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

Assessing non-consent bias with parallel randomized and nonrandomized clinical trials

S M Marcus1

  • 1Biostatistics Section, Jefferson Medical College, Philadelphia, PA 19107, USA.

Journal of Clinical Epidemiology
|July 1, 1997
PubMed
Summary

Non-consent bias can occur in clinical trials when eligible patients decline randomization. This bias may distort treatment effect estimates, necessitating careful analysis of both randomized and non-randomized groups.

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

Pediatric fatality review of the 2013 National Poison Database System (NPDS): focus on intent.

Clinical toxicology (Philadelphia, Pa.)·2015
Same author

2012 pediatric fatality review of the National Poison Database System.

Clinical toxicology (Philadelphia, Pa.)·2014
Same author

2011 Pediatric fatality review of the National Poison Data System [corrected].

Clinical toxicology (Philadelphia, Pa.)·2012
Same author

A randomized, prospective pilot study of patient expectancy and antidepressant outcome.

Psychological medicine·2012
Same author

Adverse reactions to over-the-counter cough and cold products among children: the cases managed out of hospitals.

Journal of clinical pharmacy and therapeutics·2009
Same author

Increased serotonin 2C receptor mRNA editing: a possible risk factor for suicide.

Molecular psychiatry·2007

Area of Science:

  • Clinical Trials Methodology
  • Biostatistics
  • Epidemiology

Background:

  • Randomized clinical trials (RCTs) are susceptible to non-consent bias when eligible participants refuse randomization.
  • This refusal can lead to discrepancies between the trial population and the broader eligible population, potentially affecting treatment effect generalizability.

Purpose of the Study:

  • To address non-consent bias in clinical trials.
  • To introduce a methodology for estimating treatment effects in the eligible population, accounting for participants who decline randomization.

Main Methods:

  • Proposes conducting parallel non-randomized trials alongside RCTs for patients who opt out of randomization.
  • Utilizes observed baseline covariate data to adjust for differences between randomized and non-randomized groups.

Related Experiment Videos

  • Employs sensitivity analysis to assess the impact of unobserved covariates on treatment effect estimates.

    • Main Results:

    • Adjusted analyses can reveal hidden non-consent bias stemming from unobserved differences between trial participants and the eligible population.
    • Discrepancies in outcomes between randomized and non-randomized groups, even after adjustment, indicate potential bias.

    Conclusions:

    • Parallel non-randomized trials can supplement RCTs to mitigate non-consent bias.
    • Methodology allows for more accurate estimation of treatment effects in the entire eligible population.
    • Sensitivity analysis is crucial for understanding the potential impact of unobserved confounding factors.