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

Testing for baseline balance in clinical trials

S Senn1

  • 1Medicine and Clinical Development Department, CIBA, Basle, Switzerland.

Statistics in Medicine
|September 15, 1994
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

Prevalence of humeral head osteochondrosis in the Greater Swiss Mountain dog and the Border Collie in Switzerland.

Schweizer Archiv fur Tierheilkunde·2016
Same author

A 4-year trial of tiotropium in chronic obstructive pulmonary disease (UPLIFT trial).

Revista portuguesa de pneumologia·2015
Same author

Statisticians and pharmacokineticists: what they can still learn from each other.

Clinical pharmacology and therapeutics·2010
Same author

Ethical considerations concerning treatment allocation in drug development trials.

Statistical methods in medical research·2002
Same author

Actions other than smooth muscle relaxation may play a role in the protective effects of formoterol on the allergen-induced late asthmatic reaction.

Pulmonary pharmacology & therapeutics·2002
Same author

Screening for breast cancer with mammography.

Lancet (London, England)·2002

Clinical trial analysis should avoid baseline homogeneity tests. Instead, identify prognostic variables in the trial plan and use analysis of covariance for robust treatment effect evaluation.

Area of Science:

  • Clinical Trials
  • Biostatistics
  • Medical Research Methodology

Background:

  • Current clinical trial analysis often includes baseline homogeneity tests for prognostic covariates.
  • These tests assess if covariate distributions are similar across treatment groups before outcome analysis.

Purpose of the Study:

  • To critically evaluate the practice of baseline homogeneity testing in clinical trials.
  • To propose an alternative, statistically sound approach for handling prognostic variables.

Main Methods:

  • The study argues against the philosophical and practical validity of baseline homogeneity tests.
  • It recommends identifying prognostic variables during the trial design phase.
  • The proposed method involves incorporating these variables into an analysis of covariance (ANCOVA).

Related Experiment Videos

Main Results:

  • Baseline homogeneity tests are philosophically unsound, practically unhelpful, and potentially misleading.
  • The statistical significance of baseline distributions does not determine the necessity of adjusting for covariates.
  • ANCOVA provides a more robust method for analyzing treatment effects, irrespective of baseline covariate distributions.

Conclusions:

  • The practice of testing for baseline homogeneity in clinical trials should be abandoned.
  • Prognostic variables should be pre-specified and included in the analysis of covariance model.
  • This approach ensures a more rigorous and reliable assessment of treatment efficacy.