The placebo question nobody wants to answer

The question is usually a gotcha

People ask this as a gotcha. "Isn't this just placebo?" The expectation is that if the answer is yes, the intervention doesn't count. If the answer is no, it counts. That's not actually how placebo works, and pretending it is lets both sides duck the real question.

Placebo isn't nothing, even by the skeptical benchmark. Hróbjartsson and Gøtzsche's Cochrane review of placebo across all conditions concluded that placebo interventions don't have large general clinical effects, but they did find measurable effects on patient-reported outcomes like pain and nausea, with the caveat that reporting bias may explain part of what's seen. For depression, Kirsch's 2008 PLoS Medicine meta-analysis of FDA antidepressant data famously argued that placebo accounts for most of the clinical response in mild-to-moderate depression. An intervention that's "just placebo" can still help people, and dismissing that is bad reasoning.

The uncomfortable middle answer

Is hypnotherapy partly placebo? Yes, almost certainly. Is it more than placebo? For some conditions, yes. For others, we don't know yet. The trials that use active sham controls (a session that sounds like hypnosis but has no actual induction) tend to find hypnosis beats sham in IBS and chronic pain, roughly matches sham in smoking, and the jury is still out in sleep.

The honest answer, the one that neither fans nor skeptics like, is that hypnotherapy is partly a real mechanistic intervention and partly a really effective placebo delivery system. Both of those are useful. Pretending it's one or the other is how the debate stays stupid.

If you're going to try it, the thing to do is notice whether it works for you and act on that data. The question of why it works matters less than whether it does. Which is, incidentally, the same standard we apply to most psychiatric medications, whose mechanisms we also don't fully understand.