A comparison of bootstrap approaches for estimating uncertainty of parameters in linear mixed-effects models.
Identifieur interne : 000740 ( PubMed/Corpus ); précédent : 000739; suivant : 000741A comparison of bootstrap approaches for estimating uncertainty of parameters in linear mixed-effects models.
Auteurs : Hoai-Thu Thai ; France Mentré ; Nicholas H G. Holford ; Christine Veyrat-Follet ; Emmanuelle CometsSource :
- Pharmaceutical statistics [ 1539-1612 ]
English descriptors
- KwdEn :
- MESH :
Abstract
A version of the nonparametric bootstrap, which resamples the entire subjects from original data, called the case bootstrap, has been increasingly used for estimating uncertainty of parameters in mixed-effects models. It is usually applied to obtain more robust estimates of the parameters and more realistic confidence intervals (CIs). Alternative bootstrap methods, such as residual bootstrap and parametric bootstrap that resample both random effects and residuals, have been proposed to better take into account the hierarchical structure of multi-level and longitudinal data. However, few studies have been performed to compare these different approaches. In this study, we used simulation to evaluate bootstrap methods proposed for linear mixed-effect models. We also compared the results obtained by maximum likelihood (ML) and restricted maximum likelihood (REML). Our simulation studies evidenced the good performance of the case bootstrap as well as the bootstraps of both random effects and residuals. On the other hand, the bootstrap methods that resample only the residuals and the bootstraps combining case and residuals performed poorly. REML and ML provided similar bootstrap estimates of uncertainty, but there was slightly more bias and poorer coverage rate for variance parameters with ML in the sparse design. We applied the proposed methods to a real dataset from a study investigating the natural evolution of Parkinson's disease and were able to confirm that the methods provide plausible estimates of uncertainty. Given that most real-life datasets tend to exhibit heterogeneity in sampling schedules, the residual bootstraps would be expected to perform better than the case bootstrap.
DOI: 10.1002/pst.1561
PubMed: 23457061
Links to Exploration step
pubmed:23457061Le document en format XML
<record><TEI><teiHeader><fileDesc><titleStmt><title xml:lang="en">A comparison of bootstrap approaches for estimating uncertainty of parameters in linear mixed-effects models.</title><author><name sortKey="Thai, Hoai Thu" sort="Thai, Hoai Thu" uniqKey="Thai H" first="Hoai-Thu" last="Thai">Hoai-Thu Thai</name><affiliation><nlm:affiliation>Univ Paris Diderot, Sorbonne Paris Cité, UMR 738, F-75018 Paris, France. hoai-thu.thai@inserm.fr</nlm:affiliation></affiliation></author><author><name sortKey="Mentre, France" sort="Mentre, France" uniqKey="Mentre F" first="France" last="Mentré">France Mentré</name></author><author><name sortKey="Holford, Nicholas H G" sort="Holford, Nicholas H G" uniqKey="Holford N" first="Nicholas H G" last="Holford">Nicholas H G. Holford</name></author><author><name sortKey="Veyrat Follet, Christine" sort="Veyrat Follet, Christine" uniqKey="Veyrat Follet C" first="Christine" last="Veyrat-Follet">Christine Veyrat-Follet</name></author><author><name sortKey="Comets, Emmanuelle" sort="Comets, Emmanuelle" uniqKey="Comets E" first="Emmanuelle" last="Comets">Emmanuelle Comets</name></author></titleStmt><publicationStmt><idno type="wicri:source">PubMed</idno><date when="????"><PubDate><MedlineDate>2013 May-Jun</MedlineDate></PubDate></date><idno type="RBID">pubmed:23457061</idno><idno type="pmid">23457061</idno><idno type="doi">10.1002/pst.1561</idno><idno type="wicri:Area/PubMed/Corpus">000740</idno><idno type="wicri:explorRef" wicri:stream="PubMed" wicri:step="Corpus" wicri:corpus="PubMed">000740</idno></publicationStmt><sourceDesc><biblStruct><analytic><title xml:lang="en">A comparison of bootstrap approaches for estimating uncertainty of parameters in linear mixed-effects models.</title><author><name sortKey="Thai, Hoai Thu" sort="Thai, Hoai Thu" uniqKey="Thai H" first="Hoai-Thu" last="Thai">Hoai-Thu Thai</name><affiliation><nlm:affiliation>Univ Paris Diderot, Sorbonne Paris Cité, UMR 738, F-75018 Paris, France. hoai-thu.thai@inserm.fr</nlm:affiliation></affiliation></author><author><name sortKey="Mentre, France" sort="Mentre, France" uniqKey="Mentre F" first="France" last="Mentré">France Mentré</name></author><author><name sortKey="Holford, Nicholas H G" sort="Holford, Nicholas H G" uniqKey="Holford N" first="Nicholas H G" last="Holford">Nicholas H G. Holford</name></author><author><name sortKey="Veyrat Follet, Christine" sort="Veyrat Follet, Christine" uniqKey="Veyrat Follet C" first="Christine" last="Veyrat-Follet">Christine Veyrat-Follet</name></author><author><name sortKey="Comets, Emmanuelle" sort="Comets, Emmanuelle" uniqKey="Comets E" first="Emmanuelle" last="Comets">Emmanuelle Comets</name></author></analytic><series><title level="j">Pharmaceutical statistics</title><idno type="eISSN">1539-1612</idno></series></biblStruct></sourceDesc></fileDesc><profileDesc><textClass><keywords scheme="KwdEn" xml:lang="en"><term>Bias (Epidemiology)</term><term>Clinical Trials as Topic (methods)</term><term>Computer Simulation</term><term>Confidence Intervals</term><term>Data Interpretation, Statistical</term><term>Humans</term><term>Likelihood Functions</term><term>Linear Models</term><term>Models, Statistical</term><term>Parkinson Disease (drug therapy)</term><term>Research Design</term><term>Statistics, Nonparametric</term></keywords><keywords scheme="MESH" qualifier="drug therapy" xml:lang="en"><term>Parkinson Disease</term></keywords><keywords scheme="MESH" qualifier="methods" xml:lang="en"><term>Clinical Trials as Topic</term></keywords><keywords scheme="MESH" xml:lang="en"><term>Bias (Epidemiology)</term><term>Computer Simulation</term><term>Confidence Intervals</term><term>Data Interpretation, Statistical</term><term>Humans</term><term>Likelihood Functions</term><term>Linear Models</term><term>Models, Statistical</term><term>Research Design</term><term>Statistics, Nonparametric</term></keywords></textClass></profileDesc></teiHeader><front><div type="abstract" xml:lang="en">A version of the nonparametric bootstrap, which resamples the entire subjects from original data, called the case bootstrap, has been increasingly used for estimating uncertainty of parameters in mixed-effects models. It is usually applied to obtain more robust estimates of the parameters and more realistic confidence intervals (CIs). Alternative bootstrap methods, such as residual bootstrap and parametric bootstrap that resample both random effects and residuals, have been proposed to better take into account the hierarchical structure of multi-level and longitudinal data. However, few studies have been performed to compare these different approaches. In this study, we used simulation to evaluate bootstrap methods proposed for linear mixed-effect models. We also compared the results obtained by maximum likelihood (ML) and restricted maximum likelihood (REML). Our simulation studies evidenced the good performance of the case bootstrap as well as the bootstraps of both random effects and residuals. On the other hand, the bootstrap methods that resample only the residuals and the bootstraps combining case and residuals performed poorly. REML and ML provided similar bootstrap estimates of uncertainty, but there was slightly more bias and poorer coverage rate for variance parameters with ML in the sparse design. We applied the proposed methods to a real dataset from a study investigating the natural evolution of Parkinson's disease and were able to confirm that the methods provide plausible estimates of uncertainty. Given that most real-life datasets tend to exhibit heterogeneity in sampling schedules, the residual bootstraps would be expected to perform better than the case bootstrap.</div></front></TEI><pubmed><MedlineCitation Status="MEDLINE" Owner="NLM"><PMID Version="1">23457061</PMID><DateCreated><Year>2013</Year><Month>05</Month><Day>24</Day></DateCreated><DateCompleted><Year>2013</Year><Month>12</Month><Day>27</Day></DateCompleted><DateRevised><Year>2013</Year><Month>05</Month><Day>24</Day></DateRevised><Article PubModel="Print-Electronic"><Journal><ISSN IssnType="Electronic">1539-1612</ISSN><JournalIssue CitedMedium="Internet"><Volume>12</Volume><Issue>3</Issue><PubDate><MedlineDate>2013 May-Jun</MedlineDate></PubDate></JournalIssue><Title>Pharmaceutical statistics</Title><ISOAbbreviation>Pharm Stat</ISOAbbreviation></Journal><ArticleTitle>A comparison of bootstrap approaches for estimating uncertainty of parameters in linear mixed-effects models.</ArticleTitle><Pagination><MedlinePgn>129-40</MedlinePgn></Pagination><ELocationID EIdType="doi" ValidYN="Y">10.1002/pst.1561</ELocationID><Abstract><AbstractText>A version of the nonparametric bootstrap, which resamples the entire subjects from original data, called the case bootstrap, has been increasingly used for estimating uncertainty of parameters in mixed-effects models. It is usually applied to obtain more robust estimates of the parameters and more realistic confidence intervals (CIs). Alternative bootstrap methods, such as residual bootstrap and parametric bootstrap that resample both random effects and residuals, have been proposed to better take into account the hierarchical structure of multi-level and longitudinal data. However, few studies have been performed to compare these different approaches. In this study, we used simulation to evaluate bootstrap methods proposed for linear mixed-effect models. We also compared the results obtained by maximum likelihood (ML) and restricted maximum likelihood (REML). Our simulation studies evidenced the good performance of the case bootstrap as well as the bootstraps of both random effects and residuals. On the other hand, the bootstrap methods that resample only the residuals and the bootstraps combining case and residuals performed poorly. REML and ML provided similar bootstrap estimates of uncertainty, but there was slightly more bias and poorer coverage rate for variance parameters with ML in the sparse design. We applied the proposed methods to a real dataset from a study investigating the natural evolution of Parkinson's disease and were able to confirm that the methods provide plausible estimates of uncertainty. Given that most real-life datasets tend to exhibit heterogeneity in sampling schedules, the residual bootstraps would be expected to perform better than the case bootstrap.</AbstractText><CopyrightInformation>Copyright © 2013 John Wiley & Sons, Ltd.</CopyrightInformation></Abstract><AuthorList CompleteYN="Y"><Author ValidYN="Y"><LastName>Thai</LastName><ForeName>Hoai-Thu</ForeName><Initials>HT</Initials><AffiliationInfo><Affiliation>Univ Paris Diderot, Sorbonne Paris Cité, UMR 738, F-75018 Paris, France. hoai-thu.thai@inserm.fr</Affiliation></AffiliationInfo></Author><Author ValidYN="Y"><LastName>Mentré</LastName><ForeName>France</ForeName><Initials>F</Initials></Author><Author ValidYN="Y"><LastName>Holford</LastName><ForeName>Nicholas H G</ForeName><Initials>NH</Initials></Author><Author ValidYN="Y"><LastName>Veyrat-Follet</LastName><ForeName>Christine</ForeName><Initials>C</Initials></Author><Author ValidYN="Y"><LastName>Comets</LastName><ForeName>Emmanuelle</ForeName><Initials>E</Initials></Author></AuthorList><Language>eng</Language><PublicationTypeList><PublicationType UI="D003160">Comparative Study</PublicationType><PublicationType UI="D016428">Journal Article</PublicationType></PublicationTypeList><ArticleDate DateType="Electronic"><Year>2013</Year><Month>03</Month><Day>04</Day></ArticleDate></Article><MedlineJournalInfo><Country>England</Country><MedlineTA>Pharm Stat</MedlineTA><NlmUniqueID>101201192</NlmUniqueID><ISSNLinking>1539-1604</ISSNLinking></MedlineJournalInfo><CitationSubset>IM</CitationSubset><MeshHeadingList><MeshHeading><DescriptorName UI="D015982" MajorTopicYN="N">Bias (Epidemiology)</DescriptorName></MeshHeading><MeshHeading><DescriptorName UI="D002986" MajorTopicYN="N">Clinical Trials as Topic</DescriptorName><QualifierName UI="Q000379" MajorTopicYN="Y">methods</QualifierName></MeshHeading><MeshHeading><DescriptorName UI="D003198" MajorTopicYN="N">Computer Simulation</DescriptorName></MeshHeading><MeshHeading><DescriptorName UI="D016001" MajorTopicYN="N">Confidence Intervals</DescriptorName></MeshHeading><MeshHeading><DescriptorName UI="D003627" MajorTopicYN="N">Data Interpretation, Statistical</DescriptorName></MeshHeading><MeshHeading><DescriptorName UI="D006801" MajorTopicYN="N">Humans</DescriptorName></MeshHeading><MeshHeading><DescriptorName UI="D016013" MajorTopicYN="N">Likelihood Functions</DescriptorName></MeshHeading><MeshHeading><DescriptorName UI="D016014" MajorTopicYN="N">Linear Models</DescriptorName></MeshHeading><MeshHeading><DescriptorName UI="D015233" MajorTopicYN="Y">Models, Statistical</DescriptorName></MeshHeading><MeshHeading><DescriptorName UI="D010300" MajorTopicYN="N">Parkinson Disease</DescriptorName><QualifierName UI="Q000188" MajorTopicYN="N">drug therapy</QualifierName></MeshHeading><MeshHeading><DescriptorName UI="D012107" MajorTopicYN="Y">Research Design</DescriptorName></MeshHeading><MeshHeading><DescriptorName UI="D018709" MajorTopicYN="N">Statistics, Nonparametric</DescriptorName></MeshHeading></MeshHeadingList></MedlineCitation><PubmedData><History><PubMedPubDate PubStatus="received"><Year>2012</Year><Month>01</Month><Day>19</Day></PubMedPubDate><PubMedPubDate PubStatus="revised"><Year>2013</Year><Month>01</Month><Day>31</Day></PubMedPubDate><PubMedPubDate PubStatus="accepted"><Year>2013</Year><Month>02</Month><Day>04</Day></PubMedPubDate><PubMedPubDate PubStatus="entrez"><Year>2013</Year><Month>3</Month><Day>5</Day><Hour>6</Hour><Minute>0</Minute></PubMedPubDate><PubMedPubDate PubStatus="pubmed"><Year>2013</Year><Month>3</Month><Day>5</Day><Hour>6</Hour><Minute>0</Minute></PubMedPubDate><PubMedPubDate PubStatus="medline"><Year>2013</Year><Month>12</Month><Day>29</Day><Hour>6</Hour><Minute>0</Minute></PubMedPubDate></History><PublicationStatus>ppublish</PublicationStatus><ArticleIdList><ArticleId IdType="pubmed">23457061</ArticleId><ArticleId IdType="doi">10.1002/pst.1561</ArticleId></ArticleIdList></PubmedData></pubmed></record>Pour manipuler ce document sous Unix (Dilib)
EXPLOR_STEP=$WICRI_ROOT/Wicri/Sante/explor/ParkinsonFranceV1/Data/PubMed/Corpus
HfdSelect -h $EXPLOR_STEP/biblio.hfd -nk 000740 | SxmlIndent | more
Ou
HfdSelect -h $EXPLOR_AREA/Data/PubMed/Corpus/biblio.hfd -nk 000740 | SxmlIndent | more
Pour mettre un lien sur cette page dans le réseau Wicri
{{Explor lien
|wiki= Wicri/Sante
|area= ParkinsonFranceV1
|flux= PubMed
|étape= Corpus
|type= RBID
|clé= pubmed:23457061
|texte= A comparison of bootstrap approaches for estimating uncertainty of parameters in linear mixed-effects models.
}}
Pour générer des pages wiki
HfdIndexSelect -h $EXPLOR_AREA/Data/PubMed/Corpus/RBID.i -Sk "pubmed:23457061" \
| HfdSelect -Kh $EXPLOR_AREA/Data/PubMed/Corpus/biblio.hfd \
| NlmPubMed2Wicri -a ParkinsonFranceV1
| This area was generated with Dilib version V0.6.29. |  |