Cluster analysis for repeated data with dropout: Sensitivity analysis using a distal event

Liesbeth Bruckers; Geert Molenberghs; Bianca Pulinx; Femina Hellenthal; Geert Schurink

doi:10.1080/10543406.2018.1428612

Cluster analysis for repeated data with dropout: Sensitivity analysis using a distal event

Liesbeth Bruckers^*, Geert Molenberghs, Bianca Pulinx, Femina Hellenthal, Geert Schurink

^*Corresponding author for this work

Research output: Contribution to journal › Article › Academic › peer-review

Abstract

Degeneration of the aortic wall becomes life-threatening when the risk of rupture increases. Cluster analysis on repeated measures of the diameter of the artery revealed two subgroups of patients included in a surveillance program. These results were obtained under the assumption of missingness at random. In this article, we study the vulnerability of the cluster analysis results - the estimated trajectories and the posterior membership probabilities - by applying different missing-data models for non-ignorable dropout, as proposed by Muthen et al. (2011) to the growth of the diameter of the artery.

Original language	English
Pages (from-to)	983-1004
Number of pages	22
Journal	Journal of Biopharmaceutical Statistics
Volume	28
Issue number	5
DOIs	https://doi.org/10.1080/10543406.2018.1428612
Publication status	Published - 1 Jan 2018

Keywords

Distal event
incomplete data
latent-class growth models
pattern-mixture models
selection models
sensitivity analysis
PATTERN-MIXTURE MODELS
INCOMPLETE DATA
LONGITUDINAL DATA
MISSING DATA
EM ALGORITHM
LIKELIHOOD
FIT

Access to Document

10.1080/10543406.2018.1428612

Cite this

@article{a209c089841444d8b85c7fb568d9acc0,

title = "Cluster analysis for repeated data with dropout: Sensitivity analysis using a distal event",

abstract = "Degeneration of the aortic wall becomes life-threatening when the risk of rupture increases. Cluster analysis on repeated measures of the diameter of the artery revealed two subgroups of patients included in a surveillance program. These results were obtained under the assumption of missingness at random. In this article, we study the vulnerability of the cluster analysis results - the estimated trajectories and the posterior membership probabilities - by applying different missing-data models for non-ignorable dropout, as proposed by Muthen et al. (2011) to the growth of the diameter of the artery.",

keywords = "Distal event, incomplete data, latent-class growth models, pattern-mixture models, selection models, sensitivity analysis, PATTERN-MIXTURE MODELS, INCOMPLETE DATA, LONGITUDINAL DATA, MISSING DATA, EM ALGORITHM, LIKELIHOOD, FIT",

author = "Liesbeth Bruckers and Geert Molenberghs and Bianca Pulinx and Femina Hellenthal and Geert Schurink",

year = "2018",

month = jan,

day = "1",

doi = "10.1080/10543406.2018.1428612",

language = "English",

volume = "28",

pages = "983--1004",

journal = "Journal of Biopharmaceutical Statistics",

issn = "1054-3406",

publisher = "Routledge/Taylor & Francis Group",

number = "5",

}

TY - JOUR

T1 - Cluster analysis for repeated data with dropout

T2 - Sensitivity analysis using a distal event

AU - Bruckers, Liesbeth

AU - Molenberghs, Geert

AU - Pulinx, Bianca

AU - Hellenthal, Femina

AU - Schurink, Geert

PY - 2018/1/1

Y1 - 2018/1/1

N2 - Degeneration of the aortic wall becomes life-threatening when the risk of rupture increases. Cluster analysis on repeated measures of the diameter of the artery revealed two subgroups of patients included in a surveillance program. These results were obtained under the assumption of missingness at random. In this article, we study the vulnerability of the cluster analysis results - the estimated trajectories and the posterior membership probabilities - by applying different missing-data models for non-ignorable dropout, as proposed by Muthen et al. (2011) to the growth of the diameter of the artery.

AB - Degeneration of the aortic wall becomes life-threatening when the risk of rupture increases. Cluster analysis on repeated measures of the diameter of the artery revealed two subgroups of patients included in a surveillance program. These results were obtained under the assumption of missingness at random. In this article, we study the vulnerability of the cluster analysis results - the estimated trajectories and the posterior membership probabilities - by applying different missing-data models for non-ignorable dropout, as proposed by Muthen et al. (2011) to the growth of the diameter of the artery.

KW - Distal event

KW - incomplete data

KW - latent-class growth models

KW - pattern-mixture models

KW - selection models

KW - sensitivity analysis

KW - PATTERN-MIXTURE MODELS

KW - INCOMPLETE DATA

KW - LONGITUDINAL DATA

KW - MISSING DATA

KW - EM ALGORITHM

KW - LIKELIHOOD

KW - FIT

U2 - 10.1080/10543406.2018.1428612

DO - 10.1080/10543406.2018.1428612

M3 - Article

C2 - 29376762

SN - 1054-3406

VL - 28

SP - 983

EP - 1004

JO - Journal of Biopharmaceutical Statistics

JF - Journal of Biopharmaceutical Statistics

IS - 5

ER -

Cluster analysis for repeated data with dropout: Sensitivity analysis using a distal event

Abstract

Keywords

Access to Document

Fingerprint

Cite this