Loading…
Performance of instrumental variable methods in cohort and nested case-control studies: a simulation study
ABSTRACT Purpose Instrumental variable (IV) analysis is becoming increasingly popular to adjust for confounding in observational pharmacoepidemiologic research. One of the prerequisites of an IV is that it is strongly associated with exposure; if it is weakly associated with exposure, IV estimates a...
Saved in:
| Published in: | Pharmacoepidemiology and drug safety 2014-02, Vol.23 (2), p.165-177 |
|---|---|
| Main Authors: | , , , , , , |
| Format: | Article |
| Language: | English |
| Subjects: | |
| Citations: | Items that this one cites Items that cite this one |
| Online Access: | Get full text |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| cited_by | cdi_FETCH-LOGICAL-c4175-6937d880eb41388028e37858f6acf9194e6099377e90d952ee5f744b356ef57d3 |
|---|---|
| cites | cdi_FETCH-LOGICAL-c4175-6937d880eb41388028e37858f6acf9194e6099377e90d952ee5f744b356ef57d3 |
| container_end_page | 177 |
| container_issue | 2 |
| container_start_page | 165 |
| container_title | Pharmacoepidemiology and drug safety |
| container_volume | 23 |
| creator | Uddin, Md. Jamal Groenwold, Rolf H. H. de Boer, Anthonius Belitser, Svetlana V. Roes, Kit C. B. Hoes, Arno W. Klungel, Olaf H. |
| description | ABSTRACT
Purpose
Instrumental variable (IV) analysis is becoming increasingly popular to adjust for confounding in observational pharmacoepidemiologic research. One of the prerequisites of an IV is that it is strongly associated with exposure; if it is weakly associated with exposure, IV estimates are reported to be biased. We aimed to assess the performance of IV estimates in various (pharmaco‐)epidemiologic settings.
Methods
Data were simulated for continuous/binary exposure, outcome and IV in cohort and nested case–control (NCC) designs with different incidences of the outcome. Pearson's correlation, point bi‐serial correlation, odds ratio (OR), and F‐statistic were used to assess the IV‐exposure association. Two‐stage analysis was performed to estimate the exposure effect.
Results
For all types of IV and exposure in the cohort and NCC designs, IV estimates were extremely unstable and biased when the IV was very weakly associated with exposure (e.g. Pearson's correlation |
| doi_str_mv | 10.1002/pds.3555 |
| format | article |
| fullrecord | <record><control><sourceid>proquest_cross</sourceid><recordid>TN_cdi_proquest_miscellaneous_1492685090</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>3188238441</sourcerecordid><originalsourceid>FETCH-LOGICAL-c4175-6937d880eb41388028e37858f6acf9194e6099377e90d952ee5f744b356ef57d3</originalsourceid><addsrcrecordid>eNp1kV1rFDEUhoNYbK2Cv0ACIngzbTL5mngnVddCqYUqXoZs5gzNmknWZEbdf2_Wji0IvToh5-E9J08QekHJCSWkPd325YQJIR6hI0q0bqgQ6vH-LFjTCakP0dNSNoTUnuZP0GHLGZFaiiO0uYI8pDza6ACnAftYpjyPECcb8E-bvV0HwCNMN6kvtYtdukl5wjb2OEKZoMfOFmhcilNOAZdp7j2Ut9ji4sc52Mmn-Pd29wwdDDYUeL7UY_T144cvZ5-ai8-r87N3F43jVIlGaqb6riOw5pTV2nbAVCe6QVo3aKo5yPoMphRo0mvRAohBcb5mQsIgVM-O0Zvb3G1OP-a6oxl9cRCCjZDmYijXrewE0aSir_5DN2nOsW63pyiRgkp6H-hyKiXDYLbZjzbvDCVm799U_2bvv6Ivl8B5PUJ_B_4TXoHXC2CLs2HIVbwv91xHJWOcV6655X75ALsHB5qr99fL4IX39U9-3_E2fzdSMSXMt8uVudYrpdQlN4z9AXWqqgU</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>1491065161</pqid></control><display><type>article</type><title>Performance of instrumental variable methods in cohort and nested case-control studies: a simulation study</title><source>Wiley</source><creator>Uddin, Md. Jamal ; Groenwold, Rolf H. H. ; de Boer, Anthonius ; Belitser, Svetlana V. ; Roes, Kit C. B. ; Hoes, Arno W. ; Klungel, Olaf H.</creator><creatorcontrib>Uddin, Md. Jamal ; Groenwold, Rolf H. H. ; de Boer, Anthonius ; Belitser, Svetlana V. ; Roes, Kit C. B. ; Hoes, Arno W. ; Klungel, Olaf H.</creatorcontrib><description>ABSTRACT
Purpose
Instrumental variable (IV) analysis is becoming increasingly popular to adjust for confounding in observational pharmacoepidemiologic research. One of the prerequisites of an IV is that it is strongly associated with exposure; if it is weakly associated with exposure, IV estimates are reported to be biased. We aimed to assess the performance of IV estimates in various (pharmaco‐)epidemiologic settings.
Methods
Data were simulated for continuous/binary exposure, outcome and IV in cohort and nested case–control (NCC) designs with different incidences of the outcome. Pearson's correlation, point bi‐serial correlation, odds ratio (OR), and F‐statistic were used to assess the IV‐exposure association. Two‐stage analysis was performed to estimate the exposure effect.
Results
For all types of IV and exposure in the cohort and NCC designs, IV estimates were extremely unstable and biased when the IV was very weakly associated with exposure (e.g. Pearson's correlation < 0.15 for continuous or OR < 2.0 for binary IV and exposure; although specific cut‐off values depend on simulation settings). For stronger IVs, estimates were unbiased and become less variable compared with weaker IVs in the case of continuous and binary (risk difference scale) outcomes. For a similar IV‐exposure association (e.g. OR = 1.4 and 5% incidence of the outcome), the variability of the estimates was more pronounced in the NCC (standard deviation = 2.37, case : control = 1:5) compared with the cohort design (standard deviation = 1.14). The variability was even more pronounced for rare (≤1%) outcomes. However, IV estimates from the NCC design became less variable with an increasing number of controls per case. Moreover, estimates were biased when the IV was related to confounders even with strong IVs.
Conclusions
Instrumental variable analysis performs poorly when the IV‐exposure association is extremely weak, especially in the NCC design. IV estimates in the NCC design become less variable when the number of control increases. As NCC does not use the entire cohort, in order to achieve stable estimates, this design requires a stronger IV‐exposure association than the cohort design. Copyright © 2013 John Wiley & Sons, Ltd.</description><identifier>ISSN: 1053-8569</identifier><identifier>EISSN: 1099-1557</identifier><identifier>DOI: 10.1002/pds.3555</identifier><identifier>PMID: 24306965</identifier><identifier>CODEN: PDSAEA</identifier><language>eng</language><publisher>Chichester: Blackwell Publishing Ltd</publisher><subject>Bias ; Biological and medical sciences ; Case-Control Studies ; Clinical trial. Drug monitoring ; cohort ; Cohort Studies ; Computer Simulation ; confounding in epidemiology ; Economic models ; Epidemiologic Methods ; Epidemiology ; General pharmacology ; Humans ; instrumental variable ; Medical sciences ; nested case-control ; pharmacoepidemiology ; Pharmacoepidemiology - methods ; Pharmacology ; Pharmacology. Drug treatments ; rare outcome ; Research Design ; simulation ; variability ; weak IV</subject><ispartof>Pharmacoepidemiology and drug safety, 2014-02, Vol.23 (2), p.165-177</ispartof><rights>Copyright © 2013 John Wiley & Sons, Ltd.</rights><rights>2015 INIST-CNRS</rights><rights>Copyright © 2014 John Wiley & Sons, Ltd.</rights><lds50>peer_reviewed</lds50><woscitedreferencessubscribed>false</woscitedreferencessubscribed><citedby>FETCH-LOGICAL-c4175-6937d880eb41388028e37858f6acf9194e6099377e90d952ee5f744b356ef57d3</citedby><cites>FETCH-LOGICAL-c4175-6937d880eb41388028e37858f6acf9194e6099377e90d952ee5f744b356ef57d3</cites></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>314,777,781,27905,27906</link.rule.ids><backlink>$$Uhttp://pascal-francis.inist.fr/vibad/index.php?action=getRecordDetail&idt=28163344$$DView record in Pascal Francis$$Hfree_for_read</backlink><backlink>$$Uhttps://www.ncbi.nlm.nih.gov/pubmed/24306965$$D View this record in MEDLINE/PubMed$$Hfree_for_read</backlink></links><search><creatorcontrib>Uddin, Md. Jamal</creatorcontrib><creatorcontrib>Groenwold, Rolf H. H.</creatorcontrib><creatorcontrib>de Boer, Anthonius</creatorcontrib><creatorcontrib>Belitser, Svetlana V.</creatorcontrib><creatorcontrib>Roes, Kit C. B.</creatorcontrib><creatorcontrib>Hoes, Arno W.</creatorcontrib><creatorcontrib>Klungel, Olaf H.</creatorcontrib><title>Performance of instrumental variable methods in cohort and nested case-control studies: a simulation study</title><title>Pharmacoepidemiology and drug safety</title><addtitle>Pharmacoepidemiol Drug Saf</addtitle><description>ABSTRACT
Purpose
Instrumental variable (IV) analysis is becoming increasingly popular to adjust for confounding in observational pharmacoepidemiologic research. One of the prerequisites of an IV is that it is strongly associated with exposure; if it is weakly associated with exposure, IV estimates are reported to be biased. We aimed to assess the performance of IV estimates in various (pharmaco‐)epidemiologic settings.
Methods
Data were simulated for continuous/binary exposure, outcome and IV in cohort and nested case–control (NCC) designs with different incidences of the outcome. Pearson's correlation, point bi‐serial correlation, odds ratio (OR), and F‐statistic were used to assess the IV‐exposure association. Two‐stage analysis was performed to estimate the exposure effect.
Results
For all types of IV and exposure in the cohort and NCC designs, IV estimates were extremely unstable and biased when the IV was very weakly associated with exposure (e.g. Pearson's correlation < 0.15 for continuous or OR < 2.0 for binary IV and exposure; although specific cut‐off values depend on simulation settings). For stronger IVs, estimates were unbiased and become less variable compared with weaker IVs in the case of continuous and binary (risk difference scale) outcomes. For a similar IV‐exposure association (e.g. OR = 1.4 and 5% incidence of the outcome), the variability of the estimates was more pronounced in the NCC (standard deviation = 2.37, case : control = 1:5) compared with the cohort design (standard deviation = 1.14). The variability was even more pronounced for rare (≤1%) outcomes. However, IV estimates from the NCC design became less variable with an increasing number of controls per case. Moreover, estimates were biased when the IV was related to confounders even with strong IVs.
Conclusions
Instrumental variable analysis performs poorly when the IV‐exposure association is extremely weak, especially in the NCC design. IV estimates in the NCC design become less variable when the number of control increases. As NCC does not use the entire cohort, in order to achieve stable estimates, this design requires a stronger IV‐exposure association than the cohort design. Copyright © 2013 John Wiley & Sons, Ltd.</description><subject>Bias</subject><subject>Biological and medical sciences</subject><subject>Case-Control Studies</subject><subject>Clinical trial. Drug monitoring</subject><subject>cohort</subject><subject>Cohort Studies</subject><subject>Computer Simulation</subject><subject>confounding in epidemiology</subject><subject>Economic models</subject><subject>Epidemiologic Methods</subject><subject>Epidemiology</subject><subject>General pharmacology</subject><subject>Humans</subject><subject>instrumental variable</subject><subject>Medical sciences</subject><subject>nested case-control</subject><subject>pharmacoepidemiology</subject><subject>Pharmacoepidemiology - methods</subject><subject>Pharmacology</subject><subject>Pharmacology. Drug treatments</subject><subject>rare outcome</subject><subject>Research Design</subject><subject>simulation</subject><subject>variability</subject><subject>weak IV</subject><issn>1053-8569</issn><issn>1099-1557</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2014</creationdate><recordtype>article</recordtype><recordid>eNp1kV1rFDEUhoNYbK2Cv0ACIngzbTL5mngnVddCqYUqXoZs5gzNmknWZEbdf2_Wji0IvToh5-E9J08QekHJCSWkPd325YQJIR6hI0q0bqgQ6vH-LFjTCakP0dNSNoTUnuZP0GHLGZFaiiO0uYI8pDza6ACnAftYpjyPECcb8E-bvV0HwCNMN6kvtYtdukl5wjb2OEKZoMfOFmhcilNOAZdp7j2Ut9ji4sc52Mmn-Pd29wwdDDYUeL7UY_T144cvZ5-ai8-r87N3F43jVIlGaqb6riOw5pTV2nbAVCe6QVo3aKo5yPoMphRo0mvRAohBcb5mQsIgVM-O0Zvb3G1OP-a6oxl9cRCCjZDmYijXrewE0aSir_5DN2nOsW63pyiRgkp6H-hyKiXDYLbZjzbvDCVm799U_2bvv6Ivl8B5PUJ_B_4TXoHXC2CLs2HIVbwv91xHJWOcV6655X75ALsHB5qr99fL4IX39U9-3_E2fzdSMSXMt8uVudYrpdQlN4z9AXWqqgU</recordid><startdate>201402</startdate><enddate>201402</enddate><creator>Uddin, Md. Jamal</creator><creator>Groenwold, Rolf H. H.</creator><creator>de Boer, Anthonius</creator><creator>Belitser, Svetlana V.</creator><creator>Roes, Kit C. B.</creator><creator>Hoes, Arno W.</creator><creator>Klungel, Olaf H.</creator><general>Blackwell Publishing Ltd</general><general>Wiley</general><general>Wiley Subscription Services, Inc</general><scope>BSCLL</scope><scope>IQODW</scope><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>7TK</scope><scope>K9.</scope><scope>7X8</scope></search><sort><creationdate>201402</creationdate><title>Performance of instrumental variable methods in cohort and nested case-control studies: a simulation study</title><author>Uddin, Md. Jamal ; Groenwold, Rolf H. H. ; de Boer, Anthonius ; Belitser, Svetlana V. ; Roes, Kit C. B. ; Hoes, Arno W. ; Klungel, Olaf H.</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c4175-6937d880eb41388028e37858f6acf9194e6099377e90d952ee5f744b356ef57d3</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2014</creationdate><topic>Bias</topic><topic>Biological and medical sciences</topic><topic>Case-Control Studies</topic><topic>Clinical trial. Drug monitoring</topic><topic>cohort</topic><topic>Cohort Studies</topic><topic>Computer Simulation</topic><topic>confounding in epidemiology</topic><topic>Economic models</topic><topic>Epidemiologic Methods</topic><topic>Epidemiology</topic><topic>General pharmacology</topic><topic>Humans</topic><topic>instrumental variable</topic><topic>Medical sciences</topic><topic>nested case-control</topic><topic>pharmacoepidemiology</topic><topic>Pharmacoepidemiology - methods</topic><topic>Pharmacology</topic><topic>Pharmacology. Drug treatments</topic><topic>rare outcome</topic><topic>Research Design</topic><topic>simulation</topic><topic>variability</topic><topic>weak IV</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Uddin, Md. Jamal</creatorcontrib><creatorcontrib>Groenwold, Rolf H. H.</creatorcontrib><creatorcontrib>de Boer, Anthonius</creatorcontrib><creatorcontrib>Belitser, Svetlana V.</creatorcontrib><creatorcontrib>Roes, Kit C. B.</creatorcontrib><creatorcontrib>Hoes, Arno W.</creatorcontrib><creatorcontrib>Klungel, Olaf H.</creatorcontrib><collection>Istex</collection><collection>Pascal-Francis</collection><collection>Medline</collection><collection>MEDLINE</collection><collection>MEDLINE (Ovid)</collection><collection>MEDLINE</collection><collection>MEDLINE</collection><collection>PubMed</collection><collection>CrossRef</collection><collection>Neurosciences Abstracts</collection><collection>ProQuest Health & Medical Complete (Alumni)</collection><collection>MEDLINE - Academic</collection><jtitle>Pharmacoepidemiology and drug safety</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Uddin, Md. Jamal</au><au>Groenwold, Rolf H. H.</au><au>de Boer, Anthonius</au><au>Belitser, Svetlana V.</au><au>Roes, Kit C. B.</au><au>Hoes, Arno W.</au><au>Klungel, Olaf H.</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Performance of instrumental variable methods in cohort and nested case-control studies: a simulation study</atitle><jtitle>Pharmacoepidemiology and drug safety</jtitle><addtitle>Pharmacoepidemiol Drug Saf</addtitle><date>2014-02</date><risdate>2014</risdate><volume>23</volume><issue>2</issue><spage>165</spage><epage>177</epage><pages>165-177</pages><issn>1053-8569</issn><eissn>1099-1557</eissn><coden>PDSAEA</coden><abstract>ABSTRACT
Purpose
Instrumental variable (IV) analysis is becoming increasingly popular to adjust for confounding in observational pharmacoepidemiologic research. One of the prerequisites of an IV is that it is strongly associated with exposure; if it is weakly associated with exposure, IV estimates are reported to be biased. We aimed to assess the performance of IV estimates in various (pharmaco‐)epidemiologic settings.
Methods
Data were simulated for continuous/binary exposure, outcome and IV in cohort and nested case–control (NCC) designs with different incidences of the outcome. Pearson's correlation, point bi‐serial correlation, odds ratio (OR), and F‐statistic were used to assess the IV‐exposure association. Two‐stage analysis was performed to estimate the exposure effect.
Results
For all types of IV and exposure in the cohort and NCC designs, IV estimates were extremely unstable and biased when the IV was very weakly associated with exposure (e.g. Pearson's correlation < 0.15 for continuous or OR < 2.0 for binary IV and exposure; although specific cut‐off values depend on simulation settings). For stronger IVs, estimates were unbiased and become less variable compared with weaker IVs in the case of continuous and binary (risk difference scale) outcomes. For a similar IV‐exposure association (e.g. OR = 1.4 and 5% incidence of the outcome), the variability of the estimates was more pronounced in the NCC (standard deviation = 2.37, case : control = 1:5) compared with the cohort design (standard deviation = 1.14). The variability was even more pronounced for rare (≤1%) outcomes. However, IV estimates from the NCC design became less variable with an increasing number of controls per case. Moreover, estimates were biased when the IV was related to confounders even with strong IVs.
Conclusions
Instrumental variable analysis performs poorly when the IV‐exposure association is extremely weak, especially in the NCC design. IV estimates in the NCC design become less variable when the number of control increases. As NCC does not use the entire cohort, in order to achieve stable estimates, this design requires a stronger IV‐exposure association than the cohort design. Copyright © 2013 John Wiley & Sons, Ltd.</abstract><cop>Chichester</cop><pub>Blackwell Publishing Ltd</pub><pmid>24306965</pmid><doi>10.1002/pds.3555</doi><tpages>13</tpages></addata></record> |
| fulltext | fulltext |
| identifier | ISSN: 1053-8569 |
| ispartof | Pharmacoepidemiology and drug safety, 2014-02, Vol.23 (2), p.165-177 |
| issn | 1053-8569 1099-1557 |
| language | eng |
| recordid | cdi_proquest_miscellaneous_1492685090 |
| source | Wiley |
| subjects | Bias Biological and medical sciences Case-Control Studies Clinical trial. Drug monitoring cohort Cohort Studies Computer Simulation confounding in epidemiology Economic models Epidemiologic Methods Epidemiology General pharmacology Humans instrumental variable Medical sciences nested case-control pharmacoepidemiology Pharmacoepidemiology - methods Pharmacology Pharmacology. Drug treatments rare outcome Research Design simulation variability weak IV |
| title | Performance of instrumental variable methods in cohort and nested case-control studies: a simulation study |
| url | http://sfxeu10.hosted.exlibrisgroup.com/loughborough?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-20T05%3A42%3A42IST&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=Performance%20of%20instrumental%20variable%20methods%20in%20cohort%20and%20nested%20case-control%20studies:%20a%20simulation%20study&rft.jtitle=Pharmacoepidemiology%20and%20drug%20safety&rft.au=Uddin,%20Md.%20Jamal&rft.date=2014-02&rft.volume=23&rft.issue=2&rft.spage=165&rft.epage=177&rft.pages=165-177&rft.issn=1053-8569&rft.eissn=1099-1557&rft.coden=PDSAEA&rft_id=info:doi/10.1002/pds.3555&rft_dat=%3Cproquest_cross%3E3188238441%3C/proquest_cross%3E%3Cgrp_id%3Ecdi_FETCH-LOGICAL-c4175-6937d880eb41388028e37858f6acf9194e6099377e90d952ee5f744b356ef57d3%3C/grp_id%3E%3Coa%3E%3C/oa%3E%3Curl%3E%3C/url%3E&rft_id=info:oai/&rft_pqid=1491065161&rft_id=info:pmid/24306965&rfr_iscdi=true |