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

Fallibility in estimating direct effects.

Stephen R Cole1, Miguel A Hernán

  • 1Department of Epidemiology, Johns Hopkins Bloomberg School of Public Health, Baltimore, MD 21205, USA. scole@jhsph.edu

International Journal of Epidemiology
|March 27, 2002
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

Methodological Approaches to Real-World Evidence Generation for Glucagon-like Peptide-1-Based Therapies: Synopsis of a National Institute of Diabetes and Digestive and Kidney Diseases Workshop.

Annals of internal medicine·2026
Same author

Global Sensitivity Analysis for Studies Extending Inferences From a Randomized Trial to a Target Population.

Statistics in medicine·2026
Same author

Statins and risk of cardiovascular disease: Emulating a primary prevention trial in breast cancer survivors.

Journal of the National Cancer Institute·2026
Same author

Long-term effects of colonoscopy screening on colorectal cancer incidence and mortality: a multicountry, population-based randomised controlled trial.

Lancet (London, England)·2026
Same author

Extending inferences from a randomized trial to trial-eligible and treatment-candidate target populations: examples of generalizability and transportability.

Epidemiology (Cambridge, Mass.)·2026
Same author

The TARGET guideline for reporting observational studies of interventions.

Nature medicine·2026
Same journal

Age at menarche and adverse pregnancy and perinatal outcomes: triangulating evidence from multivariable and Mendelian randomization analyses.

International journal of epidemiology·2026
Same journal

Life-course trajectories of cardiovascular disease risk factors in rural India: Andhra Pradesh Children and Parents Study (APCAPS) 2003-2023.

International journal of epidemiology·2026
Same journal

Cohort Profile Update: The Young Lives study.

International journal of epidemiology·2026
Same journal

From the departing Editors in Chief.

International journal of epidemiology·2026
Same journal

Data Resource Profile: Cheeloo Lifespan Electronic-health reseArch Data-library (Cheeloo LEAD).

International journal of epidemiology·2026
Same journal

Cohort Profile Update: The Swiss Childhood Cancer Survivor Cohort.

International journal of epidemiology·2026
See all related articles

A standard method for estimating direct effects may be flawed. Ensuring no unmeasured confounding for both exposure-outcome and mediator-outcome relationships is crucial for unbiased direct effect estimation.

Area of Science:

  • Epidemiology
  • Causal Inference
  • Biostatistics

Background:

  • Direct effect estimation is vital in observational studies.
  • Standard methods like stratification on mediators are commonly used.
  • Potential biases in these methods require careful examination.

Purpose of the Study:

  • To explain the potential flaws in a common method for quantifying direct effects.
  • To highlight the assumptions required for unbiased direct effect estimation.
  • To provide recommendations for improving causal effect quantification.

Main Methods:

  • Utilized causal graphs to model relationships between exposure, mediator, and outcome.
  • Employed a hypothetical example inspired by the Physicians' Health Study.

Related Experiment Videos

  • Analyzed the conditions under which direct effect estimation can be biased.

    • Main Results:

    • Demonstrated that stratifying on an intermediate variable can lead to biased direct effect estimates.
    • Identified two key assumptions for unbiased direct effect estimation: absence of unmeasured confounding for exposure-outcome and mediator-outcome.
    • Showcased how unmeasured confounding can compromise the validity of the method.

    Conclusions:

    • The standard method of stratifying on the intermediate variable for direct effects may be flawed.
    • Unbiased direct effect estimation necessitates the absence of unmeasured confounding for both exposure-outcome and mediator-outcome.
    • Researchers should collect data on potential confounders and clearly state assumptions regarding unmeasured confounding.