Heroes Not Zombies

The placebo effect is not a zero effect. It’s an effect. It induces change.
Irving Kirsch in his The Emperor’s New Clothes prints a simple bar chart comparing antidepressants, placebo and doing nothing in terms of clinical improvement in patients with depression. That chart shows two things.
The difference between antidepressants and placebo is so small he argues there is no clinically significant difference (apart from in patients with severe depression).
The difference between doing nothing and the effects of giving a patient an antidepressant or a placebo is huge.
In other words, despite the fact there is a barely discernible difference between antidepressants and placebo, he does not recommend doing nothing instead.
Some claim that if a treatment shows no significantly greater effects than placebo we should stop providing that treatment. But if the treatment is replaced with doing nothing there will be a reduction in the benefits experienced by the patients.
All drugs work partially through the placebo effect. In fact, every treatment contains a placebo effect. Randomised controlled trials (RCTs) comparing one group which receives the drug, and one which receives a chemically inert substance (termed placebo) mislead people into thinking there is no placebo effect in the drug group. But there is. The placebo effect is embedded, inextricably.
It is impossible to discern, in an individual case, what proportion of the improvement is due to the direct effects of the treatment and which are due to the placebo effect.
RCTs don’t help clarify that in individual cases. Don’t we need more comparative studies instead? Comparing treatment A to treatment B raises a different set of questions from comparing either to a “placebo group”. As does comparing either treatment group to a “no treatment” group.
I repeat. The placebo effect is not a zero effect.