Propensity score model overfitting led to inflated variance of estimated odds ratios.
Identifieur interne : 001548 ( PubMed/Corpus ); précédent : 001547; suivant : 001549Propensity score model overfitting led to inflated variance of estimated odds ratios.
Auteurs : Tibor Schuster ; Wilfrid Kouokam Lowe ; Robert W. PlattSource :
- Journal of clinical epidemiology [ 1878-5921 ] ; 2016.
English descriptors
- KwdEn :
- MESH :
Abstract
Simulation studies suggest that the ratio of the number of events to the number of estimated parameters in a logistic regression model should be not less than 10 or 20 to 1 to achieve reliable effect estimates. Applications of propensity score approaches for confounding control in practice, however, do often not consider these recommendations.
DOI: 10.1016/j.jclinepi.2016.05.017
PubMed: 27498378
Links to Exploration step
pubmed:27498378Le document en format XML
<record><TEI><teiHeader><fileDesc><titleStmt><title xml:lang="en">Propensity score model overfitting led to inflated variance of estimated odds ratios.</title><author><name sortKey="Schuster, Tibor" sort="Schuster, Tibor" uniqKey="Schuster T" first="Tibor" last="Schuster">Tibor Schuster</name><affiliation><nlm:affiliation>Centre for Clinical Epidemiology, Lady Davis Institute for Medical Research, 3755 Chemin de la Côte-Sainte-Catherine, Montréal, Québec H3T 1E2, Canada; Department of Epidemiology, Biostatistics and Occupational Health, McGill University, Purvis Hall, 1020 Pine Avenue West, Montréal, Québec H3A 1A2, Canada; Clinical Epidemiology and Biostatistics Unit and the Melbourne Children's Trial Centre, Murdoch Childrens Research Institute, Royal Children's Hospital, 50 Flemington Road, Parkville, Victoria 3052, Australia; Department of Paediatrics, University of Melbourne, Melbourne, Victoria 3010, Australia. Electronic address: Tibor.Schuster@mcri.edu.au.</nlm:affiliation></affiliation></author><author><name sortKey="Lowe, Wilfrid Kouokam" sort="Lowe, Wilfrid Kouokam" uniqKey="Lowe W" first="Wilfrid Kouokam" last="Lowe">Wilfrid Kouokam Lowe</name><affiliation><nlm:affiliation>Centre for Clinical Epidemiology, Lady Davis Institute for Medical Research, 3755 Chemin de la Côte-Sainte-Catherine, Montréal, Québec H3T 1E2, Canada; Department of Epidemiology, Biostatistics and Occupational Health, McGill University, Purvis Hall, 1020 Pine Avenue West, Montréal, Québec H3A 1A2, Canada; UFR de Mathématique et d'Informatique, Université de Strasbourg, 7 Rue René Descartes, 67084 Strasbourg, France.</nlm:affiliation></affiliation></author><author><name sortKey="Platt, Robert W" sort="Platt, Robert W" uniqKey="Platt R" first="Robert W" last="Platt">Robert W. Platt</name><affiliation><nlm:affiliation>Department of Epidemiology, Biostatistics and Occupational Health, McGill University, Purvis Hall, 1020 Pine Avenue West, Montréal, Québec H3A 1A2, Canada; Department of Pediatrics, McGill University, Montreal Children's Hospital, 1001 Décarie Boulevard, Montreal, Québec H4A 3J1, Canada.</nlm:affiliation></affiliation></author></titleStmt><publicationStmt><idno type="wicri:source">PubMed</idno><date when="2016">2016</date><idno type="RBID">pubmed:27498378</idno><idno type="pmid">27498378</idno><idno type="doi">10.1016/j.jclinepi.2016.05.017</idno><idno type="wicri:Area/PubMed/Corpus">001548</idno><idno type="wicri:explorRef" wicri:stream="PubMed" wicri:step="Corpus" wicri:corpus="PubMed">001548</idno></publicationStmt><sourceDesc><biblStruct><analytic><title xml:lang="en">Propensity score model overfitting led to inflated variance of estimated odds ratios.</title><author><name sortKey="Schuster, Tibor" sort="Schuster, Tibor" uniqKey="Schuster T" first="Tibor" last="Schuster">Tibor Schuster</name><affiliation><nlm:affiliation>Centre for Clinical Epidemiology, Lady Davis Institute for Medical Research, 3755 Chemin de la Côte-Sainte-Catherine, Montréal, Québec H3T 1E2, Canada; Department of Epidemiology, Biostatistics and Occupational Health, McGill University, Purvis Hall, 1020 Pine Avenue West, Montréal, Québec H3A 1A2, Canada; Clinical Epidemiology and Biostatistics Unit and the Melbourne Children's Trial Centre, Murdoch Childrens Research Institute, Royal Children's Hospital, 50 Flemington Road, Parkville, Victoria 3052, Australia; Department of Paediatrics, University of Melbourne, Melbourne, Victoria 3010, Australia. Electronic address: Tibor.Schuster@mcri.edu.au.</nlm:affiliation></affiliation></author><author><name sortKey="Lowe, Wilfrid Kouokam" sort="Lowe, Wilfrid Kouokam" uniqKey="Lowe W" first="Wilfrid Kouokam" last="Lowe">Wilfrid Kouokam Lowe</name><affiliation><nlm:affiliation>Centre for Clinical Epidemiology, Lady Davis Institute for Medical Research, 3755 Chemin de la Côte-Sainte-Catherine, Montréal, Québec H3T 1E2, Canada; Department of Epidemiology, Biostatistics and Occupational Health, McGill University, Purvis Hall, 1020 Pine Avenue West, Montréal, Québec H3A 1A2, Canada; UFR de Mathématique et d'Informatique, Université de Strasbourg, 7 Rue René Descartes, 67084 Strasbourg, France.</nlm:affiliation></affiliation></author><author><name sortKey="Platt, Robert W" sort="Platt, Robert W" uniqKey="Platt R" first="Robert W" last="Platt">Robert W. Platt</name><affiliation><nlm:affiliation>Department of Epidemiology, Biostatistics and Occupational Health, McGill University, Purvis Hall, 1020 Pine Avenue West, Montréal, Québec H3A 1A2, Canada; Department of Pediatrics, McGill University, Montreal Children's Hospital, 1001 Décarie Boulevard, Montreal, Québec H4A 3J1, Canada.</nlm:affiliation></affiliation></author></analytic><series><title level="j">Journal of clinical epidemiology</title><idno type="eISSN">1878-5921</idno><imprint><date when="2016" type="published">2016</date></imprint></series></biblStruct></sourceDesc></fileDesc><profileDesc><textClass><keywords scheme="KwdEn" xml:lang="en"><term>Epidemiologic Research Design</term><term>Humans</term><term>Odds Ratio</term><term>Propensity Score</term><term>Reproducibility of Results</term></keywords><keywords scheme="MESH" xml:lang="en"><term>Epidemiologic Research Design</term><term>Humans</term><term>Odds Ratio</term><term>Propensity Score</term><term>Reproducibility of Results</term></keywords></textClass></profileDesc></teiHeader><front><div type="abstract" xml:lang="en">Simulation studies suggest that the ratio of the number of events to the number of estimated parameters in a logistic regression model should be not less than 10 or 20 to 1 to achieve reliable effect estimates. Applications of propensity score approaches for confounding control in practice, however, do often not consider these recommendations.</div></front></TEI><pubmed><MedlineCitation Status="MEDLINE" Owner="NLM"><PMID Version="1">27498378</PMID><DateCreated><Year>2016</Year><Month>09</Month><Day>19</Day></DateCreated><DateCompleted><Year>2017</Year><Month>06</Month><Day>26</Day></DateCompleted><DateRevised><Year>2017</Year><Month>06</Month><Day>26</Day></DateRevised><Article PubModel="Print-Electronic"><Journal><ISSN IssnType="Electronic">1878-5921</ISSN><JournalIssue CitedMedium="Internet"><Volume>80</Volume><PubDate><Year>2016</Year><Month>Dec</Month></PubDate></JournalIssue><Title>Journal of clinical epidemiology</Title><ISOAbbreviation>J Clin Epidemiol</ISOAbbreviation></Journal><ArticleTitle>Propensity score model overfitting led to inflated variance of estimated odds ratios.</ArticleTitle><Pagination><MedlinePgn>97-106</MedlinePgn></Pagination><ELocationID EIdType="pii" ValidYN="Y">S0895-4356(16)30277-3</ELocationID><ELocationID EIdType="doi" ValidYN="Y">10.1016/j.jclinepi.2016.05.017</ELocationID><Abstract><AbstractText Label="OBJECTIVE" NlmCategory="OBJECTIVE">Simulation studies suggest that the ratio of the number of events to the number of estimated parameters in a logistic regression model should be not less than 10 or 20 to 1 to achieve reliable effect estimates. Applications of propensity score approaches for confounding control in practice, however, do often not consider these recommendations.</AbstractText><AbstractText Label="STUDY DESIGN AND SETTING" NlmCategory="METHODS">We conducted extensive Monte Carlo and plasmode simulation studies to investigate the impact of propensity score model overfitting on the performance in estimating conditional and marginal odds ratios using different established propensity score inference approaches. We assessed estimate accuracy and precision as well as associated type I error and type II error rates in testing the null hypothesis of no exposure effect.</AbstractText><AbstractText Label="RESULTS" NlmCategory="RESULTS">For all inference approaches considered, our simulation study revealed considerably inflated standard errors of effect estimates when using overfitted propensity score models. Overfitting did not considerably affect type I error rates for most inference approaches. However, because of residual confounding, estimation performance and type I error probabilities were unsatisfactory when using propensity score quintile adjustment.</AbstractText><AbstractText Label="CONCLUSION" NlmCategory="CONCLUSIONS">Overfitting of propensity score models should be avoided to obtain reliable estimates of treatment or exposure effects in individual studies.</AbstractText><CopyrightInformation>Copyright © 2016 Elsevier Inc. All rights reserved.</CopyrightInformation></Abstract><AuthorList CompleteYN="Y"><Author ValidYN="Y"><LastName>Schuster</LastName><ForeName>Tibor</ForeName><Initials>T</Initials><AffiliationInfo><Affiliation>Centre for Clinical Epidemiology, Lady Davis Institute for Medical Research, 3755 Chemin de la Côte-Sainte-Catherine, Montréal, Québec H3T 1E2, Canada; Department of Epidemiology, Biostatistics and Occupational Health, McGill University, Purvis Hall, 1020 Pine Avenue West, Montréal, Québec H3A 1A2, Canada; Clinical Epidemiology and Biostatistics Unit and the Melbourne Children's Trial Centre, Murdoch Childrens Research Institute, Royal Children's Hospital, 50 Flemington Road, Parkville, Victoria 3052, Australia; Department of Paediatrics, University of Melbourne, Melbourne, Victoria 3010, Australia. Electronic address: Tibor.Schuster@mcri.edu.au.</Affiliation></AffiliationInfo></Author><Author ValidYN="Y"><LastName>Lowe</LastName><ForeName>Wilfrid Kouokam</ForeName><Initials>WK</Initials><AffiliationInfo><Affiliation>Centre for Clinical Epidemiology, Lady Davis Institute for Medical Research, 3755 Chemin de la Côte-Sainte-Catherine, Montréal, Québec H3T 1E2, Canada; Department of Epidemiology, Biostatistics and Occupational Health, McGill University, Purvis Hall, 1020 Pine Avenue West, Montréal, Québec H3A 1A2, Canada; UFR de Mathématique et d'Informatique, Université de Strasbourg, 7 Rue René Descartes, 67084 Strasbourg, France.</Affiliation></AffiliationInfo></Author><Author ValidYN="Y"><LastName>Platt</LastName><ForeName>Robert W</ForeName><Initials>RW</Initials><AffiliationInfo><Affiliation>Department of Epidemiology, Biostatistics and Occupational Health, McGill University, Purvis Hall, 1020 Pine Avenue West, Montréal, Québec H3A 1A2, Canada; Department of Pediatrics, McGill University, Montreal Children's Hospital, 1001 Décarie Boulevard, Montreal, Québec H4A 3J1, Canada.</Affiliation></AffiliationInfo></Author></AuthorList><Language>eng</Language><PublicationTypeList><PublicationType UI="D016428">Journal Article</PublicationType></PublicationTypeList><ArticleDate DateType="Electronic"><Year>2016</Year><Month>09</Month><Day>14</Day></ArticleDate></Article><MedlineJournalInfo><Country>United States</Country><MedlineTA>J Clin Epidemiol</MedlineTA><NlmUniqueID>8801383</NlmUniqueID><ISSNLinking>0895-4356</ISSNLinking></MedlineJournalInfo><CitationSubset>IM</CitationSubset><MeshHeadingList><MeshHeading><DescriptorName UI="D015340" MajorTopicYN="Y">Epidemiologic Research Design</DescriptorName></MeshHeading><MeshHeading><DescriptorName UI="D006801" MajorTopicYN="N">Humans</DescriptorName></MeshHeading><MeshHeading><DescriptorName UI="D016017" MajorTopicYN="Y">Odds Ratio</DescriptorName></MeshHeading><MeshHeading><DescriptorName UI="D057216" MajorTopicYN="Y">Propensity Score</DescriptorName></MeshHeading><MeshHeading><DescriptorName UI="D015203" MajorTopicYN="N">Reproducibility of Results</DescriptorName></MeshHeading></MeshHeadingList><KeywordList Owner="NOTNLM"><Keyword MajorTopicYN="N">Confounder adjustment</Keyword><Keyword MajorTopicYN="N">Inverse probability weighting</Keyword><Keyword MajorTopicYN="N">Logistic regression</Keyword><Keyword MajorTopicYN="N">Odds ratio</Keyword><Keyword MajorTopicYN="N">Overfitting</Keyword><Keyword MajorTopicYN="N">Propensity score</Keyword></KeywordList></MedlineCitation><PubmedData><History><PubMedPubDate PubStatus="received"><Year>2016</Year><Month>01</Month><Day>06</Day></PubMedPubDate><PubMedPubDate PubStatus="revised"><Year>2016</Year><Month>04</Month><Day>30</Day></PubMedPubDate><PubMedPubDate PubStatus="accepted"><Year>2016</Year><Month>05</Month><Day>11</Day></PubMedPubDate><PubMedPubDate PubStatus="pubmed"><Year>2016</Year><Month>8</Month><Day>9</Day><Hour>6</Hour><Minute>0</Minute></PubMedPubDate><PubMedPubDate PubStatus="medline"><Year>2017</Year><Month>6</Month><Day>27</Day><Hour>6</Hour><Minute>0</Minute></PubMedPubDate><PubMedPubDate PubStatus="entrez"><Year>2016</Year><Month>8</Month><Day>8</Day><Hour>6</Hour><Minute>0</Minute></PubMedPubDate></History><PublicationStatus>ppublish</PublicationStatus><ArticleIdList><ArticleId IdType="pubmed">27498378</ArticleId><ArticleId IdType="pii">S0895-4356(16)30277-3</ArticleId><ArticleId IdType="doi">10.1016/j.jclinepi.2016.05.017</ArticleId></ArticleIdList></PubmedData></pubmed></record>Pour manipuler ce document sous Unix (Dilib)
EXPLOR_STEP=$WICRI_ROOT/Wicri/Asie/explor/AustralieFrV1/Data/PubMed/Corpus
HfdSelect -h $EXPLOR_STEP/biblio.hfd -nk 001548 | SxmlIndent | more
Ou
HfdSelect -h $EXPLOR_AREA/Data/PubMed/Corpus/biblio.hfd -nk 001548 | SxmlIndent | more
Pour mettre un lien sur cette page dans le réseau Wicri
{{Explor lien
|wiki= Wicri/Asie
|area= AustralieFrV1
|flux= PubMed
|étape= Corpus
|type= RBID
|clé= pubmed:27498378
|texte= Propensity score model overfitting led to inflated variance of estimated odds ratios.
}}
Pour générer des pages wiki
HfdIndexSelect -h $EXPLOR_AREA/Data/PubMed/Corpus/RBID.i -Sk "pubmed:27498378" \
| HfdSelect -Kh $EXPLOR_AREA/Data/PubMed/Corpus/biblio.hfd \
| NlmPubMed2Wicri -a AustralieFrV1
| This area was generated with Dilib version V0.6.33. |  |