Design of randomized controlled confirmatory trials using historical control data to augment sample size for...

Area of Science:

• Biostatistics
• Clinical Trials
• Epidemiology

Background:

• Augmenting concurrent controls (CC) in randomized controlled trials (RCTs) with historical data is crucial for clinical study efficiency.
• Existing methods, like Stuart and Rubin's matching approach, may not sufficiently save sample sizes when historical data is used as a secondary control group.

Purpose of the Study:

• To propose a novel matching method for RCTs that effectively utilizes historical data to save sample size.
• To introduce a two-stage design for validating the proposed matching method's assumptions before full implementation.

Main Methods:

• A new propensity score estimation technique is developed, excluding concurrent randomized control patients to maximize sample size savings.
• A two-stage adaptive design is proposed to assess the validity of the new matching method.
• The proposed methods are illustrated using previous clinical trial data and validated through simulation studies.

Main Results:

• The novel matching method demonstrates potential for significant sample size reduction in RCTs augmented with historical data.
• Simulation studies confirm the operating characteristics and feasibility of the proposed approach.
• The two-stage design provides a framework for reliable implementation of the sample-saving method.

Conclusions:

• The proposed matching method offers an effective strategy for augmenting concurrent controls with historical data, leading to improved sample size efficiency in clinical trials.
• The two-stage design ensures the robustness and applicability of the method in real-world clinical research settings.