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

Monte Carlo methods in clinical research: applications in multivariable analysis

J Concato1, A R Feinstein

  • 1Department of Medicine, Yale University School of Medicine, New Haven, CT 06510, USA.

Journal of Investigative Medicine : the Official Publication of the American Federation for Clinical Research
|August 1, 1997
PubMed
Summary

Monte Carlo simulations revealed that logistic regression models may be unreliable with fewer than 10-20 outcome events per independent variable (EPV). This finding is crucial for clinical researchers using statistical analysis to ensure accurate results.

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

External control arms in oncology: current use and future directions.

Annals of oncology : official journal of the European Society for Medical Oncology·2022
Same author

THE PRE-THERAPEUTIC CLASSIFICATION OF CO-MORBIDITY IN CHRONIC DISEASE.

Journal of chronic diseases·2015
Same author

NATURAL HISTORY AND TREATMENT OF RHEUMATIC FEVER.

A listing of research in the cardiovascular field·2014
Same author

II. Treatment and Prevention.

The Yale journal of biology and medicine·2011
Same author

Renal artery revascularization in patients with atherosclerotic renal artery stenosis and impaired renal function: conservative management versus renal artery stenting.

Clinical nephrology·2010
Same author

Biomarkers for the diagnosis and risk stratification of acute kidney injury: a systematic review.

Kidney international·2007

Area of Science:

  • Statistical modeling
  • Computational statistics
  • Clinical research methodology

Background:

  • Monte Carlo methods utilize random numbers for complex problem-solving, especially when traditional mathematical approaches fall short.
  • While widely used in science, the principles of Monte Carlo simulations may not be familiar to all clinical researchers.
  • This study focuses on the application of Monte Carlo methods in multivariable statistical analysis, specifically examining the outcome events per independent variable (EPV) ratio.

Purpose of the Study:

  • To elucidate the history and fundamental principles of Monte Carlo methods.
  • To demonstrate the practical application of Monte Carlo simulations in clinical research.
  • To investigate the impact of the number of outcome events per independent variable (EPV) on statistical model reliability.

Related Experiment Videos

Main Methods:

  • Real-world data from a clinical trial of 673 patients with 252 deaths and 7 predictor variables (EPV=36) were used.
  • Monte Carlo simulations were performed using proportional hazards and logistic regression models.
  • Simulations were conducted across a range of EPV values (2, 5, 10, 15, 20, and 25) to assess model performance.

Main Results:

  • Monte Carlo simulations confirmed a critical "rule of thumb" regarding EPV.
  • Statistical models, including logistic and proportional hazards regression, may yield imprecise or spurious results when the EPV falls below 10-20.
  • This highlights a potential pitfall in multivariable statistical analysis in clinical research.

Conclusions:

  • Monte Carlo techniques provide valuable tools for clinical investigators facing complex analytical challenges.
  • These simulation methods offer an alternative when standard mathematical approaches are insufficient.
  • The findings underscore the importance of considering EPV for reliable statistical inference in clinical studies.