A ramdom-effects Markov transition model for poisson-distributed repeated measures with non-ignorable missing values

Jinhui Li, Xiaowei Yang, Yingnian Wu, Steven Shoptaw

Research output: Contribution to journalArticle

9 Citations (Scopus)

Abstract

In biomedical research with longitudinal designs, missing values due to intermittent non-response or premature withdrawal are usually 'non-ignorable' in the sense that unobserved values are related to the patterns of missingness. By drawing the framework of a shared-parameter mechanism, the process yielding the repeated count measures and the process yielding missing values can be modelled separately, conditionally on a group of shared parameters. For chronic diseases, Markov transition models can be used to study the transitional features of the pathologic processes. In this paper, Markov Chain Monte Carlo algorithms are developed to fit a random-effects Markov transition model for incomplete count repeated measures, within which random effects are shared by the counting process and the missing-data mechanism. Assuming a Poisson distribution for the count measures, the transition probabilities are estimated using a Poisson regression model. The missingness mechanism is modelled with a multinomial-logit regression to calculate the transition probabilities of the missingness indicators. The method is demonstrated using both simulated data sets and a practical data set from a smoking cessation clinical trial.

Original languageEnglish (US)
Pages (from-to)2519-2532
Number of pages14
JournalStatistics in Medicine
Volume26
Issue number12
DOIs
StatePublished - May 30 2007

Fingerprint

Transition Model
Repeated Measures
Missing Values
Markov Model
Siméon Denis Poisson
Count
Poisson Distribution
Random Effects
Transition Probability
Markov Chains
Process Assessment (Health Care)
Smoking Cessation
Pathologic Processes
Multinomial Logit
Missing Data Mechanism
Biomedical Research
Poisson Regression
Chronic Disease
Non-response
Counting Process

Keywords

  • Markov transition models
  • Non-ignorable missing values
  • Poisson regression model
  • Repeated measures
  • Shared-parameter missingness

ASJC Scopus subject areas

  • Epidemiology

Cite this

A ramdom-effects Markov transition model for poisson-distributed repeated measures with non-ignorable missing values. / Li, Jinhui; Yang, Xiaowei; Wu, Yingnian; Shoptaw, Steven.

In: Statistics in Medicine, Vol. 26, No. 12, 30.05.2007, p. 2519-2532.

Research output: Contribution to journalArticle

Li, Jinhui ; Yang, Xiaowei ; Wu, Yingnian ; Shoptaw, Steven. / A ramdom-effects Markov transition model for poisson-distributed repeated measures with non-ignorable missing values. In: Statistics in Medicine. 2007 ; Vol. 26, No. 12. pp. 2519-2532.
@article{07d592299ed244be85d5def54642613b,
title = "A ramdom-effects Markov transition model for poisson-distributed repeated measures with non-ignorable missing values",
abstract = "In biomedical research with longitudinal designs, missing values due to intermittent non-response or premature withdrawal are usually 'non-ignorable' in the sense that unobserved values are related to the patterns of missingness. By drawing the framework of a shared-parameter mechanism, the process yielding the repeated count measures and the process yielding missing values can be modelled separately, conditionally on a group of shared parameters. For chronic diseases, Markov transition models can be used to study the transitional features of the pathologic processes. In this paper, Markov Chain Monte Carlo algorithms are developed to fit a random-effects Markov transition model for incomplete count repeated measures, within which random effects are shared by the counting process and the missing-data mechanism. Assuming a Poisson distribution for the count measures, the transition probabilities are estimated using a Poisson regression model. The missingness mechanism is modelled with a multinomial-logit regression to calculate the transition probabilities of the missingness indicators. The method is demonstrated using both simulated data sets and a practical data set from a smoking cessation clinical trial.",
keywords = "Markov transition models, Non-ignorable missing values, Poisson regression model, Repeated measures, Shared-parameter missingness",
author = "Jinhui Li and Xiaowei Yang and Yingnian Wu and Steven Shoptaw",
year = "2007",
month = "5",
day = "30",
doi = "10.1002/sim.2717",
language = "English (US)",
volume = "26",
pages = "2519--2532",
journal = "Statistics in Medicine",
issn = "0277-6715",
publisher = "John Wiley and Sons Ltd",
number = "12",

}

TY - JOUR

T1 - A ramdom-effects Markov transition model for poisson-distributed repeated measures with non-ignorable missing values

AU - Li, Jinhui

AU - Yang, Xiaowei

AU - Wu, Yingnian

AU - Shoptaw, Steven

PY - 2007/5/30

Y1 - 2007/5/30

N2 - In biomedical research with longitudinal designs, missing values due to intermittent non-response or premature withdrawal are usually 'non-ignorable' in the sense that unobserved values are related to the patterns of missingness. By drawing the framework of a shared-parameter mechanism, the process yielding the repeated count measures and the process yielding missing values can be modelled separately, conditionally on a group of shared parameters. For chronic diseases, Markov transition models can be used to study the transitional features of the pathologic processes. In this paper, Markov Chain Monte Carlo algorithms are developed to fit a random-effects Markov transition model for incomplete count repeated measures, within which random effects are shared by the counting process and the missing-data mechanism. Assuming a Poisson distribution for the count measures, the transition probabilities are estimated using a Poisson regression model. The missingness mechanism is modelled with a multinomial-logit regression to calculate the transition probabilities of the missingness indicators. The method is demonstrated using both simulated data sets and a practical data set from a smoking cessation clinical trial.

AB - In biomedical research with longitudinal designs, missing values due to intermittent non-response or premature withdrawal are usually 'non-ignorable' in the sense that unobserved values are related to the patterns of missingness. By drawing the framework of a shared-parameter mechanism, the process yielding the repeated count measures and the process yielding missing values can be modelled separately, conditionally on a group of shared parameters. For chronic diseases, Markov transition models can be used to study the transitional features of the pathologic processes. In this paper, Markov Chain Monte Carlo algorithms are developed to fit a random-effects Markov transition model for incomplete count repeated measures, within which random effects are shared by the counting process and the missing-data mechanism. Assuming a Poisson distribution for the count measures, the transition probabilities are estimated using a Poisson regression model. The missingness mechanism is modelled with a multinomial-logit regression to calculate the transition probabilities of the missingness indicators. The method is demonstrated using both simulated data sets and a practical data set from a smoking cessation clinical trial.

KW - Markov transition models

KW - Non-ignorable missing values

KW - Poisson regression model

KW - Repeated measures

KW - Shared-parameter missingness

UR - http://www.scopus.com/inward/record.url?scp=34249036469&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=34249036469&partnerID=8YFLogxK

U2 - 10.1002/sim.2717

DO - 10.1002/sim.2717

M3 - Article

C2 - 17106918

AN - SCOPUS:34249036469

VL - 26

SP - 2519

EP - 2532

JO - Statistics in Medicine

JF - Statistics in Medicine

SN - 0277-6715

IS - 12

ER -