A comparison of the Polytomous logistic regression and joint Cox proportional hazards models for evaluating multiple disease subtypes in prospective cohort studies

Xiaonan (Nan) Xue, Mimi Kim, Mia M. Gaudet, Yikyung Park, Moonseong Heo, Albert R. Hollenbeck, Howard Strickler, Marc J. Gunter

Research output: Contribution to journalArticle

21 Citations (Scopus)

Abstract

Background: Polytomous logistic regression models are commonly used in case-control studies of cancer to directly compare the risks associated with an exposure variable across multiple cancer subtypes. However, the validity, accuracy, and efficiency of this approach for prospective cohort studies have not been formally evaluated. Methods: We investigated the performance of the polytomous logistic regression model and compared it with an alternative approach based on a joint Cox proportional hazards model using simulation studies. We then applied both methods to a prospective cohort study to assess whether the association of breast cancer with body size differs according to estrogen and progesterone receptor-defined subtypes. Results: Our simulations showed that the polytomous logistic regression model but not the joint Cox regression model yielded biased results in comparing exposure and disease subtype associations when the baseline hazards for different disease subtypes are nonproportional. For this reason, an analysis of a real data set was based on the joint Cox proportional hazards model and showed that body size has a significantly greater association with estrogen- and progesterone-positive breast cancer than with other subtypes. Conclusions: Because of the limitations of the polytomous logistic regression model for the comparison of exposure-disease associations across disease subtypes, the joint Cox proportional hazards model is recommended over the polytomous logistic regression model in prospective cohort studies. Impact: The article will promote the use of the joint Cox model in a prospective cohort study. Examples of SAS and S-plus programming codes are provided to facilitate use by nonstatisticians.

Original languageEnglish (US)
Pages (from-to)275-285
Number of pages11
JournalCancer Epidemiology Biomarkers and Prevention
Volume22
Issue number2
DOIs
StatePublished - Feb 2013

Fingerprint

Proportional Hazards Models
Cohort Studies
Joints
Logistic Models
Prospective Studies
Body Size
Breast Neoplasms
Joint Diseases
Progesterone Receptors
Estrogen Receptors
Progesterone
Case-Control Studies
Neoplasms
Estrogens

ASJC Scopus subject areas

  • Epidemiology
  • Oncology

Cite this

A comparison of the Polytomous logistic regression and joint Cox proportional hazards models for evaluating multiple disease subtypes in prospective cohort studies. / Xue, Xiaonan (Nan); Kim, Mimi; Gaudet, Mia M.; Park, Yikyung; Heo, Moonseong; Hollenbeck, Albert R.; Strickler, Howard; Gunter, Marc J.

In: Cancer Epidemiology Biomarkers and Prevention, Vol. 22, No. 2, 02.2013, p. 275-285.

Research output: Contribution to journalArticle

@article{9127ca5778c94297a08a76644265b440,
title = "A comparison of the Polytomous logistic regression and joint Cox proportional hazards models for evaluating multiple disease subtypes in prospective cohort studies",
abstract = "Background: Polytomous logistic regression models are commonly used in case-control studies of cancer to directly compare the risks associated with an exposure variable across multiple cancer subtypes. However, the validity, accuracy, and efficiency of this approach for prospective cohort studies have not been formally evaluated. Methods: We investigated the performance of the polytomous logistic regression model and compared it with an alternative approach based on a joint Cox proportional hazards model using simulation studies. We then applied both methods to a prospective cohort study to assess whether the association of breast cancer with body size differs according to estrogen and progesterone receptor-defined subtypes. Results: Our simulations showed that the polytomous logistic regression model but not the joint Cox regression model yielded biased results in comparing exposure and disease subtype associations when the baseline hazards for different disease subtypes are nonproportional. For this reason, an analysis of a real data set was based on the joint Cox proportional hazards model and showed that body size has a significantly greater association with estrogen- and progesterone-positive breast cancer than with other subtypes. Conclusions: Because of the limitations of the polytomous logistic regression model for the comparison of exposure-disease associations across disease subtypes, the joint Cox proportional hazards model is recommended over the polytomous logistic regression model in prospective cohort studies. Impact: The article will promote the use of the joint Cox model in a prospective cohort study. Examples of SAS and S-plus programming codes are provided to facilitate use by nonstatisticians.",
author = "Xue, {Xiaonan (Nan)} and Mimi Kim and Gaudet, {Mia M.} and Yikyung Park and Moonseong Heo and Hollenbeck, {Albert R.} and Howard Strickler and Gunter, {Marc J.}",
year = "2013",
month = "2",
doi = "10.1158/1055-9965.EPI-12-1050",
language = "English (US)",
volume = "22",
pages = "275--285",
journal = "Cancer Epidemiology Biomarkers and Prevention",
issn = "1055-9965",
publisher = "American Association for Cancer Research Inc.",
number = "2",

}

TY - JOUR

T1 - A comparison of the Polytomous logistic regression and joint Cox proportional hazards models for evaluating multiple disease subtypes in prospective cohort studies

AU - Xue, Xiaonan (Nan)

AU - Kim, Mimi

AU - Gaudet, Mia M.

AU - Park, Yikyung

AU - Heo, Moonseong

AU - Hollenbeck, Albert R.

AU - Strickler, Howard

AU - Gunter, Marc J.

PY - 2013/2

Y1 - 2013/2

N2 - Background: Polytomous logistic regression models are commonly used in case-control studies of cancer to directly compare the risks associated with an exposure variable across multiple cancer subtypes. However, the validity, accuracy, and efficiency of this approach for prospective cohort studies have not been formally evaluated. Methods: We investigated the performance of the polytomous logistic regression model and compared it with an alternative approach based on a joint Cox proportional hazards model using simulation studies. We then applied both methods to a prospective cohort study to assess whether the association of breast cancer with body size differs according to estrogen and progesterone receptor-defined subtypes. Results: Our simulations showed that the polytomous logistic regression model but not the joint Cox regression model yielded biased results in comparing exposure and disease subtype associations when the baseline hazards for different disease subtypes are nonproportional. For this reason, an analysis of a real data set was based on the joint Cox proportional hazards model and showed that body size has a significantly greater association with estrogen- and progesterone-positive breast cancer than with other subtypes. Conclusions: Because of the limitations of the polytomous logistic regression model for the comparison of exposure-disease associations across disease subtypes, the joint Cox proportional hazards model is recommended over the polytomous logistic regression model in prospective cohort studies. Impact: The article will promote the use of the joint Cox model in a prospective cohort study. Examples of SAS and S-plus programming codes are provided to facilitate use by nonstatisticians.

AB - Background: Polytomous logistic regression models are commonly used in case-control studies of cancer to directly compare the risks associated with an exposure variable across multiple cancer subtypes. However, the validity, accuracy, and efficiency of this approach for prospective cohort studies have not been formally evaluated. Methods: We investigated the performance of the polytomous logistic regression model and compared it with an alternative approach based on a joint Cox proportional hazards model using simulation studies. We then applied both methods to a prospective cohort study to assess whether the association of breast cancer with body size differs according to estrogen and progesterone receptor-defined subtypes. Results: Our simulations showed that the polytomous logistic regression model but not the joint Cox regression model yielded biased results in comparing exposure and disease subtype associations when the baseline hazards for different disease subtypes are nonproportional. For this reason, an analysis of a real data set was based on the joint Cox proportional hazards model and showed that body size has a significantly greater association with estrogen- and progesterone-positive breast cancer than with other subtypes. Conclusions: Because of the limitations of the polytomous logistic regression model for the comparison of exposure-disease associations across disease subtypes, the joint Cox proportional hazards model is recommended over the polytomous logistic regression model in prospective cohort studies. Impact: The article will promote the use of the joint Cox model in a prospective cohort study. Examples of SAS and S-plus programming codes are provided to facilitate use by nonstatisticians.

UR - http://www.scopus.com/inward/record.url?scp=84873394888&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84873394888&partnerID=8YFLogxK

U2 - 10.1158/1055-9965.EPI-12-1050

DO - 10.1158/1055-9965.EPI-12-1050

M3 - Article

C2 - 23292084

AN - SCOPUS:84873394888

VL - 22

SP - 275

EP - 285

JO - Cancer Epidemiology Biomarkers and Prevention

JF - Cancer Epidemiology Biomarkers and Prevention

SN - 1055-9965

IS - 2

ER -