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

Consequences of event rate heterogeneity across non-randomized study sub-groups.

Lynn E Eberly1

  • 1University of Minnesota, 420 Delaware St. SE, MMC 303, Minneapolis 55455, USA. lynn@biostat.umn.edu

Statistics in Medicine
|June 24, 2004
PubMed
Summary

Ignoring differential event rate variability in non-randomized group comparisons can bias results. This impacts statistical power and confidence interval accuracy, potentially leading to incorrect conclusions in clinical trial and observational data analyses.

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

Neurochemical Endpoints to Inform Early-Stage Trials of Spinocerebellar Ataxia 2 and 3 in a Multisite Setting.

Annals of clinical and translational neurology·2026
Same author

Racial and Ethnic Disparities in the Time to Dual Incontinence and its Treatment in Older Adults after Nursing Home Admission.

Annals of gerontology and geriatric research·2026
Same author

Factors associated with severe neurological sequelae of COVID-19: findings from the multicenter COVID-BRAIN imaging cohort.

Frontiers in human neuroscience·2026
Same author

Disrupted neurovascular-endocrine coupling in type 1 diabetes with impaired awareness of hypoglycemia.

The Journal of clinical investigation·2026
Same author

Examination of oxidative stress and glutamate as potential mechanisms of N-acetylcysteine in the treatment of non-suicidal self-injury in young people assigned female at birth: randomised trial.

BJPsych open·2025
Same author

Reduced removal of waste products from energy metabolism takes center stage in human brain aging.

Scientific reports·2025

Area of Science:

  • Biostatistics
  • Epidemiology
  • Clinical Trials

Background:

  • Non-randomized group comparisons are increasingly used in clinical research.
  • Unknown sources of heterogeneity in event rates are common in such analyses.
  • Ignoring heterogeneity can bias covariate effect estimates and reduce statistical power.

Purpose of the Study:

  • To investigate the impact of differential event rate variability between non-randomized groups.
  • To demonstrate how ignoring this heterogeneity affects regression models and statistical inference.
  • To highlight potential biases in post hoc analyses of clinical trials and observational studies.

Main Methods:

  • Analysis of post hoc clinical trials data.
  • Simulation studies to model the effects of ignored heterogeneity.

Related Experiment Videos

  • Evaluation of regression effect estimates, statistical power, and confidence interval coverage.

    • Main Results:

    • Differential event rate variability, when ignored, attenuates regression effect estimates.
    • This leads to lower power for hypothesis testing and reduced confidence interval coverage.
    • Ignoring heterogeneity can even produce significant negative comparisons when groups are not truly different.

    Conclusions:

    • Failure to account for differential event rate variability in non-randomized groups introduces significant bias.
    • Statistical models must incorporate heterogeneity to ensure accurate inference in clinical and observational research.
    • Careful consideration of variability is crucial for reliable comparisons between non-randomized groups.