A shared random effect parameter approach for longitudinal dementia data with non-ignorable missing data

Research output: Contribution to journalArticle

37 Citations (Scopus)

Abstract

A significant source of missing data in longitudinal epidemiologic studies on elderly individuals is death. It is generally believed that these missing data by death are non-ignorable to likelihood based inference. Inference based on data only from surviving participants in the study may lead to biased results. In this paper we model both the probability of disease and the probability of death using shared random effect parameters. We also propose to use the Laplace approximation for obtaining an approximate likelihood function so that high dimensional integration over the distributions of the random effect parameters is not necessary. Parameter estimates can be obtained by maximizing the approximate log-likelihood function. Data from a longitudinal dementia study will be used to illustrate the approach. A small simulation is conducted to compare parameter estimates from the proposed method to the 'naive' method where missing data is considered at random.

Original languageEnglish (US)
Pages (from-to)211-219
Number of pages9
JournalStatistics in Medicine
Volume23
Issue number2
DOIs
StatePublished - Jan 30 2004

Fingerprint

Nonignorable Missing Data
Dementia
Random Effects
Likelihood Functions
Missing Data
Longitudinal Studies
Likelihood Function
Information Storage and Retrieval
Laplace Approximation
Epidemiologic Studies
Estimate
Biased
Likelihood
High-dimensional
Necessary
Simulation

ASJC Scopus subject areas

  • Epidemiology
  • Statistics and Probability

Cite this

A shared random effect parameter approach for longitudinal dementia data with non-ignorable missing data. / Gao, Sujuan.

In: Statistics in Medicine, Vol. 23, No. 2, 30.01.2004, p. 211-219.

Research output: Contribution to journalArticle

@article{e590e562bac3422dad7f62f2d11128bd,
title = "A shared random effect parameter approach for longitudinal dementia data with non-ignorable missing data",
abstract = "A significant source of missing data in longitudinal epidemiologic studies on elderly individuals is death. It is generally believed that these missing data by death are non-ignorable to likelihood based inference. Inference based on data only from surviving participants in the study may lead to biased results. In this paper we model both the probability of disease and the probability of death using shared random effect parameters. We also propose to use the Laplace approximation for obtaining an approximate likelihood function so that high dimensional integration over the distributions of the random effect parameters is not necessary. Parameter estimates can be obtained by maximizing the approximate log-likelihood function. Data from a longitudinal dementia study will be used to illustrate the approach. A small simulation is conducted to compare parameter estimates from the proposed method to the 'naive' method where missing data is considered at random.",
author = "Sujuan Gao",
year = "2004",
month = "1",
day = "30",
doi = "10.1002/sim.1710",
language = "English (US)",
volume = "23",
pages = "211--219",
journal = "Statistics in Medicine",
issn = "0277-6715",
publisher = "John Wiley and Sons Ltd",
number = "2",

}

TY - JOUR

T1 - A shared random effect parameter approach for longitudinal dementia data with non-ignorable missing data

AU - Gao, Sujuan

PY - 2004/1/30

Y1 - 2004/1/30

N2 - A significant source of missing data in longitudinal epidemiologic studies on elderly individuals is death. It is generally believed that these missing data by death are non-ignorable to likelihood based inference. Inference based on data only from surviving participants in the study may lead to biased results. In this paper we model both the probability of disease and the probability of death using shared random effect parameters. We also propose to use the Laplace approximation for obtaining an approximate likelihood function so that high dimensional integration over the distributions of the random effect parameters is not necessary. Parameter estimates can be obtained by maximizing the approximate log-likelihood function. Data from a longitudinal dementia study will be used to illustrate the approach. A small simulation is conducted to compare parameter estimates from the proposed method to the 'naive' method where missing data is considered at random.

AB - A significant source of missing data in longitudinal epidemiologic studies on elderly individuals is death. It is generally believed that these missing data by death are non-ignorable to likelihood based inference. Inference based on data only from surviving participants in the study may lead to biased results. In this paper we model both the probability of disease and the probability of death using shared random effect parameters. We also propose to use the Laplace approximation for obtaining an approximate likelihood function so that high dimensional integration over the distributions of the random effect parameters is not necessary. Parameter estimates can be obtained by maximizing the approximate log-likelihood function. Data from a longitudinal dementia study will be used to illustrate the approach. A small simulation is conducted to compare parameter estimates from the proposed method to the 'naive' method where missing data is considered at random.

UR - http://www.scopus.com/inward/record.url?scp=0347093540&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0347093540&partnerID=8YFLogxK

U2 - 10.1002/sim.1710

DO - 10.1002/sim.1710

M3 - Article

C2 - 14716723

AN - SCOPUS:0347093540

VL - 23

SP - 211

EP - 219

JO - Statistics in Medicine

JF - Statistics in Medicine

SN - 0277-6715

IS - 2

ER -