sample size

June 19, 2020: Living Textbook Grand Rounds Series: Part 4-Demystifying Biostatistical Concepts for Embedded Pragmatic Clinical Trials (Elizabeth Turner, PhD; Patrick Heagerty, PhD; David Murray, PhD)

Posted on by Gina Uhlenbrauck

Speakers

Elizabeth Turner, PhD
Associate Professor
Department of Biostatistics & Bioinformatics
Duke Global Health Institute
Duke University  

Patrick Heagerty, PhD
Professor Department of Biostatistics
University of Washington  

David Murray, PhD
Associate Director for Prevention
Director, Office of Disease Prevention National Institutes of Health

Topic

Demystifying Biostatistical Concepts for Embedded Pragmatic Clinical Trials

Keywords

Embedded PCTs; Biostatistics; Trial design; Cluster-randomized trial (CRT); Stepped-wedge; Intraclass correlation coefficient; NIH Collaboratory Trial; Sample size; Individually randomized group treatment

Key Points

  • Focus on the research question, because that will drive the design, and the design will drive the analysis.
  • Select design features with analysis in mind, and collaborate early with a statistician. Weigh statistical choices against the challenges of implementation.
  • If possible, choose individual randomization. However, sometimes there is a strong rationale for choosing cluster/group randomization. Clustering must be accounted for in both design and analysis for CRTs and individually randomized group treatment (IRGT) trials.
  • The intraclass correlation coefficient (ICC) is a common measure of outcome clustering. Estimating the ICC is needed for study planning and power.
  • Increasing the number of clusters has more impact on power than increasing the number of patients per cluster.

Discussion Themes

With the move to virtual healthcare, the boundaries between clinic-based clusters have become more fluid. What approaches should trials use to describe contamination and estimate the impact of contamination on outcomes?

Read more about ICC in a Living Textbook resource and visit the Training Resources page for practical help on how to plan and conduct ePCTs.

Learn more in the Living Textbook about considerations for trial design and analysis for ePCTs.

Visit the NIH Collaboratory’s Biostatistics and Study Design Core webpage for more resources around design and analysis issues in ePCTs.

The NIH hosts a Research Methods Resources website with materials on this topic.

Tags

#pctGR, @Collaboratory1

March 17, 2020: Cheat Sheet on the Intraclass Correlation Coefficient

Posted on by Karen Staman

The NIH Collaboratory Biostatistics and Study Design Core has created an Intraclass Correlation Coefficient (ICC) Cheat Sheet to provide an introductory description of the ICC, which is important for the design and analysis of cluster-randomized trials.

“The intraclass correlation coefficient (ICC) is a descriptive statistic that describes the extent to which outcomes 1) within each cluster are likely to be similar or 2) between different clusters are likely to be different from each other, relative to outcomes from other clusters. The ICC is an important tool for cluster-randomized pragmatic trials because this value helps determine the sample size needed to detect a treatment effect.” —from the ICC Cheat Sheet

The tool is a 2-page handout that can be used in trainings or classes regarding pragmatic clinical trials involving cluster randomization.

For more on the ICC, see the Intraclass Correlation section in the Living Textbook or this in-depth working document on the ICC from the Biostatistics and Study Design Core. If you have questions, feedback or suggestions regarding this tool, please contact us at nih-collaboratory@dm.duke.edu.