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

A multiple imputation method for missing covariates in non-linear mixed-effects models with application to HIV

H Wu1, L Wu

  • 1Statistical and Data Analysis Center, Harvard School of Public Health, Frontier Science & Technology Research Foundation, Inc., 1244 Boylston Street, Suite 303, Chestnut Hill, Massachusetts 02467, USA. wu@sdac.harvard.edu

Statistics in Medicine
|June 15, 2001
PubMed
Summary

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

Analogs of the antimicrobial peptide trichogin having opposite membrane properties.

European journal of biochemistry·2001
Same author

Estimate of the frequency of true incomplete exchanges in human lymphocytes exposed to 1 GeV/u Fe ions in vitro.

International journal of radiation biology·2001
Same author

Genetic deletion of the Pten tumor suppressor gene promotes cell motility by activation of Rac1 and Cdc42 GTPases.

Current biology : CB·2001
Same author

Mammotome core biopsy for mammary microcalcification: analysis of 160 biopsies from 142 women with surgical and radiologic followup.

Cancer·2001
Same author

Clinical and angiographic outcome in the laser angioplasty for restenotic stents (LARS) multicenter registry.

Catheterization and cardiovascular interventions : official journal of the Society for Cardiac Angiography & Interventions·2001
Same author

Comparison of F ratios generated from interphase and metaphase chromosome damage induced by high doses of low- and high-LET radiation.

Radiation research·2000

This study introduces a novel three-step multiple imputation method for handling missing covariate data in non-linear mixed-effects models. The proposed method demonstrates superior accuracy and reliability compared to traditional imputation techniques.

Area of Science:

  • Biostatistics
  • Statistical Modeling
  • Data Science

Background:

  • Missing data in covariates poses challenges for accurate parameter estimation in non-linear mixed-effects models.
  • Traditional methods like mean-value imputation and complete-case analysis can introduce bias and reduce statistical power.

Purpose of the Study:

  • To develop and evaluate a robust three-step multiple imputation method for addressing missing covariates in non-linear mixed-effects models.
  • To compare the performance of the proposed method against mean-value imputation and complete-case analysis.

Main Methods:

  • A three-step multiple imputation technique implemented using a Gibbs sampler was developed.
  • Simulations were conducted to compare the proposed method with mean-value imputation and complete-case methods.

Related Experiment Videos

  • The methods were applied to model HIV viral dynamics using data from an AIDS clinical trial.

    • Main Results:

    • The proposed multiple imputation method yielded estimates with smaller biases and lower mean-squared errors for covariate coefficients compared to the other two methods.
    • The application to HIV viral dynamics data suggested more reliable results from the proposed method.

    Conclusions:

    • The three-step multiple imputation method provides a more accurate and reliable approach for parameter estimation in non-linear mixed-effects models with missing covariates.
    • This method is particularly valuable in complex modeling scenarios, such as analyzing HIV viral dynamics.