Partner Introduction (SPOR Evidence Alliance): 2-min video.

• Summary: Sharmila Sreetharan (Research Coordinator, Central Coordinating Office) discusses the SPOR Evidence Alliance, a pan-Canadian research initiative supported by the Canadian Institutes of Health Research under Canada’s Strategy for Patient-Oriented Research (SPOR).

Pragmatic and Group-Randomized Trials in Public Health and Medicine Online Course (Part 1: Randomized Designs for Clinical Trials) - National Institutes of Health (Office of Disease Prevention): 23-min webinar.

• Summary: Dr. David Murray overviews three types of randomized trials: (1) individual-level randomized trials, (2) individually randomized group treatment trials, and (3) group-randomized trials (also known as cluster randomized trials); includes both parallel- and stepped-wedge randomized trials. Guidance on how to select among these designs is provided, as well as a review of allocation/randomization techniques that may be used in cluster randomized trials (to reduce covariate imbalance at baseline): i.e., matching, stratification, and constrained randomization.

Hemming K, et al. What type of cluster randomized trial for which setting? J Epidemiol Popul Health. 2024 Feb;72(1):202195. (4-page paper)

• Summary: The cluster trial allows a randomized evaluation when (1) it is either not possible to randomize individuals or (2) randomizing individuals would put the trial at high risk of contamination across the intervention groups. This paper outlines the different possible options when considering the cluster design and weighs the pros and cons (in relation to statistical efficiency, study logistics, and assumptions required) of the different design choices.

Advances in the design and analysis of individually randomized group treatment trials (HDRN Canada Pragmatic Trials Training Program): 5-min video.

• Summary: Dr. Guangyu Tong discusses advances in the design and analysis of individually randomized group treatment trials. Connections are made to related topics, such as cluster randomized trials, and more specifically, stepped-wedged cluster trials. The need for biostatisticians to get involved in this space is also emphasized. A related technical paper can be found here. * An institutional login (e.g., university or research institute e-mail address) is required to access this material.

NIH Pragmatic Trials Collaboratory (Living Textbook): Section 5 - Stepped-Wedge Designs (Chapter: Design - Experimental Designs and Randomization Schemes): 1-page website.

• Summary: Overviews stepped-wedge designs, whereby clusters are randomized into several groups (“waves”) that define when the intervention will begin. All clusters start the trial in the control state (no intervention), and groups of clusters cross over to experience the intervention over time on a staggered schedule. All groups ultimately receive the intervention before the end of the trial. The advantages of stepped-wedge designs, as well as the ethical considerations and analytical challenges are also discussed.

Ivers NM, et al. Allocation techniques for balance at baseline in cluster randomized trials: a methodological review. Trials. 2012 Aug 1;13:120. (9-page paper)

• Summary: The risk for imbalance of covariates at baseline requires special attention in cluster trials. Imbalance at baseline decreases statistical power, the precision of the results, and reduces face validity and credibility of the findings. A variety of restricted randomization methods have been proposed to minimize the risk of imbalance. This paper summarizes the advantages and limitations of different allocation/randomization techniques, i.e., stratification, matching, minimization, and covariate-constrained randomization.

Hemming K, et al. How to design efficient cluster randomised trials. BMJ. 2017 Jul 14;358:j3064. (5-page paper)

• Summary: Cluster randomized trials have diminishing returns in statistical power and precision of the effect estimates as the cluster size increases. This paper overviews the trade-offs that are made when determining the number of clusters vs. the size of clusters and introduces methods that will enable both (1) researchers to design efficient cluster trials and (2) the scientific community (including funders of trials) to better appraise the efficiency of cluster trials that are being completed.

NIH Pragmatic Trials Collaboratory - What Are the Arguments For and Against the Stepped-Wedge Design? (October 17, 2023): 2-min video.

• Summary: Dr. Liz Turner discusses arguments for and against using the stepped-wedge design. It is ultimately recommended to use a parallel design, when possible, but if a stepped-wedge design is necessary, plan to account for the impact of time effects in the design and analysis.

NIH Pragmatic Trials Collaboratory - Pair-Matching vs Stratification in Cluster Randomized Trials (June 26, 2014): 3-page document.

• Summary: Pair matching and stratification are two allocation/randomization techniques in cluster trials. With pair matching, clusters are paired in terms of their potential confounders and, subsequently, within each pair, one cluster is randomized to receive either the intervention or the control condition, and the other cluster receives the opposite. When doing stratification, strata are formed based on the potential confounders, and within each stratum, a randomization scheme that ensures balance is developed.

NIH Pragmatic Trials Collaboratory (Living Textbook): Section 7 - Covariate-Constrained Randomization (Chapter: Design - Experimental Designs and Randomization Schemes): 1-page website.

• Summary: Overviews covariate-constrained randomization, which is one of the  techniques that can provide better baseline balance of covariates in cluster trials than other methods (e.g., simple random allocation, stratification, and minimization); this technique is particularly useful when the number of randomized units is small (e.g., less than 20 clusters).