Evidence-based practice, according to authoritative statements by the founders of this approach to health care, involves using the "best available" evidence in addition to clinical expertise and patient preferences to make decisions on the care of patients. However, many systematic reviewers interpret "best available" as "best possible" and exclude from their reviews any evidence produced by research of a grade less than the highest possible (e.g., the randomized clinical trial [RCT] for interventions), even if that means making no recommendations at all. Voltaire's comment that "the best is the enemy of the good" is applicable here. Rehabilitation would be disadvantaged especially, as it can boast few RCTs, because of its nature. The myopic focus on the "strongest" research designs may also steer researchers away from asking, "What is the best design to answer 'this' research question?" Lastly, rehabilitation and other clinicians need to know not just "which" interventions are effective, but also "how" these interventions need to be delivered; information relevant to this latter aspect of knowledge translation is typically produced using "weak" research designs.