Hypnotherapy for generalised anxiety: the evidence

Generalised anxiety disorder (GAD) (ongoing, hard-to-control worry) is Grade B for hypnotherapy. That is the question this hub answers. Trials exist. They do not yet add up to the independent, protocol-stable programme that grades IBS as A. Procedure-day distress is a different, stronger pile and lives on its own Grade A page. It is not a substitute grade for GAD. Grades are explained on how we evaluate evidence.

Key evidence

Valentine, Milling, Clark, and Moriarty (2019) meta-analysed 15 studies (17 trials) that compared hypnosis with a control. The mean weighted effect size at the end of treatment was 0.79. That number is a standardised mean difference, not a clinic success rate. In the authors’ percentile reading, the average hypnosis participant improved more than about 79% of control participants. It is not “79% of patients were cured.”

Those 17 trials were not GAD. Three were dental, three surgery or medical, four general anxiety, and seven test or performance anxiety. The authors themselves note that only four trials addressed anxiety that could be called chronic. That four-trial subgroup (effect 0.93) is the closest published pile to GAD, and it is still a subgroup, not a multi-group GAD programme. Effect sizes were larger against wait-list or no-treatment (1.12) than against standard care or attention (0.60). Combining hypnosis with another psychological intervention beat stand-alone hypnosis.

Coelho, Canter, and Ernst (2007), as abstracted by DARE, reviewed 14 RCTs of clinical or subclinical anxiety and concluded that the evidence was insufficient. The included trials were small and mostly low quality. That older filter is why this hub does not jump to Grade A on the back of Valentine’s mixed pile.

NICE CG113 is the UK adult guideline for generalised anxiety disorder and panic. This wiki does not cite it as a hypnotherapy recommendation. Stepped psychological care in that pathway is not a hypnosis protocol.

Suggestion and expectancy are defined on suggestion, and response expectancy. They are methods words, not an outcome grade.

Limitations

“Anxiety” in a trial title can mean a STAI score before an exam, a dental visit, or a GAD diagnosis. Pooling them inflates confidence in GAD. Heterogeneity in Valentine was significant. Clinic session counts and clinic percentages are not trial endpoints.

Social anxiety, health anxiety, panic, and child anxiety are separate clinical entities. They are not 301’d into this hub.

What we don't know yet

A large, independently run RCT of one named hypnotherapy protocol versus CBT for DSM-diagnosed GAD is still the missing trial. See what we don’t know.

This page is informational only. It is not personal medical advice. Hypnotherapy does not replace assessment, first-line psychological care, or medicines a clinician has prescribed.