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

Weighting in instrumental variables and G-estimation.

Marshall M Joffe1, Colleen Brensinger

  • 1Department of Biostatistics and Epidemiology, University of Pennsylvania School of Medicine, Room 602 Blockley Hall, 423 Guardian Drive, Philadelphia, PA 19104-6021, USA. mjoffe@cceb.upenn.edu

Statistics in Medicine
|April 11, 2003
PubMed
Summary

This study introduces a method to enhance randomized trial analysis using compliance and covariate data. The approach improves treatment effect estimates and increases statistical power, even when assumptions are not fully met.

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

Polycystic ovary syndrome and its association with reproductive cancers in women.

Fertility and sterility·2026
Same author

ASO Visual Abstract: What is the Cost of Selectivity? Selective and Nonselective Alpha Blockade Costs Associated with Adrenalectomy for Pheochromocytoma.

Annals of surgical oncology·2026
Same author

Fewer Pills, Lower Bills: Antihypertensive and Cost Outcomes of Adrenalectomy in Primary Aldosteronism.

Annals of surgical oncology·2026
Same author

What is the Cost of Selectivity? Selective and Nonselective Alpha Blockade Costs Associated with Adrenalectomy for Pheochromocytoma.

Annals of surgical oncology·2026
Same author

State-based variations in risk of institutional placement among dually eligible non-elderly adults with intellectual / developmental disabilities.

Disability and health journal·2025
Same author

Impact of a Longitudinal Virtual Education Series on Home Dialysis Education: A Mixed-Methods Evaluation.

Clinical journal of the American Society of Nephrology : CJASN·2025

Area of Science:

  • Biostatistics
  • Clinical Trials Methodology
  • Epidemiology

Background:

  • Randomized trials are crucial for establishing treatment efficacy.
  • Non-compliance with assigned treatment complicates the interpretation of randomized trial results.
  • Pre-randomization covariates can influence both compliance and outcomes.

Purpose of the Study:

  • To propose a statistical method for analyzing randomized trials with non-compliance.
  • To leverage information on compliance and pre-randomization covariates for improved analysis.
  • To enhance the precision of treatment effect estimates and the power of hypothesis tests.

Main Methods:

  • Utilizing compliance data and pre-randomization covariates.
  • Stratifying participants based on pretreatment covariates.

Related Experiment Videos

  • Determining the effect of randomization on treatment received within strata.
  • Weighting estimating functions by the effect of randomization on treatment received.

    • Main Results:

    • The proposed weighting scheme can improve the precision of explanatory treatment effect estimates.
    • The method can increase the power of intent-to-treat tests for the null hypothesis.
    • Efficiency gains are directly related to the variability of the weights, which are derived from covariate-predicted non-compliance.

    Conclusions:

    • The proposed method offers a simple yet effective way to improve the analysis of randomized trials with non-compliance.
    • The statistical gains are substantial when pretreatment covariates predict non-compliance and treatment effects are homogeneous across strata.
    • The weighting scheme demonstrates potential for improvement even when ideal conditions are not met, offering broad applicability.