A latent-variable marginal method for multi-level incomplete binary data

Baojiang Chen, Xiao Hua Zhou

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

Incomplete multi-level data arise commonly in many clinical trials and observational studies. Because of multi-level variations in this type of data, appropriate data analysis should take these variations into account. A random effects model can allow for the multi-level variations by assuming random effects at each level, but the computation is intensive because high-dimensional integrations are often involved in fitting models. Marginal methods such as the inverse probability weighted generalized estimating equations can involve simple estimation computation, but it is hard to specify the working correlation matrix for multi-level data. In this paper, we introduce a latent variable method to deal with incomplete multi-level data when the missing mechanism is missing at random, which fills the gap between the random effects model and marginal models. Latent variable models are built for both the response and missing data processes to incorporate the variations that arise at each level. Simulation studies demonstrate that this method performs well in various situations. We apply the proposed method to an Alzheimer's disease study.

Original languageEnglish (US)
Pages (from-to)3211-3222
Number of pages12
JournalStatistics in Medicine
Volume31
Issue number26
DOIs
StatePublished - Nov 20 2012

Fingerprint

Binary Data
Incomplete Data
Latent Variables
Random Effects Model
Weighted Estimating Equations
Marginal Model
Latent Variable Models
Missing at Random
Observational Study
Generalized Estimating Equations
Alzheimer's Disease
Model Fitting
Observational Studies
Correlation Matrix
Alzheimer Disease
Random Effects
Missing Data
Clinical Trials
Data analysis
High-dimensional

Keywords

  • Estimating equation
  • Latent variable
  • Missing at random
  • Missing response
  • Multi-level

ASJC Scopus subject areas

  • Epidemiology
  • Statistics and Probability

Cite this

A latent-variable marginal method for multi-level incomplete binary data. / Chen, Baojiang; Zhou, Xiao Hua.

In: Statistics in Medicine, Vol. 31, No. 26, 20.11.2012, p. 3211-3222.

Research output: Contribution to journalArticle

Chen, Baojiang ; Zhou, Xiao Hua. / A latent-variable marginal method for multi-level incomplete binary data. In: Statistics in Medicine. 2012 ; Vol. 31, No. 26. pp. 3211-3222.
@article{844467a11eab42788a70da6cd3d17466,
title = "A latent-variable marginal method for multi-level incomplete binary data",
abstract = "Incomplete multi-level data arise commonly in many clinical trials and observational studies. Because of multi-level variations in this type of data, appropriate data analysis should take these variations into account. A random effects model can allow for the multi-level variations by assuming random effects at each level, but the computation is intensive because high-dimensional integrations are often involved in fitting models. Marginal methods such as the inverse probability weighted generalized estimating equations can involve simple estimation computation, but it is hard to specify the working correlation matrix for multi-level data. In this paper, we introduce a latent variable method to deal with incomplete multi-level data when the missing mechanism is missing at random, which fills the gap between the random effects model and marginal models. Latent variable models are built for both the response and missing data processes to incorporate the variations that arise at each level. Simulation studies demonstrate that this method performs well in various situations. We apply the proposed method to an Alzheimer's disease study.",
keywords = "Estimating equation, Latent variable, Missing at random, Missing response, Multi-level",
author = "Baojiang Chen and Zhou, {Xiao Hua}",
year = "2012",
month = "11",
day = "20",
doi = "10.1002/sim.5394",
language = "English (US)",
volume = "31",
pages = "3211--3222",
journal = "Statistics in Medicine",
issn = "0277-6715",
publisher = "John Wiley and Sons Ltd",
number = "26",

}

TY - JOUR

T1 - A latent-variable marginal method for multi-level incomplete binary data

AU - Chen, Baojiang

AU - Zhou, Xiao Hua

PY - 2012/11/20

Y1 - 2012/11/20

N2 - Incomplete multi-level data arise commonly in many clinical trials and observational studies. Because of multi-level variations in this type of data, appropriate data analysis should take these variations into account. A random effects model can allow for the multi-level variations by assuming random effects at each level, but the computation is intensive because high-dimensional integrations are often involved in fitting models. Marginal methods such as the inverse probability weighted generalized estimating equations can involve simple estimation computation, but it is hard to specify the working correlation matrix for multi-level data. In this paper, we introduce a latent variable method to deal with incomplete multi-level data when the missing mechanism is missing at random, which fills the gap between the random effects model and marginal models. Latent variable models are built for both the response and missing data processes to incorporate the variations that arise at each level. Simulation studies demonstrate that this method performs well in various situations. We apply the proposed method to an Alzheimer's disease study.

AB - Incomplete multi-level data arise commonly in many clinical trials and observational studies. Because of multi-level variations in this type of data, appropriate data analysis should take these variations into account. A random effects model can allow for the multi-level variations by assuming random effects at each level, but the computation is intensive because high-dimensional integrations are often involved in fitting models. Marginal methods such as the inverse probability weighted generalized estimating equations can involve simple estimation computation, but it is hard to specify the working correlation matrix for multi-level data. In this paper, we introduce a latent variable method to deal with incomplete multi-level data when the missing mechanism is missing at random, which fills the gap between the random effects model and marginal models. Latent variable models are built for both the response and missing data processes to incorporate the variations that arise at each level. Simulation studies demonstrate that this method performs well in various situations. We apply the proposed method to an Alzheimer's disease study.

KW - Estimating equation

KW - Latent variable

KW - Missing at random

KW - Missing response

KW - Multi-level

UR - http://www.scopus.com/inward/record.url?scp=84867544389&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84867544389&partnerID=8YFLogxK

U2 - 10.1002/sim.5394

DO - 10.1002/sim.5394

M3 - Article

VL - 31

SP - 3211

EP - 3222

JO - Statistics in Medicine

JF - Statistics in Medicine

SN - 0277-6715

IS - 26

ER -