Metacognitive therapy (MCT): research, outcomes and limitations

T There is a difference between symptom change during treatment, results compared with a control group and results maintained at follow-up. The three research sources below, published in 2018–2021, shed light on different parts of this. They are a selection, not a comprehensive overview of the current evidence.

The meta-analysis: which treatments were compared?

Normann and Morina’s systematic review from 2018 included 25 studies and a total of 780 participants. Fifteen studies were randomised controlled trials. The studies’ primary symptom measures were pooled across several diagnoses.

In eight comparisons with cognitive and behavioural treatments, the between-group effect after treatment was Hedges g = 0.69, with a 95% confidence interval of 0.36–1.03, favouring MCT. At follow-up, the corresponding effect was 0.37, confidence interval 0.07–0.66, based on seven comparisons. The small number of studies, their small size, heterogeneity and possible publication bias limit the certainty of the conclusions.

Depression: results after one year

Hjemdal and colleagues’ follow-up from 2019 concerned 39 people with unipolar depression. They had originally been randomised to immediate MCT or a ten-week waiting list. The waiting-list group was then also offered MCT, so no control group remained at the one-year follow-up.

The primary measure was the Beck Depression Inventory, BDI. Recovery was defined as an improvement of at least nine points and a score no higher than 14. In the intention-to-treat analysis, 26 of 39 participants, 67%, met that criterion. Five were missing at follow-up; their last observed values were carried forward in the analysis. The study supports the possibility that improvements after MCT can be maintained, but cannot show that the one-year outcome is better than after another treatment.

GAD: the nine-year follow-up

In a follow-up published in 2021, Solem and colleagues followed up a randomised comparison of MCT and CBT for generalised anxiety disorder, GAD. Of the original 60 participants, 39 took part after an average of just over nine years; 37 provided self-reports.

The primary measure was the Penn State Worry Questionnaire, PSWQ. Recovery required an improvement of at least seven points and a score no higher than 47. In the completer analysis, 12 of 21 MCT participants, 57%, and 6 of 16 CBT participants, 38%, met the criterion.

This is an interesting long-term finding, but the groups are small and 35% did not take part in the follow-up. The percentages describe those who returned questionnaires. They do not in themselves establish a statistically significant difference between the treatments.

Research findings and guidelines

NICE guideline CG113 specifies CBT or applied relaxation when high-intensity psychological treatment is chosen for GAD at step 3. MCT is not listed as a separate option in that recommendation. Research findings and formal guidelines need to be reported separately.

For me, the follow-ups are particularly interesting: what remains when treatment has ended? These studies give us reasons to keep investigating MCT, while also pointing out what larger, independent comparisons need to test.