Background: The majority of modern acute stroke trials utilize trial designs that require the prioritization of a single primary outcome over other, potentially equally important, outcomes.
Aims: To demonstrate how a Desirability of Outcome Ranking (DOOR) design can facilitate comparison of a telemedicine neurological assessment to an onboard assessment whilst evaluating multiple outcomes simultaneously, within the upcoming Melbourne Mobile Stroke Unit (MSU) Telemedicine Trial.
Methods: A DOOR design combines multiple ranked measures into an overall outcome. All participants within each arm of the trial will be compared to those in the other using a Win Odds method, resulting in a "win, loss, or draw" for telemedicine, compared with the onboard arm. Our primary study endpoint will be the odds that a random participant treated through a MSU telemedicine assessment will have a more desirable DOOR scale outcome than a random participant treated by an onboard MSU neurologist. Recruiting 242 participants (121 per arm) would yield a power of 0.8 to observe such treatment effect against the null hypothesis (Win Odds = 1; 2-sided alpha level=0.05).
Results: The design evaluates, in order of importance: safety, scene-to-decision time metrics, and resource utilization. If a participant in one treatment arm is achieving better safety than the comparator, this is defined as a “win” for that participant and a “loss” for the comparator. If there is no difference in safety, time to treatment decision is compared. If no clinically meaningful difference is observed, then resource utilization is compared. If there is no difference in resource utilization, the two participants are declared as tied for the overall outcome.
Conclusion: A DOOR design clearly articulates the relative importance of multiple critical variables included in a primary outcome measure, whilst allowing for the development of a trial with a feasible sample size.