Machine learning methods are comparable to logistic regression techniques in predicting severe walking limitation following total knee arthroplasty

Purpose Machine-learning methods are flexible prediction algorithms with potential advantages over conventional regression. This study aimed to use machine learning methods to predict post-total knee arthroplasty (TKA) walking limitation, and to compare their performance with that of logistic regres...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Knee surgery, sports traumatology, arthroscopy : official journal of the ESSKA sports traumatology, arthroscopy : official journal of the ESSKA, 2020-10, Vol.28 (10), p.3207-3216
Hauptverfasser: Pua, Yong-Hao, Kang, Hakmook, Thumboo, Julian, Clark, Ross Allan, Chew, Eleanor Shu-Xian, Poon, Cheryl Lian-Li, Chong, Hwei-Chi, Yeo, Seng-Jin
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page 3216
container_issue 10
container_start_page 3207
container_title Knee surgery, sports traumatology, arthroscopy : official journal of the ESSKA
container_volume 28
creator Pua, Yong-Hao
Kang, Hakmook
Thumboo, Julian
Clark, Ross Allan
Chew, Eleanor Shu-Xian
Poon, Cheryl Lian-Li
Chong, Hwei-Chi
Yeo, Seng-Jin
description Purpose Machine-learning methods are flexible prediction algorithms with potential advantages over conventional regression. This study aimed to use machine learning methods to predict post-total knee arthroplasty (TKA) walking limitation, and to compare their performance with that of logistic regression. Methods From the department’s clinical registry, a cohort of 4026 patients who underwent elective, primary TKA between July 2013 and July 2017 was identified. Candidate predictors included demographics and preoperative clinical, psychosocial, and outcome measures. The primary outcome was severe walking limitation at 6 months post-TKA, defined as a maximum walk time ≤ 15 min. Eight common regression (logistic, penalized logistic, and ordinal logistic with natural splines) and ensemble machine learning (random forest, extreme gradient boosting, and SuperLearner) methods were implemented to predict the probability of severe walking limitation. Models were compared on discrimination and calibration metrics. Results At 6 months post-TKA, 13% of patients had severe walking limitation. Machine learning and logistic regression models performed moderately [mean area under the ROC curves (AUC) 0.73–0.75]. Overall, the ordinal logistic regression model performed best while the SuperLearner performed best among machine learning methods, with negligible differences between them (Brier score difference,
doi_str_mv 10.1007/s00167-019-05822-7
format Article
fullrecord <record><control><sourceid>proquest_cross</sourceid><recordid>TN_cdi_proquest_miscellaneous_2337000472</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2337000472</sourcerecordid><originalsourceid>FETCH-LOGICAL-c441t-dde686f6d1c8ff7ef5797a2b9504e6883e33936e447f098dbd4f62afd8f340523</originalsourceid><addsrcrecordid>eNp9Uc2OFCEQJkbjjqMv4MGQePHSyl833UezUddkjRc9EwaKGXZpaIHZzT6HLyy9s2riwRMF9f1Q9SH0kpK3lBD5rhBCB9kROnWkHxnr5CO0oYLzTnIhH6MNmQTrGOmHM_SslCtCWimmp-iM05GzYZIb9POLNgcfAQfQOfq4xzPUQ7IF6wzYpHnRWe8C4JpwSHtfqjc4wz5DKT5FXMEcov9xhIJ9xEsG601dZQrcQFO41eF6vQY_-6rrSnEphHS7PtZUdcDXEaC51UNOS9Cl3j1HT5wOBV48nFv0_eOHb-cX3eXXT5_P3192RghaO2thGAc3WGpG5yS4Xk5Ss93UE9E6IwfOJz6AENKRabQ7K9zAtLOj44L0jG_Rm5PuktM6QVWzLwZC0BHSsSjGuWw7E3KFvv4HepWOObbfKTayvm1z9doidkKZnErJ4NSS_azznaJErZGpU2SqRabuI1OykV49SB93M9g_lN8ZNQA_AUprxT3kv97_kf0Fp5OlQw</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>2825832339</pqid></control><display><type>article</type><title>Machine learning methods are comparable to logistic regression techniques in predicting severe walking limitation following total knee arthroplasty</title><source>MEDLINE</source><source>Access via Wiley Online Library</source><source>SpringerLink Journals (MCLS)</source><creator>Pua, Yong-Hao ; Kang, Hakmook ; Thumboo, Julian ; Clark, Ross Allan ; Chew, Eleanor Shu-Xian ; Poon, Cheryl Lian-Li ; Chong, Hwei-Chi ; Yeo, Seng-Jin</creator><creatorcontrib>Pua, Yong-Hao ; Kang, Hakmook ; Thumboo, Julian ; Clark, Ross Allan ; Chew, Eleanor Shu-Xian ; Poon, Cheryl Lian-Li ; Chong, Hwei-Chi ; Yeo, Seng-Jin</creatorcontrib><description>Purpose Machine-learning methods are flexible prediction algorithms with potential advantages over conventional regression. This study aimed to use machine learning methods to predict post-total knee arthroplasty (TKA) walking limitation, and to compare their performance with that of logistic regression. Methods From the department’s clinical registry, a cohort of 4026 patients who underwent elective, primary TKA between July 2013 and July 2017 was identified. Candidate predictors included demographics and preoperative clinical, psychosocial, and outcome measures. The primary outcome was severe walking limitation at 6 months post-TKA, defined as a maximum walk time ≤ 15 min. Eight common regression (logistic, penalized logistic, and ordinal logistic with natural splines) and ensemble machine learning (random forest, extreme gradient boosting, and SuperLearner) methods were implemented to predict the probability of severe walking limitation. Models were compared on discrimination and calibration metrics. Results At 6 months post-TKA, 13% of patients had severe walking limitation. Machine learning and logistic regression models performed moderately [mean area under the ROC curves (AUC) 0.73–0.75]. Overall, the ordinal logistic regression model performed best while the SuperLearner performed best among machine learning methods, with negligible differences between them (Brier score difference, &lt; 0.001; 95% CI [− 0.0025, 0.002]). Conclusions When predicting post-TKA physical function, several machine learning methods did not outperform logistic regression—in particular, ordinal logistic regression that does not assume linearity in its predictors. Level of evidence Prognostic level II</description><identifier>ISSN: 0942-2056</identifier><identifier>EISSN: 1433-7347</identifier><identifier>DOI: 10.1007/s00167-019-05822-7</identifier><identifier>PMID: 31832697</identifier><language>eng</language><publisher>Berlin/Heidelberg: Springer Berlin Heidelberg</publisher><subject>Aged ; Algorithms ; Arthroplasty (knee) ; Arthroplasty, Replacement, Knee - adverse effects ; Calibration ; Female ; Humans ; Joint replacement surgery ; Knee ; Learning algorithms ; Logistic Models ; Machine Learning ; Male ; Medicine ; Medicine &amp; Public Health ; Middle Aged ; Mobility Limitation ; Orthopedics ; Postoperative Complications - physiopathology ; Prognosis ; Registries ; Regression analysis ; Regression models ; Spline functions ; Statistical analysis ; Treatment Outcome ; Walking</subject><ispartof>Knee surgery, sports traumatology, arthroscopy : official journal of the ESSKA, 2020-10, Vol.28 (10), p.3207-3216</ispartof><rights>European Society of Sports Traumatology, Knee Surgery, Arthroscopy (ESSKA) 2019</rights><rights>European Society of Sports Traumatology, Knee Surgery, Arthroscopy (ESSKA) 2019.</rights><lds50>peer_reviewed</lds50><woscitedreferencessubscribed>false</woscitedreferencessubscribed><citedby>FETCH-LOGICAL-c441t-dde686f6d1c8ff7ef5797a2b9504e6883e33936e447f098dbd4f62afd8f340523</citedby><cites>FETCH-LOGICAL-c441t-dde686f6d1c8ff7ef5797a2b9504e6883e33936e447f098dbd4f62afd8f340523</cites><orcidid>0000-0003-2313-9665</orcidid></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktopdf>$$Uhttps://link.springer.com/content/pdf/10.1007/s00167-019-05822-7$$EPDF$$P50$$Gspringer$$H</linktopdf><linktohtml>$$Uhttps://link.springer.com/10.1007/s00167-019-05822-7$$EHTML$$P50$$Gspringer$$H</linktohtml><link.rule.ids>314,780,784,27924,27925,41488,42557,51319</link.rule.ids><backlink>$$Uhttps://www.ncbi.nlm.nih.gov/pubmed/31832697$$D View this record in MEDLINE/PubMed$$Hfree_for_read</backlink></links><search><creatorcontrib>Pua, Yong-Hao</creatorcontrib><creatorcontrib>Kang, Hakmook</creatorcontrib><creatorcontrib>Thumboo, Julian</creatorcontrib><creatorcontrib>Clark, Ross Allan</creatorcontrib><creatorcontrib>Chew, Eleanor Shu-Xian</creatorcontrib><creatorcontrib>Poon, Cheryl Lian-Li</creatorcontrib><creatorcontrib>Chong, Hwei-Chi</creatorcontrib><creatorcontrib>Yeo, Seng-Jin</creatorcontrib><title>Machine learning methods are comparable to logistic regression techniques in predicting severe walking limitation following total knee arthroplasty</title><title>Knee surgery, sports traumatology, arthroscopy : official journal of the ESSKA</title><addtitle>Knee Surg Sports Traumatol Arthrosc</addtitle><addtitle>Knee Surg Sports Traumatol Arthrosc</addtitle><description>Purpose Machine-learning methods are flexible prediction algorithms with potential advantages over conventional regression. This study aimed to use machine learning methods to predict post-total knee arthroplasty (TKA) walking limitation, and to compare their performance with that of logistic regression. Methods From the department’s clinical registry, a cohort of 4026 patients who underwent elective, primary TKA between July 2013 and July 2017 was identified. Candidate predictors included demographics and preoperative clinical, psychosocial, and outcome measures. The primary outcome was severe walking limitation at 6 months post-TKA, defined as a maximum walk time ≤ 15 min. Eight common regression (logistic, penalized logistic, and ordinal logistic with natural splines) and ensemble machine learning (random forest, extreme gradient boosting, and SuperLearner) methods were implemented to predict the probability of severe walking limitation. Models were compared on discrimination and calibration metrics. Results At 6 months post-TKA, 13% of patients had severe walking limitation. Machine learning and logistic regression models performed moderately [mean area under the ROC curves (AUC) 0.73–0.75]. Overall, the ordinal logistic regression model performed best while the SuperLearner performed best among machine learning methods, with negligible differences between them (Brier score difference, &lt; 0.001; 95% CI [− 0.0025, 0.002]). Conclusions When predicting post-TKA physical function, several machine learning methods did not outperform logistic regression—in particular, ordinal logistic regression that does not assume linearity in its predictors. Level of evidence Prognostic level II</description><subject>Aged</subject><subject>Algorithms</subject><subject>Arthroplasty (knee)</subject><subject>Arthroplasty, Replacement, Knee - adverse effects</subject><subject>Calibration</subject><subject>Female</subject><subject>Humans</subject><subject>Joint replacement surgery</subject><subject>Knee</subject><subject>Learning algorithms</subject><subject>Logistic Models</subject><subject>Machine Learning</subject><subject>Male</subject><subject>Medicine</subject><subject>Medicine &amp; Public Health</subject><subject>Middle Aged</subject><subject>Mobility Limitation</subject><subject>Orthopedics</subject><subject>Postoperative Complications - physiopathology</subject><subject>Prognosis</subject><subject>Registries</subject><subject>Regression analysis</subject><subject>Regression models</subject><subject>Spline functions</subject><subject>Statistical analysis</subject><subject>Treatment Outcome</subject><subject>Walking</subject><issn>0942-2056</issn><issn>1433-7347</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2020</creationdate><recordtype>article</recordtype><sourceid>EIF</sourceid><sourceid>ABUWG</sourceid><sourceid>AFKRA</sourceid><sourceid>BENPR</sourceid><sourceid>CCPQU</sourceid><recordid>eNp9Uc2OFCEQJkbjjqMv4MGQePHSyl833UezUddkjRc9EwaKGXZpaIHZzT6HLyy9s2riwRMF9f1Q9SH0kpK3lBD5rhBCB9kROnWkHxnr5CO0oYLzTnIhH6MNmQTrGOmHM_SslCtCWimmp-iM05GzYZIb9POLNgcfAQfQOfq4xzPUQ7IF6wzYpHnRWe8C4JpwSHtfqjc4wz5DKT5FXMEcov9xhIJ9xEsG601dZQrcQFO41eF6vQY_-6rrSnEphHS7PtZUdcDXEaC51UNOS9Cl3j1HT5wOBV48nFv0_eOHb-cX3eXXT5_P3192RghaO2thGAc3WGpG5yS4Xk5Ss93UE9E6IwfOJz6AENKRabQ7K9zAtLOj44L0jG_Rm5PuktM6QVWzLwZC0BHSsSjGuWw7E3KFvv4HepWOObbfKTayvm1z9doidkKZnErJ4NSS_azznaJErZGpU2SqRabuI1OykV49SB93M9g_lN8ZNQA_AUprxT3kv97_kf0Fp5OlQw</recordid><startdate>20201001</startdate><enddate>20201001</enddate><creator>Pua, Yong-Hao</creator><creator>Kang, Hakmook</creator><creator>Thumboo, Julian</creator><creator>Clark, Ross Allan</creator><creator>Chew, Eleanor Shu-Xian</creator><creator>Poon, Cheryl Lian-Li</creator><creator>Chong, Hwei-Chi</creator><creator>Yeo, Seng-Jin</creator><general>Springer Berlin Heidelberg</general><general>Springer Nature B.V</general><scope>CGR</scope><scope>CUY</scope><scope>CVF</scope><scope>ECM</scope><scope>EIF</scope><scope>NPM</scope><scope>AAYXX</scope><scope>CITATION</scope><scope>3V.</scope><scope>7QO</scope><scope>7RV</scope><scope>7TS</scope><scope>7X7</scope><scope>7XB</scope><scope>88E</scope><scope>8AO</scope><scope>8FD</scope><scope>8FI</scope><scope>8FJ</scope><scope>8FK</scope><scope>ABUWG</scope><scope>AFKRA</scope><scope>BENPR</scope><scope>CCPQU</scope><scope>FR3</scope><scope>FYUFA</scope><scope>GHDGH</scope><scope>K9.</scope><scope>KB0</scope><scope>M0S</scope><scope>M1P</scope><scope>NAPCQ</scope><scope>P64</scope><scope>PQEST</scope><scope>PQQKQ</scope><scope>PQUKI</scope><scope>7X8</scope><orcidid>https://orcid.org/0000-0003-2313-9665</orcidid></search><sort><creationdate>20201001</creationdate><title>Machine learning methods are comparable to logistic regression techniques in predicting severe walking limitation following total knee arthroplasty</title><author>Pua, Yong-Hao ; Kang, Hakmook ; Thumboo, Julian ; Clark, Ross Allan ; Chew, Eleanor Shu-Xian ; Poon, Cheryl Lian-Li ; Chong, Hwei-Chi ; Yeo, Seng-Jin</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c441t-dde686f6d1c8ff7ef5797a2b9504e6883e33936e447f098dbd4f62afd8f340523</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2020</creationdate><topic>Aged</topic><topic>Algorithms</topic><topic>Arthroplasty (knee)</topic><topic>Arthroplasty, Replacement, Knee - adverse effects</topic><topic>Calibration</topic><topic>Female</topic><topic>Humans</topic><topic>Joint replacement surgery</topic><topic>Knee</topic><topic>Learning algorithms</topic><topic>Logistic Models</topic><topic>Machine Learning</topic><topic>Male</topic><topic>Medicine</topic><topic>Medicine &amp; Public Health</topic><topic>Middle Aged</topic><topic>Mobility Limitation</topic><topic>Orthopedics</topic><topic>Postoperative Complications - physiopathology</topic><topic>Prognosis</topic><topic>Registries</topic><topic>Regression analysis</topic><topic>Regression models</topic><topic>Spline functions</topic><topic>Statistical analysis</topic><topic>Treatment Outcome</topic><topic>Walking</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Pua, Yong-Hao</creatorcontrib><creatorcontrib>Kang, Hakmook</creatorcontrib><creatorcontrib>Thumboo, Julian</creatorcontrib><creatorcontrib>Clark, Ross Allan</creatorcontrib><creatorcontrib>Chew, Eleanor Shu-Xian</creatorcontrib><creatorcontrib>Poon, Cheryl Lian-Li</creatorcontrib><creatorcontrib>Chong, Hwei-Chi</creatorcontrib><creatorcontrib>Yeo, Seng-Jin</creatorcontrib><collection>Medline</collection><collection>MEDLINE</collection><collection>MEDLINE (Ovid)</collection><collection>MEDLINE</collection><collection>MEDLINE</collection><collection>PubMed</collection><collection>CrossRef</collection><collection>ProQuest Central (Corporate)</collection><collection>Biotechnology Research Abstracts</collection><collection>Nursing &amp; Allied Health Database</collection><collection>Physical Education Index</collection><collection>Health &amp; Medical Collection</collection><collection>ProQuest Central (purchase pre-March 2016)</collection><collection>Medical Database (Alumni Edition)</collection><collection>ProQuest Pharma Collection</collection><collection>Technology Research Database</collection><collection>Hospital Premium Collection</collection><collection>Hospital Premium Collection (Alumni Edition)</collection><collection>ProQuest Central (Alumni) (purchase pre-March 2016)</collection><collection>ProQuest Central (Alumni Edition)</collection><collection>ProQuest Central UK/Ireland</collection><collection>ProQuest Central</collection><collection>ProQuest One Community College</collection><collection>Engineering Research Database</collection><collection>Health Research Premium Collection</collection><collection>Health Research Premium Collection (Alumni)</collection><collection>ProQuest Health &amp; Medical Complete (Alumni)</collection><collection>Nursing &amp; Allied Health Database (Alumni Edition)</collection><collection>Health &amp; Medical Collection (Alumni Edition)</collection><collection>Medical Database</collection><collection>Nursing &amp; Allied Health Premium</collection><collection>Biotechnology and BioEngineering Abstracts</collection><collection>ProQuest One Academic Eastern Edition (DO NOT USE)</collection><collection>ProQuest One Academic</collection><collection>ProQuest One Academic UKI Edition</collection><collection>MEDLINE - Academic</collection><jtitle>Knee surgery, sports traumatology, arthroscopy : official journal of the ESSKA</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Pua, Yong-Hao</au><au>Kang, Hakmook</au><au>Thumboo, Julian</au><au>Clark, Ross Allan</au><au>Chew, Eleanor Shu-Xian</au><au>Poon, Cheryl Lian-Li</au><au>Chong, Hwei-Chi</au><au>Yeo, Seng-Jin</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Machine learning methods are comparable to logistic regression techniques in predicting severe walking limitation following total knee arthroplasty</atitle><jtitle>Knee surgery, sports traumatology, arthroscopy : official journal of the ESSKA</jtitle><stitle>Knee Surg Sports Traumatol Arthrosc</stitle><addtitle>Knee Surg Sports Traumatol Arthrosc</addtitle><date>2020-10-01</date><risdate>2020</risdate><volume>28</volume><issue>10</issue><spage>3207</spage><epage>3216</epage><pages>3207-3216</pages><issn>0942-2056</issn><eissn>1433-7347</eissn><abstract>Purpose Machine-learning methods are flexible prediction algorithms with potential advantages over conventional regression. This study aimed to use machine learning methods to predict post-total knee arthroplasty (TKA) walking limitation, and to compare their performance with that of logistic regression. Methods From the department’s clinical registry, a cohort of 4026 patients who underwent elective, primary TKA between July 2013 and July 2017 was identified. Candidate predictors included demographics and preoperative clinical, psychosocial, and outcome measures. The primary outcome was severe walking limitation at 6 months post-TKA, defined as a maximum walk time ≤ 15 min. Eight common regression (logistic, penalized logistic, and ordinal logistic with natural splines) and ensemble machine learning (random forest, extreme gradient boosting, and SuperLearner) methods were implemented to predict the probability of severe walking limitation. Models were compared on discrimination and calibration metrics. Results At 6 months post-TKA, 13% of patients had severe walking limitation. Machine learning and logistic regression models performed moderately [mean area under the ROC curves (AUC) 0.73–0.75]. Overall, the ordinal logistic regression model performed best while the SuperLearner performed best among machine learning methods, with negligible differences between them (Brier score difference, &lt; 0.001; 95% CI [− 0.0025, 0.002]). Conclusions When predicting post-TKA physical function, several machine learning methods did not outperform logistic regression—in particular, ordinal logistic regression that does not assume linearity in its predictors. Level of evidence Prognostic level II</abstract><cop>Berlin/Heidelberg</cop><pub>Springer Berlin Heidelberg</pub><pmid>31832697</pmid><doi>10.1007/s00167-019-05822-7</doi><tpages>10</tpages><orcidid>https://orcid.org/0000-0003-2313-9665</orcidid></addata></record>
fulltext fulltext
identifier ISSN: 0942-2056
ispartof Knee surgery, sports traumatology, arthroscopy : official journal of the ESSKA, 2020-10, Vol.28 (10), p.3207-3216
issn 0942-2056
1433-7347
language eng
recordid cdi_proquest_miscellaneous_2337000472
source MEDLINE; Access via Wiley Online Library; SpringerLink Journals (MCLS)
subjects Aged
Algorithms
Arthroplasty (knee)
Arthroplasty, Replacement, Knee - adverse effects
Calibration
Female
Humans
Joint replacement surgery
Knee
Learning algorithms
Logistic Models
Machine Learning
Male
Medicine
Medicine & Public Health
Middle Aged
Mobility Limitation
Orthopedics
Postoperative Complications - physiopathology
Prognosis
Registries
Regression analysis
Regression models
Spline functions
Statistical analysis
Treatment Outcome
Walking
title Machine learning methods are comparable to logistic regression techniques in predicting severe walking limitation following total knee arthroplasty
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2024-12-24T22%3A40%3A09IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest_cross&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Machine%20learning%20methods%20are%20comparable%20to%20logistic%20regression%20techniques%20in%20predicting%20severe%20walking%20limitation%20following%20total%20knee%20arthroplasty&rft.jtitle=Knee%20surgery,%20sports%20traumatology,%20arthroscopy%20:%20official%20journal%20of%20the%20ESSKA&rft.au=Pua,%20Yong-Hao&rft.date=2020-10-01&rft.volume=28&rft.issue=10&rft.spage=3207&rft.epage=3216&rft.pages=3207-3216&rft.issn=0942-2056&rft.eissn=1433-7347&rft_id=info:doi/10.1007/s00167-019-05822-7&rft_dat=%3Cproquest_cross%3E2337000472%3C/proquest_cross%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_pqid=2825832339&rft_id=info:pmid/31832697&rfr_iscdi=true