A copula model for repeated measurements with non-ignorable non-monotone missing outcome

Changyu Shen, Lisa Weissfeld

Research output: Contribution to journalArticle

9 Citations (Scopus)

Abstract

A normal copula-based selection model is proposed for continuous longitudinal data with a non-ignorable non-monotone missing-data process. The normal copula is used to combine the distribution of the outcome of interest and that of the missing-data indicators given the covariates. Parameters in the model are estimated by a pseudo-likelihood method. We first use the GEE with a logistic link to estimate the parameters associated with the marginal distribution of the missing-data indicator given the covariates, assuming that covariates are always observed. Then we estimate other parameters by inserting the estimates from the first step into the full likelihood function. A simulation study is conducted to assess the robustness of the assumed model under different missing-data processes. The proposed method is then applied to one example from a community cohort study to demonstrate its capability to reduce bias.

Original languageEnglish
Pages (from-to)2427-2440
Number of pages14
JournalStatistics in Medicine
Volume25
Issue number14
DOIs
StatePublished - Jul 30 2006

Fingerprint

Copula Models
Repeated Measurements
Missing Data
Likelihood Functions
Covariates
Copula
Cohort Studies
Estimate
Pseudo-likelihood
Cohort Study
Selection Model
Likelihood Methods
Longitudinal Data
Marginal Distribution
Likelihood Function
Logistics
Simulation Study
Robustness
Model
Demonstrate

Keywords

  • Copula
  • Missing data
  • Mixed-effects model
  • Non-ignorable
  • Non-monotone

ASJC Scopus subject areas

  • Epidemiology

Cite this

A copula model for repeated measurements with non-ignorable non-monotone missing outcome. / Shen, Changyu; Weissfeld, Lisa.

In: Statistics in Medicine, Vol. 25, No. 14, 30.07.2006, p. 2427-2440.

Research output: Contribution to journalArticle

Shen, Changyu ; Weissfeld, Lisa. / A copula model for repeated measurements with non-ignorable non-monotone missing outcome. In: Statistics in Medicine. 2006 ; Vol. 25, No. 14. pp. 2427-2440.
@article{edd203943c8947638f31602975751373,
title = "A copula model for repeated measurements with non-ignorable non-monotone missing outcome",
abstract = "A normal copula-based selection model is proposed for continuous longitudinal data with a non-ignorable non-monotone missing-data process. The normal copula is used to combine the distribution of the outcome of interest and that of the missing-data indicators given the covariates. Parameters in the model are estimated by a pseudo-likelihood method. We first use the GEE with a logistic link to estimate the parameters associated with the marginal distribution of the missing-data indicator given the covariates, assuming that covariates are always observed. Then we estimate other parameters by inserting the estimates from the first step into the full likelihood function. A simulation study is conducted to assess the robustness of the assumed model under different missing-data processes. The proposed method is then applied to one example from a community cohort study to demonstrate its capability to reduce bias.",
keywords = "Copula, Missing data, Mixed-effects model, Non-ignorable, Non-monotone",
author = "Changyu Shen and Lisa Weissfeld",
year = "2006",
month = "7",
day = "30",
doi = "10.1002/sim.2355",
language = "English",
volume = "25",
pages = "2427--2440",
journal = "Statistics in Medicine",
issn = "0277-6715",
publisher = "John Wiley and Sons Ltd",
number = "14",

}

TY - JOUR

T1 - A copula model for repeated measurements with non-ignorable non-monotone missing outcome

AU - Shen, Changyu

AU - Weissfeld, Lisa

PY - 2006/7/30

Y1 - 2006/7/30

N2 - A normal copula-based selection model is proposed for continuous longitudinal data with a non-ignorable non-monotone missing-data process. The normal copula is used to combine the distribution of the outcome of interest and that of the missing-data indicators given the covariates. Parameters in the model are estimated by a pseudo-likelihood method. We first use the GEE with a logistic link to estimate the parameters associated with the marginal distribution of the missing-data indicator given the covariates, assuming that covariates are always observed. Then we estimate other parameters by inserting the estimates from the first step into the full likelihood function. A simulation study is conducted to assess the robustness of the assumed model under different missing-data processes. The proposed method is then applied to one example from a community cohort study to demonstrate its capability to reduce bias.

AB - A normal copula-based selection model is proposed for continuous longitudinal data with a non-ignorable non-monotone missing-data process. The normal copula is used to combine the distribution of the outcome of interest and that of the missing-data indicators given the covariates. Parameters in the model are estimated by a pseudo-likelihood method. We first use the GEE with a logistic link to estimate the parameters associated with the marginal distribution of the missing-data indicator given the covariates, assuming that covariates are always observed. Then we estimate other parameters by inserting the estimates from the first step into the full likelihood function. A simulation study is conducted to assess the robustness of the assumed model under different missing-data processes. The proposed method is then applied to one example from a community cohort study to demonstrate its capability to reduce bias.

KW - Copula

KW - Missing data

KW - Mixed-effects model

KW - Non-ignorable

KW - Non-monotone

UR - http://www.scopus.com/inward/record.url?scp=33745949690&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=33745949690&partnerID=8YFLogxK

U2 - 10.1002/sim.2355

DO - 10.1002/sim.2355

M3 - Article

VL - 25

SP - 2427

EP - 2440

JO - Statistics in Medicine

JF - Statistics in Medicine

SN - 0277-6715

IS - 14

ER -