Although we are sympathetic to Reynolds and Slocum's concerns, 1 their editorial errs on several points. The authors cite deference to regulatory (presumably Food and Drug Administration) approval as a constraint. The FDA has long recognized broad physician discretion in using FDAapproved drugs or devices for indications as physicians see fit and has mechanisms for the use of unapproved drugs or devices. That physicians do not venture beyond approved indications may reflect reticence in the face of insurers or peers who misunderstand evidence-based medicine (EBM) by equating EBM with randomized control trials (RCTs). EBM recognizes many levels of evidence. 2 The current version of EBM as equivalent to RCTs fails to understand the epistemic basis underlying statistical approaches in RCTs, as their very structure renders them highly problematic as sole guides to therapy. 3 The alternative of undisciplined or unsubstantiated treatments is unappealing as well; however, Reynolds and Slocum similarly err in proposing that physicians choose treatments according to first principles of physiological mechanisms. Although this would be ideal, first principles are rarely known and when known, most often are derived post hoc. Historically, ''rational therapy'' rarely has been the case, particularly for truly innovative therapies. Take the authors' example of deep brain stimulation (DBS). Its first reported use in movement disorders was in 1980, 4 long before there was any cogent rationale other than that stimulation prior to ablation seemed helpful. Theories of pathophysiology developed in the ensuing years and used as a rational basis for DBS were devised post hoc. 5 Today, there is no clear understanding of how DBS works; certainly there is little evidence to support ''rewiring.'' At best, the history of innovations is one of lucky (in retrospect) intuitive leaps (guesses), guided by prior understanding. As Reynolds and Slocum rightly point out, current interpretations of EBM often lead to radical skepticism and nihilism, particularly in rare disorders, but the alternatives they offer are little better because they are based on a misreading of history and a lack of epistemic analyses. There is ''structure'' to innovation, as evidenced by the fact that certain individuals have made unique and important contributions more consistently than either chance or sociopolitical status can explain. There is a reasonable supposition that these individuals were on the ''right track.'' How they happened onto the right track should be carefully understood and systemized, perhaps through proper philosophical analysis, and applied to facilitate new discoveries.