Placebo and Active Controls in Meditation Trials: Why Comparisons Matter

A new medicine is usually tested against a look-alike pill that nobody can tell apart. Meditation cannot be tested that way, because people know when they are meditating. This page explains how researchers cope with that, why the choice of comparison group changes what a result means, and what the better-controlled studies have found. It is information only. For a checklist you can use on any study, see how to read a meditation study.

Why blinding is the problem

Two leading researchers write that the classic double-blind, placebo-controlled design cannot be used for meditation, because participants know if they are assigned to a meditation condition. [1] They say this is partly responsible for the poor quality of many meditation trials, and that active comparison treatments are required. [1] The result is that expectation, a hope that something will help, can shape what people report. That is a real limit on many positive findings.

The main kinds of comparison

Most of the literature has used the first kind. One study team noted that most research on the health effects of mindfulness interventions relies on wait-list comparisons. [2] A large expert review cited a 2015 estimate (by Dimidjian and Segal) that, of all mindfulness-program research, about 20% compared the program with waitlist or usual care while only about 9% compared it with an active control. [3]

What a good active control looks like

A US team built and tested the Health Enhancement Program (HEP) as a comparison for mindfulness-based stress reduction (MBSR). Both were matched on group format and instructor training, [1] and on the explicit expectation that the program is beneficial and the requirement to practice daily. [1] In their trial, 63 people were randomized, 31 to MBSR and 32 to HEP. [2]

What they found:

• MBSR led to lower thermal pain ratings in a task that used mindfulness instructions, but not in a task that used HEP-related instructions. [2]
• People in both groups reported improved general distress, anxiety and other measures, but the study found MBSR no more effective than a rigorous active control on these self-reported outcomes. [2]

Plain reading: in this trial, the improvement people reported may have come largely from things the two programs shared, such as group support, a trained instructor and expectation. That does not mean the programs are worthless. It means a waitlist comparison alone would have made MBSR look better than it was.

Sham meditation and placebo comparisons

Some laboratory studies test meditation against placebo directly in healthy volunteers. In one, 75 healthy volunteers were randomly assigned to 4 days of mindfulness meditation, placebo conditioning, sham mindfulness meditation or book listening. [4] All three mental approaches lowered pain ratings compared with rest and the control condition. [4] Mindfulness meditation reduced pain intensity and unpleasantness more than placebo cream, and more than sham meditation, and brain imaging suggested different brain regions were involved. [4] A later analysis pooled two such trials of 115 healthy participants and reported that meditation and placebo engaged distinct neural pain signatures. [5]

How to read this: the participants were healthy volunteers given short-term heat pain in a laboratory, and the results show how a mechanism may differ. They do not show how well meditation treats long-term pain in patients. Evidence label: promising but limited; preclinical or mechanism-level for the brain findings. Brain-imaging changes are not the same as health outcomes.

What reviews that use active controls found

A major 2014 review in JAMA Internal Medicine included only trials with active controls and found 47 trials with 3515 participants. [6] It found moderate evidence that mindfulness programs improved anxiety, depression and pain, and low evidence for stress and quality of life. [6] It found no evidence that meditation programs were better than any active treatment such as drugs, exercise and other behavioral therapies. [6] NCCIH describes a 2018 analysis in which mindfulness-based approaches were better than no treatment for anxiety and depression and worked as well as established evidence-based therapies. [7] These fit together: meditation tends to beat nothing, and to match other reasonable treatments, rather than to beat them.

What to take from this

• "Better than a waitlist" is a lower bar than "better than another credible program".
• "No better than an active control" can still mean "useful", because the comparison may be a good treatment.
• Describing a trial as "meditation training" can itself raise expectations, so one expert source suggests plain descriptions, such as training in stabilizing attention. [1]
• Explore the main findings by condition in meditation for stress, sleep and anxiety, and for the wider bias picture see quality and bias in meditation research.