Bayesian Model Assessment in Joint Modeling of Longitudinal and Survival Data With Applications to Cancer Clinical Trials

Danjie Zhang, Ming Hui Chen, Joseph G. Ibrahim, Mark E. Boye, Wei Shen

Research output: Research - peer-reviewArticle

  • 1 Citations

Abstract

Joint models for longitudinal and survival data are routinely used in clinical trials or other studies to assess a treatment effect while accounting for longitudinal measures such as patient-reported outcomes. In the Bayesian framework, the deviance information criterion (DIC) and the logarithm of the pseudo-marginal likelihood (LPML) are two well-known Bayesian criteria for comparing joint models. However, these criteria do not provide separate assessments of each component of the joint model. In this article, we develop a novel decomposition of DIC and LPML to assess the fit of the longitudinal and survival components of the joint model, separately. Based on this decomposition, we then propose new Bayesian model assessment criteria, namely, ΔDIC and ΔLPML, to determine the importance and contribution of the longitudinal (survival) data to the model fit of the survival (longitudinal) data. Moreover, we develop an efficient Monte Carlo method for computing the conditional predictive ordinate statistics in the joint modeling setting. A simulation study is conducted to examine the empirical performance of the proposed criteria and the proposed methodology is further applied to a case study in mesothelioma. Supplementary materials for this article are available online.

LanguageEnglish (US)
Pages121-133
Number of pages13
JournalJournal of Computational and Graphical Statistics
Volume26
Issue number1
DOIs
StatePublished - Jan 2 2017

Fingerprint

Joint Modeling
Joint Model
Survival Data
Bayesian Model
Longitudinal Data
Clinical Trials
Cancer
Bayesian model
Clinical trials
Modeling
Deviance Information Criterion
Marginal Likelihood
Logarithm
Decompose
Information criterion
Decomposition
Deviance
Marginal likelihood
Ordinate
Treatment Effects

Keywords

  • CPO
  • DIC
  • LPML
  • Monte Carlo method
  • Patient-reported outcome (PRO)

ASJC Scopus subject areas

  • Statistics and Probability
  • Discrete Mathematics and Combinatorics
  • Statistics, Probability and Uncertainty

Cite this

Bayesian Model Assessment in Joint Modeling of Longitudinal and Survival Data With Applications to Cancer Clinical Trials. / Zhang, Danjie; Chen, Ming Hui; Ibrahim, Joseph G.; Boye, Mark E.; Shen, Wei.

In: Journal of Computational and Graphical Statistics, Vol. 26, No. 1, 02.01.2017, p. 121-133.

Research output: Research - peer-reviewArticle

@article{0061276ae51a407999d5e4ae81249c4f,
title = "Bayesian Model Assessment in Joint Modeling of Longitudinal and Survival Data With Applications to Cancer Clinical Trials",
abstract = "Joint models for longitudinal and survival data are routinely used in clinical trials or other studies to assess a treatment effect while accounting for longitudinal measures such as patient-reported outcomes. In the Bayesian framework, the deviance information criterion (DIC) and the logarithm of the pseudo-marginal likelihood (LPML) are two well-known Bayesian criteria for comparing joint models. However, these criteria do not provide separate assessments of each component of the joint model. In this article, we develop a novel decomposition of DIC and LPML to assess the fit of the longitudinal and survival components of the joint model, separately. Based on this decomposition, we then propose new Bayesian model assessment criteria, namely, ΔDIC and ΔLPML, to determine the importance and contribution of the longitudinal (survival) data to the model fit of the survival (longitudinal) data. Moreover, we develop an efficient Monte Carlo method for computing the conditional predictive ordinate statistics in the joint modeling setting. A simulation study is conducted to examine the empirical performance of the proposed criteria and the proposed methodology is further applied to a case study in mesothelioma. Supplementary materials for this article are available online.",
keywords = "CPO, DIC, LPML, Monte Carlo method, Patient-reported outcome (PRO)",
author = "Danjie Zhang and Chen, {Ming Hui} and Ibrahim, {Joseph G.} and Boye, {Mark E.} and Wei Shen",
year = "2017",
month = "1",
doi = "10.1080/10618600.2015.1117472",
volume = "26",
pages = "121--133",
journal = "Journal of Computational and Graphical Statistics",
issn = "1061-8600",
publisher = "American Statistical Association",
number = "1",

}

TY - JOUR

T1 - Bayesian Model Assessment in Joint Modeling of Longitudinal and Survival Data With Applications to Cancer Clinical Trials

AU - Zhang,Danjie

AU - Chen,Ming Hui

AU - Ibrahim,Joseph G.

AU - Boye,Mark E.

AU - Shen,Wei

PY - 2017/1/2

Y1 - 2017/1/2

N2 - Joint models for longitudinal and survival data are routinely used in clinical trials or other studies to assess a treatment effect while accounting for longitudinal measures such as patient-reported outcomes. In the Bayesian framework, the deviance information criterion (DIC) and the logarithm of the pseudo-marginal likelihood (LPML) are two well-known Bayesian criteria for comparing joint models. However, these criteria do not provide separate assessments of each component of the joint model. In this article, we develop a novel decomposition of DIC and LPML to assess the fit of the longitudinal and survival components of the joint model, separately. Based on this decomposition, we then propose new Bayesian model assessment criteria, namely, ΔDIC and ΔLPML, to determine the importance and contribution of the longitudinal (survival) data to the model fit of the survival (longitudinal) data. Moreover, we develop an efficient Monte Carlo method for computing the conditional predictive ordinate statistics in the joint modeling setting. A simulation study is conducted to examine the empirical performance of the proposed criteria and the proposed methodology is further applied to a case study in mesothelioma. Supplementary materials for this article are available online.

AB - Joint models for longitudinal and survival data are routinely used in clinical trials or other studies to assess a treatment effect while accounting for longitudinal measures such as patient-reported outcomes. In the Bayesian framework, the deviance information criterion (DIC) and the logarithm of the pseudo-marginal likelihood (LPML) are two well-known Bayesian criteria for comparing joint models. However, these criteria do not provide separate assessments of each component of the joint model. In this article, we develop a novel decomposition of DIC and LPML to assess the fit of the longitudinal and survival components of the joint model, separately. Based on this decomposition, we then propose new Bayesian model assessment criteria, namely, ΔDIC and ΔLPML, to determine the importance and contribution of the longitudinal (survival) data to the model fit of the survival (longitudinal) data. Moreover, we develop an efficient Monte Carlo method for computing the conditional predictive ordinate statistics in the joint modeling setting. A simulation study is conducted to examine the empirical performance of the proposed criteria and the proposed methodology is further applied to a case study in mesothelioma. Supplementary materials for this article are available online.

KW - CPO

KW - DIC

KW - LPML

KW - Monte Carlo method

KW - Patient-reported outcome (PRO)

UR - http://www.scopus.com/inward/record.url?scp=85013098571&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85013098571&partnerID=8YFLogxK

U2 - 10.1080/10618600.2015.1117472

DO - 10.1080/10618600.2015.1117472

M3 - Article

VL - 26

SP - 121

EP - 133

JO - Journal of Computational and Graphical Statistics

T2 - Journal of Computational and Graphical Statistics

JF - Journal of Computational and Graphical Statistics

SN - 1061-8600

IS - 1

ER -