Abstract:Cluster randomized trials (CRTs) are studies where treatment is randomized at the cluster level but outcomes are typically collected at the individual level. When CRTs are employed in pragmatic settings, baseline population characteristics may moderate treatment effects, leading to what is known as heterogeneous treatment effects (HTEs). Pre-specified, hypothesis-driven HTE analyses in CRTs can enable an understanding of how interventions may impact subpopulation outcomes. While closed-form sample size formulas have recently been proposed, assuming known intracluster correlation coefficients (ICCs) for both the covariate and outcome, guidance on optimal cluster randomized designs to ensure maximum power with pre-specified HTE analyses has not yet been developed. We derive new design formulas to determine the cluster size and number of clusters to achieve the locally optimal design (LOD) that minimizes variance for estimating the HTE parameter given a budget constraint. Given the LODs are based on covariate and outcome-ICC values that are usually unknown, we further develop the maximin design for assessing HTE, identifying the combination of design resources that maximize the relative efficiency of the HTE analysis in the worst case scenario. In addition, given the analysis of the average treatment effect is often of primary interest, we also establish optimal designs to accommodate multiple objectives by combining considerations for studying both the average and heterogeneous treatment effects. We illustrate our methods using the context of the Kerala Diabetes Prevention Program CRT, and provide an R Shiny app to facilitate calculation of optimal designs under a wide range of design parameters.

What problem does this paper attempt to address?

The main problem that this paper attempts to solve is how to design the optimal experimental scheme to evaluate the Heterogeneous Treatment Effects (HTE) in Cluster Randomized Trials (CRTs). Specifically, the researchers have developed new design formulas to determine the cluster size and the number of clusters required to achieve the Locally Optimal Design (LOD) under budget constraints, thereby minimizing the variance of estimating HTE parameters. Since the Intracluster Correlation Coefficient (ICC) is usually unknown in practical applications, the researchers have further developed the Maximin Design to identify the combination of design resources that maximizes the relative efficiency of HTE analysis in the worst - case scenario. In addition, considering that the Average Treatment Effect (ATE) is often the main research interest, the researchers have also established an optimal design method that can consider both ATE and HTE simultaneously. ### Key Points Summary: 1. **Research Background**: - Cluster Randomized Trials (CRTs) are becoming more and more popular in clinical medicine, public health, and implementation science research, especially in cases where treatment contamination needs to be prevented or individual randomization cannot be carried out due to logistical limitations. - When CRTs are applied to real - world scenarios, baseline population characteristics may affect treatment effects, leading to Heterogeneous Treatment Effects (HTE). 2. **Research Objectives**: - Develop new design formulas to determine the cluster size and the number of clusters required to achieve the Locally Optimal Design (LOD) under budget constraints, thereby minimizing the variance of estimating HTE parameters. - Develop the Maximin Design to identify the combination of design resources that maximizes the relative efficiency of HTE analysis in the worst - case scenario. - Considering that ATE is usually the main research interest, establish an optimal design method that can consider both ATE and HTE simultaneously. 3. **Methods and Results**: - **Locally Optimal Design (LOD)**: By deriving a closed - form solution, determine the cluster size and the number of clusters that minimize the variance of HTE estimation under a given budget and a fixed ICC value. - **Maximin Design**: Through a search process, identify the design with the highest relative efficiency (RE) within a given range of ICC values. This helps to deal with the uncertainty of ICC values in the design stage. - **Multi - objective Optimal Design**: Combining the considerations of ATE and HTE, propose a design method that can balance among multiple objectives. 4. **Application Examples**: - Use the data of the Kerala Diabetes Prevention Project (K - DPP) to illustrate how to use the newly proposed optimal design method to determine the number of clusters and the cluster size required to maximize power under a fixed total budget. - Provide a free R Shiny application for exploring optimal designs in a wider range of practical scenarios. ### Conclusion: This paper provides a systematic solution for HTE evaluation in CRTs by developing new design formulas and the Maximin Design method, especially in cases where the ICC value is uncertain. These methods not only help to improve the statistical power of research but also optimize resource allocation under a limited budget.

Maximin optimal cluster randomized designs for assessing treatment effect heterogeneity

Sample size requirements for testing treatment effect heterogeneity in cluster randomized trials with binary outcomes

Optimal designs using generalized estimating equations in cluster randomized crossover and stepped wedge trials

Sample size considerations for stratified cluster randomization design with binary outcomes and varying cluster size

Using Power Analysis to Choose the Unit of Randomization, Outcome, and Approach for Subgroup Analysis for a Multilevel Randomized Controlled Clinical Trial to Reduce Disparities in Cardiovascular Health

A Utility-Based Design for Randomized Comparative Trials with Ordinal Outcomes and Prognostic Subgroups

Covariate-constrained randomization in cluster randomized 2 × 2 factorial trials: application to a diabetes prevention study

Leveraging baseline covariates to analyze small cluster-randomized trials with a rare binary outcome

Leveraging contact network structure in the design of cluster randomized trials

Two-Stage TMLE to reduce bias and improve efficiency in cluster randomized trials

Methods for sample size determination in cluster randomized trials

Methods for dealing with unequal cluster sizes in cluster randomized trials: A scoping review

Design-Based RCT Estimators and Central Limit Theorems for Baseline Subgroup and Related Analyses

An Adaptive Trial Design to Optimize Dose-Schedule Regimes with Delayed Outcomes

Handling incomplete outcomes and covariates in cluster-randomized trials: doubly-robust estimation, efficiency considerations, and sensitivity analysis

CRTpowerdist: An R package to calculate attained power and construct the power distribution for cross-sectional stepped-wedge and parallel cluster randomized trials

Blurring cluster randomized trials and observational studies: Two-Stage TMLE for subsampling, missingness, and few independent units.

Blurring cluster randomized trials and observational studies using Two-Stage TMLE to address sub-sampling, missingness, and minimal independent units

Comparing cluster-level dynamic treatment regimens using sequential, multiple assignment, randomized trials: Regression estimation and sample size considerations

Optimizing Trial Designs for Targeted Therapies

Minimax optimal subgroup identification