Optimal design for matched pair cluster randomized trials with heterogeneous correlations and costs

Area of Science:

• Biostatistics
• Clinical Trials Methodology
• Health Economics

Background:

• Cluster randomized trials (CRTs) are essential for evaluating interventions in group-level settings but are often expensive.
• Budget constraints necessitate optimal design strategies for CRTs, particularly matched pair designs.
• Standard CRT designs often assume uniform costs and correlations, which is unrealistic in practice.

Purpose of the Study:

• To propose optimal subject allocation methods for matched pair cluster randomized trials (CRTs) under budget constraints.
• To minimize the variance of the treatment effect estimator in CRTs with heterogeneous parameters.
• To enhance the efficiency of matched pair CRTs compared to traditional balanced designs.

Main Methods:

• Developed a novel approach for optimal subject allocation within clusters for matched pair CRTs.
• Derived allocation formulas by minimizing treatment effect estimator variance under general conditions.
• Accounted for heterogeneity in intra-class correlation, matching correlation, and sampling costs.
• Explored min-max and pseudo-Bayesian optimal designs to handle unknown parameters.

Main Results:

• The proposed optimal allocation design is more efficient than the standard balanced design for matched pair CRTs.
• The method effectively minimizes variance even with heterogeneous intra-class correlation, matching correlation, and sampling costs.
• Numerical examples using real-world data validate the theoretical findings and demonstrate practical applicability.

Conclusions:

• Optimal subject allocation is crucial for efficient and cost-effective matched pair CRTs.
• The proposed methods provide a robust framework for designing CRTs under realistic conditions of heterogeneity and budget limitations.
• This research offers practical guidance for optimizing resource allocation in cluster randomized trials.