Covariate-constrained randomization in cluster randomized 2x2 factorial trials: Application to a diabetes prevention...

Area of Science:

• Clinical Trials Methodology
• Biostatistics
• Public Health Research

Background:

• Cluster randomized trials (CRTs) involve randomization at the group level, posing challenges for covariate balance with small cluster numbers.
• Imbalance in cluster-level covariates can bias treatment effect estimates, reduce statistical power, and increase outcome variability in CRTs.
• Covariate-constrained randomization (CR) is a strategy to mitigate covariate imbalance in CRTs, with methods previously undeveloped for factorial designs.

Approach:

• Developed and evaluated covariate-constrained randomization (CR) methods specifically for 2x2 factorial cluster randomized trials.
• Applied CR methods to the BEGIN study, a CRT focused on weight loss interventions for pre-diabetes.
• Utilized simulation studies to compare CR against simple randomization, assessing performance across various factors like cluster number and covariate association with outcomes.

Key Points:

• Covariate-constrained randomization (CR) effectively balances cluster-level covariates in factorial CRTs compared to simple randomization.
• CR leads to more precise treatment effect inferences in factorial CRTs.
• Including cluster-level covariates in analysis models increases power for detecting treatment effects, though power may be reduced with very few clusters.

Conclusions:

• Covariate-constrained randomization (CR) is recommended over simple randomization for factorial CRTs to prevent imbalanced designs and enhance inference precision.
• Analysis models should incorporate cluster-level covariates in factorial CRTs, except when dealing with a minimal number of clusters, to optimize power and statistical validity.
• These findings support the adoption of CR for more robust and reliable outcomes in complex factorial CRT designs.