Group sequential two-stage preference designs.

Area of Science:

• Clinical Trial Design
• Biostatistics
• Health Services Research

Background:

• The two-stage preference design (TSPD) allows for unbiased estimation of treatment efficacy, patient selection effects, and preference effects.
• A significant limitation of TSPD is its requirement for large sample sizes, impacting practical implementation.
• Patient preference significantly influences treatment outcomes, necessitating designs that account for this factor.

Purpose of the Study:

• To propose a group sequential two-stage preference design (GS-TSPD) to enhance the efficiency of TSPD.
• To enable early stopping of clinical trials when sufficient evidence of treatment, selection, or preference effects is found.
• To reduce expected resource allocation in clinical trials by incorporating sequential monitoring.

Main Methods:

• The proposed GS-TSPD integrates sequential monitoring with the traditional TSPD framework.
• Repeated hypothesis tests are conducted on accumulated data during interim analyses to assess study termination criteria.
• Group sequential stopping boundaries are applied to maintain Type I error rates, utilizing approximate sequential density functions.

Main Results:

• Simulations demonstrate that GS-TSPD can achieve desired statistical power with reduced sample sizes compared to traditional TSPD.
• The proposed design effectively controls Type I error rates while allowing for early termination.
• The applicability of GS-TSPD is illustrated through a case study involving Hepatitis C treatment modalities.

Conclusions:

• GS-TSPD offers a statistically rigorous and resource-efficient alternative to traditional TSPD for clinical trials involving patient preferences.
• The design facilitates timely decision-making by enabling early trial closure based on accumulating evidence.
• GS-TSPD is a valuable methodological advancement for studies aiming to understand complex treatment effects influenced by patient choice.