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

Cluster randomization trials in tropical medicine: a case study

N Klar1, T Gyorkos, A Donner

  • 1Department of Epidemiology and Biostatistics, University of Western Ontario, London, Canada.

Transactions of the Royal Society of Tropical Medicine and Hygiene
|July 1, 1995
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

Self-Diffusion in Amorphous Silicon by Local Bond Rearrangements.

Physical review letters·2018
Same author

Bortezomib, C1-inhibitor and plasma exchange do not prolong the survival of multi-transgenic GalT-KO pig kidney xenografts in baboons.

American journal of transplantation : official journal of the American Society of Transplantation and the American Society of Transplant Surgeons·2015
Same author

Propensity scores used for analysis of cluster randomized trials with selection bias: a simulation study.

Statistics in medicine·2013
Same author

An evaluation of utility measurement in Crohn's disease.

Inflammatory bowel diseases·2013
Same author

Estimation of the secondary attack rate for delta hepatitis coinfection among injection drug users.

The Canadian journal of infectious diseases = Journal canadien des maladies infectieuses·2012
Same author

[Endoscopic calcaneoplasty (ECP) in Haglund's syndrome. Indication, surgical technique, surgical findings and results].

Zeitschrift fur Orthopadie und Unfallchirurgie·2011

Cluster randomized trials in tropical medicine require careful analysis to avoid incorrect conclusions. This study compares methods for analyzing binary outcomes in such designs, emphasizing accurate treatment effect assessment.

Area of Science:

  • Tropical Medicine
  • Biostatistics
  • Epidemiology

Background:

  • Field trials in tropical medicine frequently employ cluster randomization, where social units (families, schools, communities) are randomized instead of individuals.
  • This design is chosen for administrative ease, minimizing treatment contamination, and addressing ethical considerations.
  • Dependencies within clusters necessitate specific statistical approaches for sample size calculation and data analysis to prevent erroneous findings.

Purpose of the Study:

  • To compare various statistical methods for testing treatment effects in cluster randomized trials with binary outcomes.
  • To highlight the importance of accounting for dependencies among cluster members in data analysis.
  • To illustrate these methods using a real-world example from a trial on intestinal parasites.

Related Experiment Videos

Main Methods:

  • The study focuses on methods for analyzing binary outcome data from cluster randomized trials.
  • It involves comparing different statistical tests designed to accommodate the clustered nature of the data.
  • The methods are demonstrated using data from a family-based trial for imported intestinal parasites.

Main Results:

  • Failure to account for cluster dependencies can lead to inaccurate conclusions regarding treatment efficacy.
  • Appropriate statistical methods are crucial for valid interpretation of results in cluster randomized trials.
  • The paper provides a comparative overview of analytical techniques suitable for binary outcomes in such designs.

Conclusions:

  • Accurate statistical analysis is essential in cluster randomized trials to correctly determine treatment effectiveness.
  • The choice of analytical method significantly impacts the validity of conclusions, especially with binary outcomes.
  • This research underscores the need for specialized statistical approaches when dealing with clustered data in public health interventions.