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

Variable selection for propensity score models.

M Alan Brookhart1, Sebastian Schneeweiss, Kenneth J Rothman

  • 1Division of Pharmacoepidemiology and Pharmacoeconomics, Department of Medicine, Brigham and Women's Hospital, Harvard Medical School, Boston, MA 02120, USA. abrookhart@rics.harvard.edu

American Journal of Epidemiology
|April 21, 2006
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

Products That Pose Health Risks - Can Litigation Protect Us When Government Fails?

The New England journal of medicine·2026
Same author

Patterns of antithrombotic treatment after left atrial appendage occlusion.

Heart rhythm·2026
Same author

Out-of-Pocket Spending for Insulin by Medicare Beneficiaries After Monthly Caps.

JAMA internal medicine·2026
Same author

Fluticasone- vs Budesonide-Based Dual Therapy for COPD.

JAMA network open·2026
Same author

List Price Reductions Among Brand-Name ICS-LABA Inhalers In 2024 Were Associated With Increased Generic Uptake.

Health affairs (Project Hope)·2026
Same author

Risk of neutropenia-related hospitalisation among clozapine initiators.

BMJ mental health·2026

Variable selection for propensity score (PS) models is crucial in epidemiology. Including variables unrelated to exposure but related to outcomes improves PS analysis by reducing variance without bias.

Area of Science:

  • Epidemiology
  • Biostatistics
  • Health Research Methods

Background:

  • Propensity score (PS) methods are increasingly used in epidemiology.
  • Variable selection for PS models remains a challenge in the epidemiologic literature.

Purpose of the Study:

  • To provide insights into variable selection for PS models through simulation studies.
  • To illustrate the impact of variable selection on bias, variance, and mean squared error of exposure effect estimates.

Main Methods:

  • Conducted two simulation studies.
  • Analyzed the effects of including different types of variables in PS models.

Main Results:

  • Variables unrelated to exposure but related to outcome should be included to decrease variance without increasing bias.

Related Experiment Videos

  • Variables related to exposure but not outcome increase variance without decreasing bias.
  • In small studies, variables strongly related to exposure but weakly to outcome can be detrimental.

    • Conclusions:

    • Standard model-building tools for exposure prediction may not yield optimal PS models.
    • Careful variable selection is essential for robust PS analysis, especially in smaller studies.