与验证
在训练集中，基于多因素logistic回归分析筛选出的14个预测因子，构建logistic、决策树、随机森林、MLP、XGBoost、Light GBM和ENet等7种复发预测模型，并在验证集中进行内部验证。结果显示，7种模型在训练集中的AUC值范围为0.705～0.762，在验证集中的AUC值范围为0.691～0.712，模型均具有较好的区分度，其中基于决策树的复发预测模型区分度最佳（AUC=0.762, 95%CI:0.758～0.766），见图1。校准曲线图表明，logistic、决策树、ENet、Light GBM和XGBoost模型的校准曲线在训练集和验证集中均接近理想对角线，显示出良好的准确性和可靠性，见图2。
2.4　SHAP模型解释
对AUC值最高的决策树模型进行SHAP解释。SHAP特征重要性排序结果显示，初诊痰培养结果、现住地结核病负担和就诊方式是初治结核病复发的前三个重要预测因子，对模型输出影响显著，见图3A。这些变量与其他变量间交互作用均较弱。SHAP蜂群图和特征依赖图直观展示了各特征值与模型预测的关系，揭示了特征对模型输出的具体贡献，见图3B、C。在14个预测因子中，初诊痰培养结果未知、现住地结核病负担严重、年龄≥60岁、管理单位为疾控中心、诊疗延误时长≥30 d、初诊痰检阳性、男性、初治两个月末痰检阳性等因素对模型有正向影响，增加复发风险；而主动就诊、治疗结局为治愈、治疗方案为2HRZE/7-10HRE或其他方式等因素对模型有负向影响，是复发的保护因素。
3　讨论
收起
结核病复发问题是我国结核病防控工作中亟待解决的重要难题，实现复发精准预测并优化干预策略具有重要的现实意义。随着大数据和人工智能技术的迅猛发展，各类机器学习算法已被广泛应用于结核病各个领域，包括结核分枝杆菌潜伏感染与活动性结核病的鉴别诊断[5]、耐药结核病的诊断与预测[6]、抗结核治疗不良反应预测[7]、结核分枝杆菌和非结核分枝杆菌感染的分类[8]以及结核病治疗转归结果的预测等方面[9]。这些研究通过整合机器学习技术与医学影像、生物标志物及临床数据，显著提升了预测准确性和效率。然而，几乎很少有研究将机器学习算法运用到结核病复发领域。本研究利用喀什地区常规监测数据，构建了基于logistic、决策树、随机森林、MLP、XGboost、Light GBM、ENet等7种机器学习算法的初治结核病患者复发预测模型。
本研究结果显示，截至2023年末，2016—2022年喀什地区成功治疗的初治结核病患者中复发者占13.59%，高于我国其他地区[10-11]。复发率男性高于女性，≥60岁者高于其他年龄段者，农牧民高于其他职业者。该特征与多个地区类似，如韩国一项回顾性国家队列研究发现，糖尿病合并结核病患者中男性、中老年者是结核复发的独立预测因素[12]；中国衡阳市一项研究发现，男性肺结核患者复发风险是女性的1.592倍，≥60岁者复发风险是其他年龄段者的8.175倍，农民复发风险是其他职业者的1.379倍[13]。男性复发率较高可能与职业暴露、吸烟及饮酒等行为因素相关，这些因素可削弱免疫功能并增加再感染风险；而≥60岁人群的高复发率可能与年龄相关的免疫衰老及合并症（如糖尿病）有关，导致潜伏感染的再激活风险显著升高;农牧民的高复发率则可能与其职业特性相关，包括医疗资源可及性低、劳动强度大导致的营养状况较差，以及居住地结核病传播压力较高等综合因素[14]。
本研究构建的7种机器学习模型均具有较好的预测效能，其中决策树模型区分度最佳，校准度良好。决策树模型在处理复杂的监测数据时具有较高的适应性和解释性，对于临床医生制定个性化治疗方案具有重要意义。对决策树模型进行SHAP解释，SHAP特征重要性排序结果显示，初诊痰培养结果、现住地结核病负担和就诊方式是初治结核病患者复发的三大重要预测因子。痰培养是确诊肺结核的金标准之一，痰培养阳性患者复发风险升高，可能与初始菌量高、病情严重（如肺部空洞）及药物难以彻底清除病灶细菌相关；而痰培养结果未知的患者复发风险也显著升高，可能因缺乏病原学证据导致治疗方案不充分，尤其潜在耐药菌株未被识别和治疗，增加了复发风险。结核病负担严重地区通常伴随着广泛的结核分枝杆菌传播，提高治疗成功患者外源性再感染风险。同时，结核病高负担地区可能存在医疗资源不足、治疗不规范及耐药菌株流行等状况，进一步加剧复发风险。就诊方式主要分为健康体检与主动筛查等主动来源方式和转诊、追踪、直接就诊与推介等被动来源方式。主动方式通常由患者或医疗机构主动发起，患者在无明显症状时通过定期体检或筛查被发现。一项系统综述表明，主动来源是患者成功治疗后复发的保护因素[1]。主动来源的患者通常在在疾病早期病情较轻时被发现，能够及时开始规范的抗结核治疗，减少了治疗延误和不规律服药的情况，这可能显著降低了复发风险。此外，管理单位为疾控中心、诊疗延误时长≥30 d、初诊痰检阳性、初治两个月末痰检阳性、使用FDC均在一定程度上增加初治结核病患者复发风险，而治愈、治疗方案为2HRZE/7-10HRE及单纯结核性胸膜炎降低复发风险，这在不同研究中均被报道[15-16]。提示在结核病防控中，优化治疗管理、缩短诊疗延误、规范治疗方案以及加强高危人群的随访管理，是降低复发率的重要策略。
本研究有以下局限性。数据来源于《结核病管理信息系统》，可能因录入质量问题低估复发率；仅分析了病案中已登记且缺失较少的变量，缺乏对合并症（如HIV感染、糖尿病）及营养状况等潜在重要因素的探索；数据来自特定地区，可能存在地域偏差，模型虽通过内部验证，但未进行外部验证，泛化能力尚不明确；模型实际临床适用性仍缺乏充分验证。
综上，本研究利用喀什地区结核病常规监测数据，构建了基于7种机器学习算法的初治结核病患者复发预测模型，其中基于决策树的模型具有较好的预测效能，能够为结核病复发的个体化防控提供科学依据。此外，SHAP模型解释分析表明，初诊痰培养结果、现住地结核病负担和就诊方式是复发的重要预测因子。
基金
收起
  • 国家自然科学基金(81860589)
  • 自治区自然科学基金(2022 D01C203)
  • 新疆维吾尔自治区”十四五“高等学校特色学科-公共卫生与预防医学
文献
收起
参考文献 引证文献
排序方式：
［1］
李涛， 赵雁林， 贾忠伟． 结核病复发及其影响因素研究进展［J］.疾病监测，2024，39（4）：403-411.
Li T， Zhao YL, Jia ZW. Progress in research of tuberculosis recurrence and influencing factors［J］. Disease Surveillance. 2024, 39（4）: 403-411.（In Chinese）
［2］
Teferi MY. El-Khatib Z, Boltena MT .et al. Tuberculosis treatment outcome and predictors in Africa: a systematic review and Meta-Analysis ［J］. International Journal of Environmental Research and Public Health, 2021，18（20）: 10678.
［3］
钟倩红．马晓慧，钟永辉，等.2014-2018年佛山市肺结核患者复发情况及影响因素分析［J］. 中国防痨杂志，2022,44（7）:690-697.
Zhong QH, Ma XH, Zhong YH, et al. Analysis of recurrence rate of pulmonary tuberculosis patients in Foshan and the influencing factors from 2014 to 2018 ［J］. Chinese Journal of Antituberculosis,2022, 44（7）: 690-697.（In Chinese）
［4］
Hung CL, Chien JY, Ou CY. Associated factors for tuberculosis recurrence in Taiwan: a nationwide nested case-control study from 1998 to 2010［J］. PLOS One. 2015, 10（5）: e0124822.
［5］
Luo Y, Xue Y, Song HJ, et al. Machine learning based on routine lahoratory indicators promoting the discrimination between active tuberculosis and latent tuberculosis infection［J］. Journal of Infection,2022, 84（5）: 648-657.
［6］
Falcao I, Cardoso DL, Coutinho DSA, et al. Model for predicting drug resistance based on the clinical profile of tuberculosis patients using machine learning techniques ［J］. PeerJ Computer Science, 2024, 10:e2246.
［7］
Kong HB, Li Y, Shen Y， et al. Predicting the risk of pulmonary embolism in patients with tuberculosis using machine learning algorithms ［J］. European Journal of Medical Research, 2024, 29（1）:618.
［8］
Anh NK. Phat NK, Thu NQ, et al. Discovery of urinary biosignatures for tuberculosis and nontuberculous mycobacteria classification using metabolomics and machine learning ［J］. Scientific Reports, 2024, 14（1）: 15312.
［9］
Chen JF, Jiang YL, Li ZH, et al. Predictive machine learning models for anticipating loss to follow-up in tuberculosis patients throughout anti-TB treatment journey［J］. Scientific Reports, 2024, 14（1）: 24685.
［10］
彭红， 虞浩， 姜洁， 等．江苏省初治结核病复发流行病学特征及影响因素［J］.江苏预防医学，2019，30（4）：355-359.
Peng H, Yu H, Jiang J, et al. Epidemiological characteristics and associated factors with tuberculosis recurrence in Jiangsu province［J］.Jiangsu Journal of Preventive Medicine, 2019, 30 （4）: 355-359.（In Chinese）
［11］
Ruan QL, Yang QL, Sun F, et al. Recurrent pulmonary tuberculosis after treatment success: a population-based retrospective study in China［J］. Clinical Microbiology and Infection, 2022, 28（5）: 684-689.
［12］
Eksombatchai D, Jeong D, Mok J, et al. Sex differences in the impact of diabetes mellitus on tuberculosis recurrence: a retrospective National cohort study［J］. International Journal of Infectious Diseases,2023, 127: 1-10.
［13］
周旭， 王巧智， 万燕萍， 等．衡阳市成功治疗的肺结核患者5年复发影响因素分析［J］.预防医学，2022,34（3）:268-271.
Zhou X, Wang QZ, Wan YP, et al. Factors affecting the five-year recurrence among successfully-treated pulmonary tuberculosis patients in Hengyang City ［J］. Journal of Preventive Medicine, 2022,34（3）: 268-271.（In Chinese）
［14］
Mujtaba MA, Richardson M, Shahzad H, et al. Demographic and clinical determinants of tuberculosis and TB recurrence: a Double-Edged retrospective study from Pakistan ［J］. Journal of Tropical Medicine, 2022, 2022（1）: 4408306.
［15］
Vega V, Cabrera-Sanchez J, Rodríguez S, et al. Risk factors for pulmonary tuberculosis recurrence, relapse and reinfection: a systematic review and meta-analysis ［J］. BMJ Open Respir Res,2024, 11（1）: e002281.
［16］
Qiu BB, Wu ZC, Tao BL, et al. Risk factors for types of recurrent tuberculosis （reactivation versus reinfection）: A global systematic review and meta-analysis ［J］. International Journal of Infectious Diseases, 2022, 116: 14-20.
2025年第52卷第13期
PDF下载
47
21
引用本文
BibTeX
文章信息
doi: 10.20043/j.cnki.MPM.202503139
  • 接收时间：2025-03-12
  • 首发时间：2026-03-19
  • 出版时间：2025-07-10
补充材料
相关文章
文章信息
作者
出版历史
  • 收稿日期：2025-03-12
基金
国家自然科学基金(81860589)
自治区自然科学基金(2022 D01C203)
新疆维吾尔自治区”十四五“高等学校特色学科-公共卫生与预防医学
作者信息
    1.新疆医科大学公共卫生学院，新疆 乌鲁木齐 830017
    2.喀什地区疾病预防控制中心
    3.新疆医科大学临床医学部

通讯作者:

向阳，E-mail:
参考文献
分享链接
https://castjournals.cast.org.cn/joweb/xdyfyx/CN/10.20043/j.cnki.MPM.202503139
分享至
全文二维码

扫描看全文

引用本文
BibTeX
本文的引用情况
2种不同金属材料的力学参数
科
Family		 属数
Number of
genus		 种数
Number of
species		 占总种数比例
Percentage of
total species (%)		 属
Genus		 种数
Number of
species		 占总种数比例
Percentage of total
species (%)
鹅膏菌科Amanitaceae 2 11 5.26 鹅膏菌属 Amanita 10 4.78
小菇科 Mycenaceae 2 12 5.74 丝盖伞属 Inocybe 5 2.39
多孔菌科 Polyporaceae 8 14 6.70 蜡蘑属 Laccaria 5 2.39
红菇科 Russulaceae 3 23 11.00 小皮伞属 Marasmius 6 2.87
小菇属 Mycena 11 5.26
光柄菇属 Pluteus 5 2.39
红菇属 Russula 17 8.13
栓菌属 Trametes 5 2.39
关闭全屏
相关链接
论文
最新文章
热读文章
热点专题
内参
蓝皮书
产业发展报告
传播力报告
期刊分级目录
资讯
学术新闻
行业新闻
学术会议
行业会议
关于
关于我们
联系我们
北京市海淀区学院南路86号
100081
010-62199257
qkjq@cast.org.cn

中国科学技术协会版权所有 京ICP 备10216604号-4 海淀分局备案 1101084647
科技导报社主办 中国科协信息中心技术支持