Bayesian hierarchical modeling based on multisource exchangeability

Bayesian hierarchical models produce shrinkage estimators that can be used as the basis for integrating supplementary data into the analysis of a primary data source. Established approaches should be considered limited, however, because posterior estimation either requires prespecification of a shri...

Full description

Saved in:
Bibliographic Details
Published in:Biostatistics (Oxford, England) Vol. 19; no. 2; p. 169
Main Authors: Kaizer, Alexander M, Koopmeiners, Joseph S, Hobbs, Brian P
Format: Journal Article
Language:English
Published: England 01.04.2018
Subjects:
ISSN:1468-4357, 1468-4357
Online Access:Get more information
Tags: Add Tag
No Tags, Be the first to tag this record!
Abstract Bayesian hierarchical models produce shrinkage estimators that can be used as the basis for integrating supplementary data into the analysis of a primary data source. Established approaches should be considered limited, however, because posterior estimation either requires prespecification of a shrinkage weight for each source or relies on the data to inform a single parameter, which determines the extent of influence or shrinkage from all sources, risking considerable bias or minimal borrowing. We introduce multisource exchangeability models (MEMs), a general Bayesian approach for integrating multiple, potentially non-exchangeable, supplemental data sources into the analysis of a primary data source. Our proposed modeling framework yields source-specific smoothing parameters that can be estimated in the presence of the data to facilitate a dynamic multi-resolution smoothed estimator that is asymptotically consistent while reducing the dimensionality of the prior space. When compared with competing Bayesian hierarchical modeling strategies, we demonstrate that MEMs achieve approximately 2.2 times larger median effective supplemental sample size when the supplemental data sources are exchangeable as well as a 56% reduction in bias when there is heterogeneity among the supplemental sources. We illustrate the application of MEMs using a recently completed randomized trial of very low nicotine content cigarettes, which resulted in a 30% improvement in efficiency compared with the standard analysis.
AbstractList Bayesian hierarchical models produce shrinkage estimators that can be used as the basis for integrating supplementary data into the analysis of a primary data source. Established approaches should be considered limited, however, because posterior estimation either requires prespecification of a shrinkage weight for each source or relies on the data to inform a single parameter, which determines the extent of influence or shrinkage from all sources, risking considerable bias or minimal borrowing. We introduce multisource exchangeability models (MEMs), a general Bayesian approach for integrating multiple, potentially non-exchangeable, supplemental data sources into the analysis of a primary data source. Our proposed modeling framework yields source-specific smoothing parameters that can be estimated in the presence of the data to facilitate a dynamic multi-resolution smoothed estimator that is asymptotically consistent while reducing the dimensionality of the prior space. When compared with competing Bayesian hierarchical modeling strategies, we demonstrate that MEMs achieve approximately 2.2 times larger median effective supplemental sample size when the supplemental data sources are exchangeable as well as a 56% reduction in bias when there is heterogeneity among the supplemental sources. We illustrate the application of MEMs using a recently completed randomized trial of very low nicotine content cigarettes, which resulted in a 30% improvement in efficiency compared with the standard analysis.Bayesian hierarchical models produce shrinkage estimators that can be used as the basis for integrating supplementary data into the analysis of a primary data source. Established approaches should be considered limited, however, because posterior estimation either requires prespecification of a shrinkage weight for each source or relies on the data to inform a single parameter, which determines the extent of influence or shrinkage from all sources, risking considerable bias or minimal borrowing. We introduce multisource exchangeability models (MEMs), a general Bayesian approach for integrating multiple, potentially non-exchangeable, supplemental data sources into the analysis of a primary data source. Our proposed modeling framework yields source-specific smoothing parameters that can be estimated in the presence of the data to facilitate a dynamic multi-resolution smoothed estimator that is asymptotically consistent while reducing the dimensionality of the prior space. When compared with competing Bayesian hierarchical modeling strategies, we demonstrate that MEMs achieve approximately 2.2 times larger median effective supplemental sample size when the supplemental data sources are exchangeable as well as a 56% reduction in bias when there is heterogeneity among the supplemental sources. We illustrate the application of MEMs using a recently completed randomized trial of very low nicotine content cigarettes, which resulted in a 30% improvement in efficiency compared with the standard analysis.
Bayesian hierarchical models produce shrinkage estimators that can be used as the basis for integrating supplementary data into the analysis of a primary data source. Established approaches should be considered limited, however, because posterior estimation either requires prespecification of a shrinkage weight for each source or relies on the data to inform a single parameter, which determines the extent of influence or shrinkage from all sources, risking considerable bias or minimal borrowing. We introduce multisource exchangeability models (MEMs), a general Bayesian approach for integrating multiple, potentially non-exchangeable, supplemental data sources into the analysis of a primary data source. Our proposed modeling framework yields source-specific smoothing parameters that can be estimated in the presence of the data to facilitate a dynamic multi-resolution smoothed estimator that is asymptotically consistent while reducing the dimensionality of the prior space. When compared with competing Bayesian hierarchical modeling strategies, we demonstrate that MEMs achieve approximately 2.2 times larger median effective supplemental sample size when the supplemental data sources are exchangeable as well as a 56% reduction in bias when there is heterogeneity among the supplemental sources. We illustrate the application of MEMs using a recently completed randomized trial of very low nicotine content cigarettes, which resulted in a 30% improvement in efficiency compared with the standard analysis.
Author Kaizer, Alexander M
Koopmeiners, Joseph S
Hobbs, Brian P
Author_xml – sequence: 1
  givenname: Alexander M
  surname: Kaizer
  fullname: Kaizer, Alexander M
  organization: Division of Biostatistics, University of Minnesota, A460 Mayo Building, MMC 303 420 Delaware St. SE, Minneapolis, MN 55455, USA
– sequence: 2
  givenname: Joseph S
  surname: Koopmeiners
  fullname: Koopmeiners, Joseph S
  organization: The University of Texas MD Anderson Cancer Center, 1515 Holcombe Blvd. Houston, TX 77030, USA
– sequence: 3
  givenname: Brian P
  surname: Hobbs
  fullname: Hobbs, Brian P
  organization: The University of Texas MD Anderson Cancer Center, 1515 Holcombe Blvd. Houston, TX 77030, USA
BackLink https://www.ncbi.nlm.nih.gov/pubmed/29036300$$D View this record in MEDLINE/PubMed
BookMark eNpNj0tLxDAUhYOMOA_9ByJduqmTR5O2Sx18wYAbXZfb5GYm2iZj00Ln31twBFfnW3ycw1mSmQ8eCblm9I7RUqxrF2IPvYu903H9NY5UsDOyYJkq0kzIfPaP52QZ4yelnAslLsicl3QCShdk8wBHjA58snfYQaf3TkOTtMFg4_wuqSGiSYJP2qGZtsLQaUxw1HvwO4TaNa4_XpJzC03Eq1OuyMfT4_vmJd2-Pb9u7repzrKiT41kFpiS2k7rhtZClBKVlbrMZQai1pQXCiZVCW1yiUZbU0xPNShurM35itz-9h668D1g7KvWRY1NAx7DECtWSs4oY7ma1JuTOtQtmurQuRa6Y_V3nP8AscZg9w
CitedBy_id crossref_primary_10_1093_biomtc_ujae118
crossref_primary_10_1093_biostatistics_kxz066
crossref_primary_10_1002_sim_9191
crossref_primary_10_1097_EDE_0000000000001690
crossref_primary_10_1016_j_cct_2021_106460
crossref_primary_10_1002_pst_2289
crossref_primary_10_1002_bimj_202200312
crossref_primary_10_1093_biomtc_ujae086
crossref_primary_10_1093_biomtc_ujae085
crossref_primary_10_1080_10543406_2023_2292211
crossref_primary_10_1177_0962280220975187
crossref_primary_10_1016_j_cct_2022_106972
crossref_primary_10_1080_10543406_2024_2358806
crossref_primary_10_1002_sim_10267
crossref_primary_10_1093_biostatistics_kxab008
crossref_primary_10_1002_sim_8946
crossref_primary_10_1080_19466315_2021_2008484
crossref_primary_10_1177_09622802251367439
crossref_primary_10_1177_17407745231158906
crossref_primary_10_1002_sim_8465
crossref_primary_10_1002_sim_9114
crossref_primary_10_1002_sim_7893
crossref_primary_10_1177_17407745211073624
crossref_primary_10_1002_sim_9514
crossref_primary_10_1111_rssc_12324
crossref_primary_10_1093_biostatistics_kxab051
crossref_primary_10_1002_sim_10077
crossref_primary_10_3390_cancers16020251
crossref_primary_10_1002_sim_10158
crossref_primary_10_1017_cts_2023_537
crossref_primary_10_1158_1078_0432_CCR_19_0820
crossref_primary_10_3390_math13162639
crossref_primary_10_1214_20_BA1229
crossref_primary_10_1002_pst_2464
crossref_primary_10_1080_10543406_2021_1998100
crossref_primary_10_1002_wics_1581
crossref_primary_10_1002_sim_70137
crossref_primary_10_1016_j_drudis_2024_104127
crossref_primary_10_1002_sim_9249
crossref_primary_10_1200_EDBK_319783
crossref_primary_10_1016_j_csda_2018_05_002
crossref_primary_10_1158_1078_0432_CCR_21_2124
crossref_primary_10_1093_biostatistics_kxaf003
crossref_primary_10_1093_biostatistics_kxad024
crossref_primary_10_1214_21_AOAS1469
crossref_primary_10_1002_bimj_202100287
crossref_primary_10_1111_biom_13927
crossref_primary_10_1371_journal_pone_0272367
crossref_primary_10_1016_j_cct_2024_107489
crossref_primary_10_1038_s41571_020_0384_0
crossref_primary_10_1002_sim_10006
crossref_primary_10_1002_bimj_202000314
crossref_primary_10_1111_biom_13253
crossref_primary_10_3389_fphar_2023_1266322
crossref_primary_10_1111_biom_13294
crossref_primary_10_1093_jnci_djy196
crossref_primary_10_1002_sim_9890
ContentType Journal Article
DBID CGR
CUY
CVF
ECM
EIF
NPM
7X8
DOI 10.1093/biostatistics/kxx031
DatabaseName Medline
MEDLINE
MEDLINE (Ovid)
MEDLINE
MEDLINE
PubMed
MEDLINE - Academic
DatabaseTitle MEDLINE
Medline Complete
MEDLINE with Full Text
PubMed
MEDLINE (Ovid)
MEDLINE - Academic
DatabaseTitleList MEDLINE - Academic
MEDLINE
Database_xml – sequence: 1
  dbid: NPM
  name: PubMed
  url: http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=PubMed
  sourceTypes: Index Database
– sequence: 2
  dbid: 7X8
  name: MEDLINE - Academic
  url: https://search.proquest.com/medline
  sourceTypes: Aggregation Database
DeliveryMethod no_fulltext_linktorsrc
Discipline Biology
EISSN 1468-4357
ExternalDocumentID 29036300
Genre Research Support, U.S. Gov't, P.H.S
Journal Article
Research Support, N.I.H., Extramural
GrantInformation_xml – fundername: NIDA NIH HHS
  grantid: U54 DA031659
– fundername: NIDA NIH HHS
  grantid: R03 DA041870
GroupedDBID ---
-E4
.2P
.I3
0R~
1TH
23N
2WC
4.4
48X
53G
5GY
5VS
5WA
6PF
70D
AAIJN
AAJKP
AAJQQ
AAMVS
AAOGV
AAPQZ
AAPXW
AARHZ
AAUAY
AAUQX
AAVAP
AAWTL
ABDFA
ABDTM
ABEJV
ABEUO
ABGNP
ABIXL
ABJNI
ABLJU
ABNKS
ABPQP
ABPTD
ABQLI
ABVGC
ABWST
ABXVV
ABZBJ
ACGFS
ACIWK
ACPRK
ACUFI
ACUXJ
ACYTK
ADBBV
ADEYI
ADEZT
ADGZP
ADHKW
ADHZD
ADIPN
ADNBA
ADOCK
ADQBN
ADRDM
ADRTK
ADVEK
ADYJX
ADYVW
ADZXQ
AECKG
AEGPL
AEJOX
AEKKA
AEKSI
AEMDU
AENEX
AENZO
AEPUE
AETBJ
AEWNT
AFFZL
AFIYH
AFOFC
AFRAH
AGINJ
AGKEF
AGORE
AGQXC
AGSYK
AHMBA
AHXPO
AIJHB
AJBYB
AJEEA
AJEUX
AJNCP
ALMA_UNASSIGNED_HOLDINGS
ALTZX
ALUQC
ALXQX
ANAKG
APIBT
APWMN
ATGXG
AXUDD
AZVOD
BAWUL
BAYMD
BCRHZ
BEYMZ
BHONS
BQUQU
BTQHN
C45
CDBKE
CGR
CS3
CUY
CVF
CZ4
DAKXR
DIK
DILTD
DU5
D~K
E3Z
EBD
EBS
ECM
EE~
EIF
EJD
EMOBN
F5P
F9B
FLIZI
FLUFQ
FOEOM
FQBLK
GAUVT
GJXCC
H13
H5~
HAR
HW0
HZ~
IOX
J21
JXSIZ
KBUDW
KOP
KQ8
KSI
KSN
M-Z
N9A
NGC
NMDNZ
NOMLY
NPM
NU-
O9-
ODMLO
OJQWA
OJZSN
OK1
OVD
P2P
PAFKI
PEELM
PQQKQ
Q1.
Q5Y
RD5
ROL
ROX
RUSNO
RW1
RXO
SV3
TEORI
TJP
TN5
TR2
W8F
WOQ
X7H
YAYTL
YKOAZ
YXANX
ZKX
~91
7X8
ID FETCH-LOGICAL-c448t-d51fa165cf903d0b3395e6f5c9754a3bc0286a44863cd75edcfd8093ca62dff72
IEDL.DBID 7X8
ISICitedReferencesCount 62
ISICitedReferencesURI http://www.webofscience.com/api/gateway?GWVersion=2&SrcApp=Summon&SrcAuth=ProQuest&DestLinkType=CitingArticles&DestApp=WOS_CPL&KeyUT=000429028200004&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
ISSN 1468-4357
IngestDate Sun Sep 28 10:04:23 EDT 2025
Mon Jul 21 06:07:30 EDT 2025
IsDoiOpenAccess false
IsOpenAccess true
IsPeerReviewed true
IsScholarly true
Issue 2
Language English
LinkModel DirectLink
MergedId FETCHMERGED-LOGICAL-c448t-d51fa165cf903d0b3395e6f5c9754a3bc0286a44863cd75edcfd8093ca62dff72
Notes ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 23
OpenAccessLink https://academic.oup.com/biostatistics/article-pdf/19/2/169/24268734/kxx031.pdf
PMID 29036300
PQID 1952101176
PQPubID 23479
ParticipantIDs proquest_miscellaneous_1952101176
pubmed_primary_29036300
PublicationCentury 2000
PublicationDate 2018-04-01
PublicationDateYYYYMMDD 2018-04-01
PublicationDate_xml – month: 04
  year: 2018
  text: 2018-04-01
  day: 01
PublicationDecade 2010
PublicationPlace England
PublicationPlace_xml – name: England
PublicationTitle Biostatistics (Oxford, England)
PublicationTitleAlternate Biostatistics
PublicationYear 2018
SSID ssj0022363
Score 2.447624
Snippet Bayesian hierarchical models produce shrinkage estimators that can be used as the basis for integrating supplementary data into the analysis of a primary data...
SourceID proquest
pubmed
SourceType Aggregation Database
Index Database
StartPage 169
SubjectTerms Bayes Theorem
Biostatistics - methods
Cigarette Smoking - prevention & control
Data Interpretation, Statistical
Humans
Models, Statistical
Nicotine
Outcome Assessment, Health Care - methods
Randomized Controlled Trials as Topic
Tobacco Products
Tobacco Use Disorder - prevention & control
Title Bayesian hierarchical modeling based on multisource exchangeability
URI https://www.ncbi.nlm.nih.gov/pubmed/29036300
https://www.proquest.com/docview/1952101176
Volume 19
WOSCitedRecordID wos000429028200004&url=https%3A%2F%2Fcvtisr.summon.serialssolutions.com%2F%23%21%2Fsearch%3Fho%3Df%26include.ft.matches%3Dt%26l%3Dnull%26q%3D
hasFullText
inHoldings 1
isFullTextHit
isPrint
link http://cvtisr.summon.serialssolutions.com/2.0.0/link/0/eLvHCXMwpV1JS8QwFA7qKHhxX8aNCF7LpE3SNCfRwcGDDnNQmVtJs4AI7eiozPx7X5KOXjwIXnJqoby-5Pvytg-hC-IkLVhWJcRQljAHi0qtTCzgkasKpmyYpfd0J4bDYjyWozbgNm3LKhdnYjioTaN9jLyXSgAaP8Asv5y8Jl41ymdXWwmNZdShQGW8V4vxdxYBkC8oqYXuIqAFYtE6J2mvem58x04chtx7mc1IKzP3K8kMYDPY_O9nbqGNlmbiq-gX22jJ1jtoLQpPzndR_1rNrW-fxF4KOyQT4F_hIIsDWIY9thnc1DjUG8YAP7az2CUcJ3vP99Dj4Oahf5u0cgqJhjvYe2J46lSac-0koYZUlEpuc8e1FJwpWmmgGrmCR3OqjeDWaGcKsJVWeWacE9k-Wqmb2h4inFWpJkZkSgrDuFKV5UwTIF-W-hsR76LzhXVKcFefg1C1bT6m5Y99uuggmricxLkaZSZ9VpmQoz-8fYzWgboUsYbmBHUcbFZ7ilb1J1jl7Sz4AazD0f0X8LO_Rg
linkProvider ProQuest
openUrl ctx_ver=Z39.88-2004&ctx_enc=info%3Aofi%2Fenc%3AUTF-8&rfr_id=info%3Asid%2Fsummon.serialssolutions.com&rft_val_fmt=info%3Aofi%2Ffmt%3Akev%3Amtx%3Ajournal&rft.genre=article&rft.atitle=Bayesian+hierarchical+modeling+based+on+multisource+exchangeability&rft.jtitle=Biostatistics+%28Oxford%2C+England%29&rft.au=Kaizer%2C+Alexander+M&rft.au=Koopmeiners%2C+Joseph+S&rft.au=Hobbs%2C+Brian+P&rft.date=2018-04-01&rft.eissn=1468-4357&rft.volume=19&rft.issue=2&rft.spage=169&rft_id=info:doi/10.1093%2Fbiostatistics%2Fkxx031&rft_id=info%3Apmid%2F29036300&rft_id=info%3Apmid%2F29036300&rft.externalDocID=29036300
thumbnail_l http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/lc.gif&issn=1468-4357&client=summon
thumbnail_m http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/mc.gif&issn=1468-4357&client=summon
thumbnail_s http://covers-cdn.summon.serialssolutions.com/index.aspx?isbn=/sc.gif&issn=1468-4357&client=summon