Beyond Randomized Clinical Trials: Use of External Controls

Area of Science:

• Clinical Trials Methodology
• Real-World Evidence (RWE) Integration
• Biostatistics

Background:

• Randomized controlled trials (RCTs) are the standard for evaluating treatment efficacy and safety.
• Ethical or feasibility constraints can hinder concurrent control group randomization in certain clinical settings.
• External data sources offer potential to reduce or replace concurrent controls.

Purpose of the Study:

• To propose and illustrate principled methods for utilizing external control data in clinical trials.
• To demonstrate how to mitigate biases arising from patient and confounding factor differences.
• To reduce reliance on concurrent control groups through robust data integration techniques.

Main Methods:

• Application of meta-analytic techniques for borrowing strength from past trials.
• Utilization of propensity score methods, including matching, for constructing external controls from real-world data (RWD).
• Bayesian meta-analytic approach for integrating historical trial data.

Main Results:

• Successful construction of external controls from RWD using propensity score matching in a rare cancer trial.
• Effective borrowing of strength from past trials via a Bayesian meta-analytic approach in a pediatric multiple sclerosis trial.
• Demonstrated validity of proposed methods for incorporating external data.

Conclusions:

• Meta-analytic and propensity score methods provide a rigorous framework for using external controls.
• External control data can effectively reduce or eliminate the need for concurrent control groups.
• These approaches enhance the feasibility and ethical conduct of clinical trials, particularly in rare diseases.