Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Marginalized transition models for longitudinal binary data with ignorable and non-ignorable drop-out.

Brenda F Kurland1, Patrick J Heagerty

  • 1National Alzheimer's Coordinating Center, University of Washington, Department of Epidemiology, 4311 11th Ave NE #300, Seattle, WA 98105, USA. kurland@u.washington.edu

Statistics in Medicine
|August 19, 2004
PubMed
Summary
This summary is machine-generated.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Data sharing in stepped-wedge cluster randomized trials: suboptimal data availability despite "data available upon request".

Journal of clinical epidemiology·2026
Same author

Spinal Manipulation and Clinician-Supported Self-Management for Preventing Chronic Low Back Pain Impact: The PACBACK Randomized Clinical Trial.

JAMA internal medicine·2026
Same author

CRT-Estimands Framework: consensus based extension of the ICH E9(R1) addendum for cluster randomised trials.

BMJ (Clinical research ed.)·2026
Same author

A tutorial on conducting sample size and power calculations for detecting treatment effect heterogeneity in cluster randomized trials with linear mixed models.

International journal of epidemiology·2026
Same author

Mechanistically informed circulating biomarkers are associated with acquired epilepsy after neonatal brain injury.

Journal of neuroinflammation·2026
Same author

Clinical trials for continuously monitored and updated AI systems.

Nature medicine·2026

This study enhances statistical models for handling non-ignorable dropout in longitudinal data. Sensitivity analysis reveals potential biases in common methods, emphasizing the need for improved model fit to reduce errors.

Area of Science:

  • Biostatistics
  • Longitudinal Data Analysis
  • Statistical Modeling

Background:

  • Longitudinal studies often encounter missing data due to participant dropout.
  • Monotone dropout, where participants drop out and do not return, presents unique analytical challenges.
  • Existing models may not adequately address non-ignorable dropout mechanisms.

Purpose of the Study:

  • To extend the marginalized transition model (MTM) to accommodate non-ignorable monotone dropout.
  • To evaluate the performance of MTM and inverse probability of censoring weighted generalized estimating equations (IPCW-GEE) under various missing data assumptions.
  • To assess the impact of model misspecification on parameter estimates.

Main Methods:

  • Utilized a selection model approach to handle non-ignorable dropout.

Related Experiment Videos

  • Conducted sensitivity analysis by holding weakly identified dropout parameters constant.
  • Compared the efficiency and bias of MTM and IPCW-GEE for data missing at random (MAR) and non-ignorable missing data.

    • Main Results:

    • The efficiency of IPCW-GEE can be as low as 40% compared to MTM for MAR data.
    • Both MTM and IPCW-GEE regression parameters exhibit misspecification bias for MAR and non-ignorable missing data.
    • Improving model fit for both MTM and IPCW-GEE noticeably reduces bias.

    Conclusions:

    • The extended MTM provides a framework for analyzing data with non-ignorable monotone dropout.
    • Sensitivity analysis is crucial for evaluating the robustness of findings to dropout assumptions.
    • Model fit is a critical factor in mitigating bias for both MTM and IPCW-GEE in longitudinal studies with missing data.