On 2016 Jul 28, John Tucker commented:

As is so often the case in Cochrane Reviews, the authors find the data supporting the efficacy of the topic medical intervention wanting. While much of the criticism found herein is valid, one is struck by the authors' proposal that the only way to accurately assess the utility of methylphenidate in ADHD is a "nocebo trial", in which the test article would be compared to a control substance having no efficacy but having an identical side effect profile.

What would such a substance be? How would one determine that its side effect profile was identical? Or that it had no intrinsic efficacy in ADHD of its own? Perhaps by running one or more RCTs? But since the nocebo is designed to have side effects, wouldn't these trials be unblinded too? Given this unblinding, could a Cochrane Review confidently reach the conclusion that the nocebo had no intrinsic efficacy? What would be the ethical issues involved in performing safety trials of an agent designed to have side effects and no benefits?

While the contributions of the Evidence-Based Medicine movement are incontrovertable, at some point its advocates need to take a deep breath and ask themselves if they have followed their logic into a world that exists only in theory. We need to apply the best standards that can reasonably be applied in evaluating any medical intervention. But holding interventions to standards of evidence that for all practical purposes are unachievable risks drifting into therapeutic nihilism.

On 2016 Jul 28, John Tucker commented:

As is so often the case in Cochrane Reviews, the authors find the data supporting the efficacy of the topic medical intervention wanting. While much of the criticism found herein is valid, one is struck by the authors' proposal that the only way to accurately assess the utility of methylphenidate in ADHD is a "nocebo trial", in which the test article would be compared to a control substance having no efficacy but having an identical side effect profile.

What would such a substance be? How would one determine that its side effect profile was identical? Or that it had no intrinsic efficacy in ADHD of its own? Perhaps by running one or more RCTs? But since the nocebo is designed to have side effects, wouldn't these trials be unblinded too? Given this unblinding, could a Cochrane Review confidently reach the conclusion that the nocebo had no intrinsic efficacy? What would be the ethical issues involved in performing safety trials of an agent designed to have side effects and no benefits?

While the contributions of the Evidence-Based Medicine movement are incontrovertable, at some point its advocates need to take a deep breath and ask themselves if they have followed their logic into a world that exists only in theory. We need to apply the best standards that can reasonably be applied in evaluating any medical intervention. But holding interventions to standards of evidence that for all practical purposes are unachievable risks drifting into therapeutic nihilism.