Preference-based controlled design: toward increased patients' engagement, efficiency, and external validity of...

Randomized controlled trials (RCTs) are the gold standard for evaluating clinical interventions, but their limitations-particularly in settings where strong patient preferences exist-can impair recruitment, compliance, and external validity. In cardiovascular research, these challenges are pronounced in trials comparing treatments that differ substantially in invasiveness, burden, or delivery mode. We describe how patient preference-controlled designs can enhance trial efficiency and generalizability by engaging patients as active participants in the allocation process. Although nonrandom allocation introduces potential for confounding, we argue that these trials can produce valid and robust treatment effect estimates when key confounding influences-such as physician bias, functional status, and health beliefs-are prospectively measured and controlled. Preference, under appropriate design conditions, may function as a valid instrumental variable. Furthermore, we introduce a hybrid trial design that incorporates both randomized and preference-based cohorts. This approach allows for separate and combined effect estimates, preserving the internal validity of randomization while expanding recruitment and improving external validity. Bayesian modeling frameworks can be used to integrate treatment effect estimates across cohorts while adjusting for differences in preference strength and confounding structure. In summary, patient preference-controlled trials represent a rigorous, underutilized methodology in cardiovascular research. By prospectively addressing and quantifying bias, they offer a practical solution to the limitations of both traditional RCTs and retrospective observational studies. We advocate for the development of formal design and reporting guidelines and encourage their consideration in trials where patient empowerment and generalizability are critical.