Bias of using odds ratio estimates in multinomial logistic regressions to estimate relative risk or prevalence ratio and alternatives

Detalhes bibliográficos
Autor(a) principal: Suzi Alves Camey
Data de Publicação: 2014
Outros Autores: Vanessa Bielefeldt Leotti Torman, Vania Naomi Hirakata, Renan Xavier Cortes, Alvaro Vigo
Tipo de documento: Artigo
Idioma: eng
Título da fonte: Cadernos de Saúde Pública
Texto Completo: https://cadernos.ensp.fiocruz.br/ojs/index.php/csp/article/view/5672
Resumo: Recent studies have emphasized that there is no justification for using the odds ratio (OR) as an approximation of the relative risk (RR) or prevalence ratio (PR). Erroneous interpretations of the OR as RR or PR must be avoided, as several studies have shown that the OR is not a good approximation for these measures when the outcome is common (> 10%). For multinomial outcomes it is usual to use the multinomial logistic regression. In this context, there are no studies showing the impact of the approximation of the OR in the estimates of RR or PR. This study aimed to present and discuss alternative methods to multinomial logistic regression based upon robust Poisson regression and the log-binomial model. The approaches were compared by simulating various possible scenarios. The results showed that the proposed models have more precise and accurate estimates for the RR or PR than the multinomial logistic regression, as in the case of the binary outcome. Thus also for multinomial outcomes the OR must not be used as an approximation of the RR or PR, since this may lead to incorrect conclusions.
id FIOCRUZ-5_a586b48678a79ec36d5315bbcd65a4db
oai_identifier_str oai:ojs.teste-cadernos.ensp.fiocruz.br:article/5672
network_acronym_str FIOCRUZ-5
network_name_str Cadernos de Saúde Pública
repository_id_str
spelling Bias of using odds ratio estimates in multinomial logistic regressions to estimate relative risk or prevalence ratio and alternativesOdds RatioPrevalence RatioLogistic ModelsRelative RiskRecent studies have emphasized that there is no justification for using the odds ratio (OR) as an approximation of the relative risk (RR) or prevalence ratio (PR). Erroneous interpretations of the OR as RR or PR must be avoided, as several studies have shown that the OR is not a good approximation for these measures when the outcome is common (> 10%). For multinomial outcomes it is usual to use the multinomial logistic regression. In this context, there are no studies showing the impact of the approximation of the OR in the estimates of RR or PR. This study aimed to present and discuss alternative methods to multinomial logistic regression based upon robust Poisson regression and the log-binomial model. The approaches were compared by simulating various possible scenarios. The results showed that the proposed models have more precise and accurate estimates for the RR or PR than the multinomial logistic regression, as in the case of the binary outcome. Thus also for multinomial outcomes the OR must not be used as an approximation of the RR or PR, since this may lead to incorrect conclusions.Los trabajos recientes han enfatizado que ya no se justifica el uso del odds ratio (OR) como una aproximación del riesgo relativo (RR) o razón de prevalencias (RP). El OR no puede ser interpretado como RR o RP, pues varios estudios han demostrado que el OR no es una buena aproximación cuando el suceso es común (> 10%). Para sucesos multinomiales se utiliza comúnmente la regresión logística multinomial. En este contexto, no hay estudios que demuestren el impacto de la aproximación del OR en las estimaciones de RR o RP. Nuestro objetivo es presentar y discutir métodos alternativos a la regresión logística multinomial, en base a la regresión de Poisson y al modelo log-binomial. Los enfoques utilizados fueron comparados en un estudio de simulación con diferentes escenarios. Así como en el caso de suceso binario, los modelos propuestos dieron como resultado estimaciones para RR o RP más precisas y esmeradas que la regresión logística multinomial. Para sucesos multinomiales el OR tampoco debe ser utilizado como aproximación del RR o de la RP, pues se puede llegar a conclusiones incorrectas.Recentes trabalhos têm enfatizado que já não há justificativa para o uso da razão de chances (RC) como aproximação do risco relativo (RR) ou razão de prevalência (RP). Deve-se evitar a interpretação equivocada da RC como RR ou RP, pois vários estudos demonstraram que a RC não é uma boa aproximação para tais medidas quando o desfecho é comum (> 10%). Para desfechos multinomiais é usual aplicar a regressão logística multinomial. Nesse contexto, não há estudos demonstrando o impacto da aproximação da RC nas estimativas de RR ou RP. O objetivo deste trabalho é apresentar e discutir métodos alternativos à regressão logística multinomial, baseados na regressão de Poisson e no modelo log-binomial. As abordagens foram comparadas por um estudo de simulação com diversos cenários. Assim como no caso do desfecho binário, os modelos propostos apresentaram estimativas mais precisas e acuradas para o RR ou RP do que a regressão logística multinomial. Então, também para os desfechos multinomiais não se deve utilizar a RC como aproximação do RR ou RP, pois conclusões incorretas podem ocorrer.Reports in Public HealthCadernos de Saúde Pública2014-01-01info:eu-repo/semantics/articleinfo:eu-repo/semantics/publishedVersiontext/htmlapplication/pdfhttps://cadernos.ensp.fiocruz.br/ojs/index.php/csp/article/view/5672Reports in Public Health; Vol. 30 No. 1 (2014): JanuaryCadernos de Saúde Pública; v. 30 n. 1 (2014): Janeiro1678-44640102-311Xreponame:Cadernos de Saúde Públicainstname:Fundação Oswaldo Cruz (FIOCRUZ)instacron:FIOCRUZenghttps://cadernos.ensp.fiocruz.br/ojs/index.php/csp/article/view/5672/11799https://cadernos.ensp.fiocruz.br/ojs/index.php/csp/article/view/5672/11800Suzi Alves CameyVanessa Bielefeldt Leotti TormanVania Naomi HirakataRenan Xavier CortesAlvaro Vigoinfo:eu-repo/semantics/openAccess2024-03-06T15:28:51Zoai:ojs.teste-cadernos.ensp.fiocruz.br:article/5672Revistahttps://cadernos.ensp.fiocruz.br/ojs/index.php/csphttps://cadernos.ensp.fiocruz.br/ojs/index.php/csp/oaicadernos@ensp.fiocruz.br||cadernos@ensp.fiocruz.br1678-44640102-311Xopendoar:2024-03-06T13:06:29.528300Cadernos de Saúde Pública - Fundação Oswaldo Cruz (FIOCRUZ)true
dc.title.none.fl_str_mv Bias of using odds ratio estimates in multinomial logistic regressions to estimate relative risk or prevalence ratio and alternatives
title Bias of using odds ratio estimates in multinomial logistic regressions to estimate relative risk or prevalence ratio and alternatives
spellingShingle Bias of using odds ratio estimates in multinomial logistic regressions to estimate relative risk or prevalence ratio and alternatives
Suzi Alves Camey
Odds Ratio
Prevalence Ratio
Logistic Models
Relative Risk
title_short Bias of using odds ratio estimates in multinomial logistic regressions to estimate relative risk or prevalence ratio and alternatives
title_full Bias of using odds ratio estimates in multinomial logistic regressions to estimate relative risk or prevalence ratio and alternatives
title_fullStr Bias of using odds ratio estimates in multinomial logistic regressions to estimate relative risk or prevalence ratio and alternatives
title_full_unstemmed Bias of using odds ratio estimates in multinomial logistic regressions to estimate relative risk or prevalence ratio and alternatives
title_sort Bias of using odds ratio estimates in multinomial logistic regressions to estimate relative risk or prevalence ratio and alternatives
author Suzi Alves Camey
author_facet Suzi Alves Camey
Vanessa Bielefeldt Leotti Torman
Vania Naomi Hirakata
Renan Xavier Cortes
Alvaro Vigo
author_role author
author2 Vanessa Bielefeldt Leotti Torman
Vania Naomi Hirakata
Renan Xavier Cortes
Alvaro Vigo
author2_role author
author
author
author
dc.contributor.author.fl_str_mv Suzi Alves Camey
Vanessa Bielefeldt Leotti Torman
Vania Naomi Hirakata
Renan Xavier Cortes
Alvaro Vigo
dc.subject.por.fl_str_mv Odds Ratio
Prevalence Ratio
Logistic Models
Relative Risk
topic Odds Ratio
Prevalence Ratio
Logistic Models
Relative Risk
description Recent studies have emphasized that there is no justification for using the odds ratio (OR) as an approximation of the relative risk (RR) or prevalence ratio (PR). Erroneous interpretations of the OR as RR or PR must be avoided, as several studies have shown that the OR is not a good approximation for these measures when the outcome is common (> 10%). For multinomial outcomes it is usual to use the multinomial logistic regression. In this context, there are no studies showing the impact of the approximation of the OR in the estimates of RR or PR. This study aimed to present and discuss alternative methods to multinomial logistic regression based upon robust Poisson regression and the log-binomial model. The approaches were compared by simulating various possible scenarios. The results showed that the proposed models have more precise and accurate estimates for the RR or PR than the multinomial logistic regression, as in the case of the binary outcome. Thus also for multinomial outcomes the OR must not be used as an approximation of the RR or PR, since this may lead to incorrect conclusions.
publishDate 2014
dc.date.none.fl_str_mv 2014-01-01
dc.type.driver.fl_str_mv info:eu-repo/semantics/article
info:eu-repo/semantics/publishedVersion
format article
status_str publishedVersion
dc.identifier.uri.fl_str_mv https://cadernos.ensp.fiocruz.br/ojs/index.php/csp/article/view/5672
url https://cadernos.ensp.fiocruz.br/ojs/index.php/csp/article/view/5672
dc.language.iso.fl_str_mv eng
language eng
dc.relation.none.fl_str_mv https://cadernos.ensp.fiocruz.br/ojs/index.php/csp/article/view/5672/11799
https://cadernos.ensp.fiocruz.br/ojs/index.php/csp/article/view/5672/11800
dc.rights.driver.fl_str_mv info:eu-repo/semantics/openAccess
eu_rights_str_mv openAccess
dc.format.none.fl_str_mv text/html
application/pdf
dc.publisher.none.fl_str_mv Reports in Public Health
Cadernos de Saúde Pública
publisher.none.fl_str_mv Reports in Public Health
Cadernos de Saúde Pública
dc.source.none.fl_str_mv Reports in Public Health; Vol. 30 No. 1 (2014): January
Cadernos de Saúde Pública; v. 30 n. 1 (2014): Janeiro
1678-4464
0102-311X
reponame:Cadernos de Saúde Pública
instname:Fundação Oswaldo Cruz (FIOCRUZ)
instacron:FIOCRUZ
instname_str Fundação Oswaldo Cruz (FIOCRUZ)
instacron_str FIOCRUZ
institution FIOCRUZ
reponame_str Cadernos de Saúde Pública
collection Cadernos de Saúde Pública
repository.name.fl_str_mv Cadernos de Saúde Pública - Fundação Oswaldo Cruz (FIOCRUZ)
repository.mail.fl_str_mv cadernos@ensp.fiocruz.br||cadernos@ensp.fiocruz.br
_version_ 1798943377917476864

Registros relacionados