Abstract

Many longitudinal studies, especially in clinical trials, suffer from missing data issues. Most estimation procedures assume that the missing values are ignorable. However, this assumption leads to unrealistic simplification and is implausible for many cases.When non-ignorable missingness is preferred, classical pattern-mixture models with the data stratified according to a variety of missing patterns and a model specified for each stratum are widely used for longitudinal data analysis. But this assumption usually results in under-identifiability because of the need to estimate many stratum-specific parameters. Further, pattern mixture models have the drawback that a large sample is usually required. In this paper, a continuous latent factor model is proposed and this novel approach overcomes limitations which exist in pattern mixture models by specifying a continuous latent factor. The advantages of this model, including small sample feasibility, are demonstrated by comparing with Roy's pattern mixturemodel using an application to a clinical study of AIDS patients with advanced immune suppression.

Original languageEnglish (US)
Title of host publicationInnovative Statistical Methods for Public Health Data
PublisherSpringer International Publishing
Pages173-199
Number of pages27
ISBN (Print)9783319185361, 9783319185354
DOIs
StatePublished - Aug 31 2015

Fingerprint

Longitudinal Studies
Acquired Immunodeficiency Syndrome
Clinical Trials
Clinical Studies

ASJC Scopus subject areas

  • Medicine(all)

Cite this

Zhang, J., & Reiser, M. (2015). A continuous latent factor model for non-ignorable missing data. In Innovative Statistical Methods for Public Health Data (pp. 173-199). Springer International Publishing. https://doi.org/10.1007/978-3-319-18536-1_9

A continuous latent factor model for non-ignorable missing data. / Zhang, Jun; Reiser, Mark.

Innovative Statistical Methods for Public Health Data. Springer International Publishing, 2015. p. 173-199.

Research output: Chapter in Book/Report/Conference proceedingChapter

Zhang, J & Reiser, M 2015, A continuous latent factor model for non-ignorable missing data. in Innovative Statistical Methods for Public Health Data. Springer International Publishing, pp. 173-199. https://doi.org/10.1007/978-3-319-18536-1_9
Zhang J, Reiser M. A continuous latent factor model for non-ignorable missing data. In Innovative Statistical Methods for Public Health Data. Springer International Publishing. 2015. p. 173-199 https://doi.org/10.1007/978-3-319-18536-1_9
Zhang, Jun ; Reiser, Mark. / A continuous latent factor model for non-ignorable missing data. Innovative Statistical Methods for Public Health Data. Springer International Publishing, 2015. pp. 173-199
@inbook{16dd395ed6234ef38915047c235f3f0c,
title = "A continuous latent factor model for non-ignorable missing data",
abstract = "Many longitudinal studies, especially in clinical trials, suffer from missing data issues. Most estimation procedures assume that the missing values are ignorable. However, this assumption leads to unrealistic simplification and is implausible for many cases.When non-ignorable missingness is preferred, classical pattern-mixture models with the data stratified according to a variety of missing patterns and a model specified for each stratum are widely used for longitudinal data analysis. But this assumption usually results in under-identifiability because of the need to estimate many stratum-specific parameters. Further, pattern mixture models have the drawback that a large sample is usually required. In this paper, a continuous latent factor model is proposed and this novel approach overcomes limitations which exist in pattern mixture models by specifying a continuous latent factor. The advantages of this model, including small sample feasibility, are demonstrated by comparing with Roy's pattern mixturemodel using an application to a clinical study of AIDS patients with advanced immune suppression.",
author = "Jun Zhang and Mark Reiser",
year = "2015",
month = "8",
day = "31",
doi = "10.1007/978-3-319-18536-1_9",
language = "English (US)",
isbn = "9783319185361",
pages = "173--199",
booktitle = "Innovative Statistical Methods for Public Health Data",
publisher = "Springer International Publishing",

}

TY - CHAP

T1 - A continuous latent factor model for non-ignorable missing data

AU - Zhang, Jun

AU - Reiser, Mark

PY - 2015/8/31

Y1 - 2015/8/31

N2 - Many longitudinal studies, especially in clinical trials, suffer from missing data issues. Most estimation procedures assume that the missing values are ignorable. However, this assumption leads to unrealistic simplification and is implausible for many cases.When non-ignorable missingness is preferred, classical pattern-mixture models with the data stratified according to a variety of missing patterns and a model specified for each stratum are widely used for longitudinal data analysis. But this assumption usually results in under-identifiability because of the need to estimate many stratum-specific parameters. Further, pattern mixture models have the drawback that a large sample is usually required. In this paper, a continuous latent factor model is proposed and this novel approach overcomes limitations which exist in pattern mixture models by specifying a continuous latent factor. The advantages of this model, including small sample feasibility, are demonstrated by comparing with Roy's pattern mixturemodel using an application to a clinical study of AIDS patients with advanced immune suppression.

AB - Many longitudinal studies, especially in clinical trials, suffer from missing data issues. Most estimation procedures assume that the missing values are ignorable. However, this assumption leads to unrealistic simplification and is implausible for many cases.When non-ignorable missingness is preferred, classical pattern-mixture models with the data stratified according to a variety of missing patterns and a model specified for each stratum are widely used for longitudinal data analysis. But this assumption usually results in under-identifiability because of the need to estimate many stratum-specific parameters. Further, pattern mixture models have the drawback that a large sample is usually required. In this paper, a continuous latent factor model is proposed and this novel approach overcomes limitations which exist in pattern mixture models by specifying a continuous latent factor. The advantages of this model, including small sample feasibility, are demonstrated by comparing with Roy's pattern mixturemodel using an application to a clinical study of AIDS patients with advanced immune suppression.

UR - http://www.scopus.com/inward/record.url?scp=84955351227&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84955351227&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-18536-1_9

DO - 10.1007/978-3-319-18536-1_9

M3 - Chapter

AN - SCOPUS:84955351227

SN - 9783319185361

SN - 9783319185354

SP - 173

EP - 199

BT - Innovative Statistical Methods for Public Health Data

PB - Springer International Publishing

ER -