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

Randomized trials, generalizability, and meta-analysis: graphical insights for binary outcomes.

Stuart G Baker1, Barnett S Kramer

  • 1Biometry Research Group, Division of Cancer Prevention, National Cancer Institute, USA. sb16i@nih.gov

BMC Medical Research Methodology
|June 18, 2003
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

Prostate cancer screening: Decision-making when MRI results are inconclusive for cancer (PI-RADS 3).

Urologic oncology·2026
Same author

Adjusting for non-compliance and contamination with a more plausible assumption.

Journal of medical screening·2026
Same author

The efficient stored specimen design for evaluating multiple screening technologies: Application to multicancer detection tests.

Journal of medical screening·2025
Same author

Minimum test tradeoff applied to uterine artery pulsatility index for antepartum prediction of preeclampsia.

Pregnancy hypertension·2025
Same author

Trends in use of the new MeSH term "overdiagnosis": A bibliometric review.

Health information and libraries journal·2025
Same author

False premises, false promises: celebrity endorsement of non-evidence-based anticancer interventions on social media.

Health promotion international·2025

Randomized trials can assess treatment effects in populations with different unobserved variables. The BK-Plot method shows that absolute difference and relative risk are robust to these unobserved variables, unlike the odds ratio.

Area of Science:

  • Biostatistics
  • Clinical Trials
  • Epidemiology

Background:

  • Randomized trials (RCTs) are crucial for determining treatment effects.
  • Generalizability of RCTs depends on random sampling of the target population.
  • Unobserved variables can influence treatment outcome generalizability.

Purpose of the Study:

  • To investigate how RCTs can assess treatment effects in populations with varying distributions of unobserved binary variables.
  • To evaluate the impact of unobserved binary variables on treatment effect measures.

Main Methods:

  • Utilized a modified BK-Plot to visualize the effect of unobserved binary variables on outcomes.
  • Assessed three outcome measures: absolute difference (DIF), relative risk (RR), and odds ratio (OR).

Related Experiment Videos

  • Assumed treatment effects are consistent across levels of the unobserved variable.

    • Main Results:

    • The BK-Plot demonstrated that DIF and RR are invariant to the proportion of subjects with a specific unobserved binary variable.
    • The odds ratio (OR) was shown to be sensitive to the distribution of the unobserved variable.
    • Graphical analysis revealed the impact of unobserved covariates on treatment effect estimation.

    Conclusions:

    • The BK-Plot offers a straightforward method for assessing generalizability in RCTs.
    • Meta-analyses using DIF or RR for binary outcomes can mitigate bias from unobserved covariates.
    • OR-based meta-analyses may be susceptible to bias from unobserved variables that do not interact with treatment.