Preference option randomized design (PORD) for comparative effectiveness research: Statistical power for testing...

Area of Science:

• Clinical Trials
• Health Services Research
• Biostatistics

Background:

• Randomized clinical trials (RCTs) are a gold standard but may limit participant choice in comparative effectiveness research.
• Forcing participants to adhere to a randomly assigned intervention can lead to practical limitations and reduced retention.

Purpose of the Study:

• To propose and validate a novel trial design, the preference option randomized design (PORD), that accommodates participant preferences post-randomization.
• To enhance participant retention and minimize forced assignments in comparative effectiveness research while preserving randomization.

Main Methods:

• The preference option randomized design (PORD) involves an informed consent process before randomization, allowing participants to opt out of their assigned intervention only if they actively express this desire after randomization.
• Participants who do not express an alternate preference are considered to have no preference and are retained in their assigned arm.
• The design defines five interconnected effects: comparative, preference, selection, intent-to-treat, and overall/as-treated.

Main Results:

• Statistical power functions were derived for testing all defined effects within the PORD framework.
• Simulations confirmed the validity of these power functions under both normal and binomial distributions.

Conclusions:

• The preference option randomized design (PORD) offers a flexible alternative to traditional RCTs for comparative effectiveness research.
• PORD aims to maximize participant retention and minimize forced assignments by integrating participant preference considerations within a randomized framework.