Research on MCT and ISTDP.
Here I bring together a selection of research on MCT and ISTDP, the two approaches I work with. You can read the summaries, search in your own words or filter by approach.
Responsible for content: Andreas Lundin, psychologist licensed in Sweden.
The summaries are based on the original publications. Each text gives the study design, sample and limitations to help you understand how the results can be interpreted. There is also a link to the original publication and to a longer discussion where one is available.
Diagnostic and symptom terms are used when they form part of the studies’ samples, measures or results. They show what the research examined, but do not by themselves describe how a person is affected or what needs to change.
-
Detached mindfulness as a stand-alone intervention: large changes in three small clinical studies, but limited evidence
Detached mindfulness as a stand-alone intervention: a Systematic Review and meta-analysis
Detached mindfulness (DM) is a core technique in metacognitive therapy. The person practises noticing an intrusive thought, stepping back and allowing it to be there without continuing to process it or fight it. The technique is usually part of a full MCT protocol. This systematic review and meta-analysis examined DM as a stand-alone intervention. Fourteen studies, representing twelve independent samples, were included. In total, 256 people aged 17 and over had received DM.
Three samples were clinical trials, two concerning obsessive-compulsive disorder and one concerning panic disorder. The pooled pre-to-post change was large on the studies’ primary outcomes: Hedges g = −1.80 (95% CI −2.84 to −0.76). The change on the depression measure, a secondary outcome, was also large: g = −1.15 (95% CI −2.23 to −0.08). The other nine samples were mainly experimental and non-clinical. Eight reported results favouring DM on at least one outcome, most often with medium or large effect sizes. The authors describe the overall picture as consistent but still limited, and see the results mainly as a basis for further research on DM as a component of MCT and as a brief, transdiagnostic intervention.
LimitationsThe authors consider the evidence limited. The studies are few, small and methodologically uneven. Only one followed up participants, after four weeks. The clinical meta-analysis is based on three studies with 48 people in total; two had no control group, and results varied considerably between studies (I² = 77.7%). Only published studies were included, so reporting bias cannot be ruled out. The pooled effects compare change within groups before and after the intervention. They therefore do not show how much better DM is than no treatment or another treatment.
-
Generic group MCT for mixed conditions: large changes in a small study without a control group
Generic Group Metacognitive Therapy for Mixed Common Mental Disorders: An Open Trial with Six-Month Follow-up in Specialized Mental Health Care
The study starts from the observation that depression and anxiety diagnoses often overlap and co-occur. MCT targets processes that may be shared across diagnoses, such as worrying and ruminating, and similar treatment elements can be used with different diagnoses. The researchers therefore examined a generic MCT programme in groups with mixed conditions. Thirty-four participants with a primary depression or anxiety diagnosis and high comorbidity received ten sessions in specialist mental health care. Change was analysed for the whole group and three diagnostic subgroups.
In the full sample, the researchers found statistically significant improvements on all outcomes, with large within-group effect sizes both from before to after treatment and at the six-month follow-up. In the diagnostic subgroups, changes on diagnosis-specific measures were large. The researchers also found improvements in broader functioning, quality of life and metacognitive beliefs. At the end of treatment, 27 of 34 participants, 79%, were classified as recovered on the measure for their primary diagnosis, based on reliable change and a clinical cut-off. None of the 34 dropped out of treatment. By the study’s calculation, the intervention amounted to an average of about seven 45-minute therapist hours per participant. The authors consider the group format feasible and potentially resource-efficient in routine specialist mental health care.
LimitationsThe trial was open and had no control group, so it cannot establish that MCT caused the changes or compare the group format with another intervention. The sample was small and drawn from one specialist clinic. Outcomes were based on self-report, therapist effects could not be ruled out, and 28 of 34 participants provided data at the six-month follow-up. Follow-up was limited to six months. No formal health economic analysis was conducted, so resource efficiency is a possibility, not established cost-effectiveness. The authors call for larger randomised controlled studies with longer follow-up.
-
ISTDP for treatment-resistant depression: large differences compared with a waiting list, but no support for the proposed chain of change
Effectiveness and mechanisms of Intensive Short-Term Dynamic Psychotherapy for treatment-resistant depression: a reanalysis of a randomized controlled trial
Two researchers at Stockholm University analysed openly available data from a randomised study with 86 participants. The study compared 20 sessions of ISTDP with a waiting list for treatment-resistant depression. The reanalysis examined both change in depression and whether it could be explained by the three process variables included in the mediation analysis: emotional suppression, negative affect and psychological distress. The researchers also tested whether process measures before treatment could predict depression after treatment, and whether the relationship ran in the opposite direction.
The difference between ISTDP and the waiting list was large at the end of treatment, Cohen’s d = 1.68, and larger at the three-month follow-up, d = 2.50 (95% CI 1.88 to 3.11). The proposed process measures also changed considerably. However, emotional suppression and negative affect did not mediate the change in depression. The distress measure initially appeared to do so, but the effect disappeared when the overlapping depression subscale was removed. The cross-lagged analyses could not demonstrate any of these relationships. The results therefore do not support the proposed stepwise chain of change in this particular reanalysis, but they do not rule out other mechanisms or changes occurring between measurement points.
LimitationsThe mechanism findings are exploratory and based on a secondary analysis of self-reports from three measurement points. This is not enough to capture rapid changes during or between sessions, and the measures may not correspond to the processes described by ISTDP theory. The study compared ISTDP with a waiting list, not an active treatment, which can produce larger effect sizes. Assessors were not blinded, the original study was registered retrospectively, and 11 of 86 participants had no post-treatment data. The reanalysis therefore supports large differences compared with a waiting list, but does not provide a definitive answer about how ISTDP works.
-
Meta-analysis: a large pooled difference favouring MCT for depressive symptoms, but considerable uncertainty
The effect of metacognitive therapy on depressive symptoms from a transdiagnostic perspective: a systematic review, meta-regression and meta-analysis
The review brought together 51 studies of full MCT programmes for adults in which depressive symptoms were measured. Studies of individual MCT techniques or MCT combined with other treatments were not included. The evidence covered both people with depression and groups in which depressive symptoms co-occurred with, for example, PTSD, GAD or long-term physical illness. Study designs varied and could be randomised, non-randomised or without a control group. The review followed Cochrane and PRISMA guidance and was preregistered in PROSPERO.
In studies with a comparison group, the pooled difference was large and favoured MCT, Hedges g = −1.28 (95% CI −1.71 to −0.86). Change within the MCT groups was also large, g = −2.29, but a pre-and-post measure without a control group cannot separate the contribution of treatment from other changes over time. The authors rated the certainty of evidence for the between-group result as moderate using GRADE. The findings suggest that MCT may reduce depressive symptoms compared with the control conditions included. However, the few active comparisons do not provide a firm answer about how MCT compares with other established treatments.
LimitationsResults varied widely between studies (I² = 93.73% in the between-group analysis). The pooled effect should therefore not be read as a uniform or expected result for an individual. Most studies were judged to have a high risk of bias, meaning methodological problems that may distort results, and none were rated low risk. Samples were often small and active comparisons few. The analyses also showed signs of small-study effects and possible publication bias. After a statistical adjustment, the between-group effect fell from g = −1.28 to g = −1.07. The difference was still large, but the adjustment does not remove the uncertainty.
-
Health anxiety in a student study: a small statistical interaction between bodily symptoms, anxiety sensitivity and metacognitive beliefs
Metacognitive beliefs about biased thinking condition the role of anxiety sensitivity in the somatic symptom-health anxiety pathway
The study used online self-reports from a sample of 564 university students. Four measures were analysed: bodily symptoms over the previous four weeks, health anxiety, anxiety sensitivity and metacognitive beliefs about health-related thinking. Anxiety sensitivity referred to fear of what bodily signals might mean or lead to, for example that palpitations signal serious illness. The metacognitive beliefs concerned the idea that negative thinking or worrying can be useful or protective, and that positive thinking can be risky.
Bodily symptoms, anxiety sensitivity and metacognitive beliefs each had a positive association with health anxiety. However, the simpler analysis did not clearly support the idea that anxiety sensitivity changed the strength of the association between bodily symptoms and health anxiety. The three-way analysis showed a small statistical interaction: at higher scores on the measure of metacognitive beliefs, greater anxiety sensitivity was associated with a stronger relationship between bodily symptoms and health anxiety. The pattern appeared above a calculated level that around 27% of participants exceeded. The three-way interaction explained an additional 0.62% of the variance in health anxiety.
LimitationsThe cross-sectional design shows concurrent associations, not what comes first or causes what. The sample came from one Canadian university, was young and non-clinical; 72.3% of participants were women. All data were collected through self-reports at the same time, and overlapping questions may have strengthened the associations. The three-way interaction was small. The study tested no treatment and cannot show that the metacognitive beliefs maintain health anxiety or that changing them would reduce it. Nor did it establish what caused the bodily symptoms or whether medical assessment was needed. The article reports 564 participants, but the reported degrees of freedom suggest that the regression models included 561 and 559 respectively; the difference is not explained. The pattern needs to be examined over time and in clinical and experimental studies.
-
Meta-analysis of MCT: greater symptom reduction than with a waiting list and CBT, but results varied
Efficacy of metacognitive interventions for psychiatric disorders: a systematic review and meta-analysis
The review brought together 49 randomised controlled trials with 3,239 adults. Twenty-one studies with 1,273 participants examined metacognitive therapy, MCT, developed by Adrian Wells. The other 28 studies with 1,966 participants examined metacognitive training, MCTraining, developed by Steffen Moritz. This is a different intervention, mainly studied in schizophrenia and psychosis. Results were analysed separately and the interventions were not compared with each other. The MCT part is summarised here. The review followed PRISMA, was preregistered in PROSPERO and covered published English-language studies up to May 2024.
On the studies’ primary outcome measures, the pooled difference was very large and favoured MCT over a waiting list: Hedges g = 1.84 (95% CI 1.18 to 2.50) in eight comparisons. Compared with the CBT treatments included, the difference was moderate immediately after treatment: g = 0.43 (95% CI 0.11 to 0.74) in 14 comparisons. At follow-up, an average of nine months later, the difference compared with CBT was smaller but still favoured MCT: g = 0.23 (95% CI 0.01 to 0.45). Waiting-list groups were not included at follow-up because they had by then received treatment.
LimitationsResults varied greatly between the MCT studies. The prediction intervals, which estimate what a new comparable study might show, included zero in both the waiting-list and CBT comparisons. The average advantage therefore cannot be expected in every new study or situation. Studies with moderate risk of bias, meaning methodological problems that may affect results, showed larger effects than studies with low risk. Smaller studies also showed larger effects.
The analysis of the MCT studies as a whole showed signs of possible publication bias. A statistical adjustment reduced the pooled effect from g = 0.90 to g = 0.38. However, tests for the separate comparisons with CBT and waiting lists did not show significant publication bias. Only 9 of 21 MCT studies had samples large enough from the outset to detect a moderate or large effect. Many also lacked information about study registration, treatment adherence, therapist competence, inter-rater agreement or whether blinding had worked. Overall, the results favour MCT, but do not say that MCT is always better than CBT or what an individual can expect.
-
MCT compared with ERP for OCD: non-inferiority after treatment, but not after six months
Metacognitive therapy versus exposure and response prevention for obsessive-compulsive disorder — A non-inferiority randomized controlled trial
The study compared metacognitive therapy (MCT) with exposure and response prevention (ERP), an established first-line treatment for obsessive-compulsive disorder. Seventy-four adults in outpatient care were randomly assigned to twelve weekly sessions of either treatment. OCD symptoms were assessed using the Y-BOCS before, during and after treatment and six months later. The assessors did not know which treatment participants had received.
A non-inferiority trial tests whether a treatment is no more than a specified margin worse than the comparison treatment. It does not mean that the treatments have been shown to be equivalent. The researchers used a margin of d = 0.38, which they described as approximately three Y-BOCS points. Immediately after treatment, the upper limit of the study’s two-sided 90% confidence interval was 0.37, just within the criterion. At the six-month follow-up it was 0.57, and the criterion was not met. This does not in itself show that ERP was superior, but at that point the study could not rule out a greater disadvantage for MCT. Y-BOCS scores fell in both groups, and dropout rates were low and similar, below 14%.
LimitationsThe study was small, and the post-treatment result was very close to the chosen margin. The criterion was not met at six months. Without a waiting list or another inactive comparison, the study can compare MCT with ERP, but cannot on its own establish how much of the change was due to each treatment.
A published corrigendum corrected an error in the abstract: the ERP group, not the MCT group, had a greater reduction in state anxiety after treatment and at follow-up. A methodological commentary published in 2025 questioned, among other things, the comparison margin, sample-size calculation and choice of confidence interval. The authors later published a response on the methodology and its rationale. The result should therefore not be read as conclusive evidence that MCT and ERP are equivalent or suit the same people.
-
Nine years later: the results continued to favour MCT for GAD
Metacognitive therapy versus cognitive-behavioral therapy in adults with generalized anxiety disorder: A 9-year follow-up study
The study followed up adults with generalised anxiety disorder, GAD, who had previously been randomly assigned to metacognitive therapy, MCT, or a manual-based CBT programme. Follow-up took place between eight and eleven years after treatment. Among those who completed the questionnaires, 12 of 21 in the MCT group, 57%, and 6 of 16 in the CBT group, 38%, were classified as recovered according to the study’s measure of worrying.
Across the four measurement points, the MCT group showed greater improvement on the measures of worrying and anxiety. Post-treatment improvements had largely been maintained in both groups. In the diagnostic interviews, 2 of 21 in the MCT group and 3 of 13 in the CBT group again met the criteria for GAD. Overall, the results continued to favour MCT compared with the CBT programme used in the study.
LimitationsThe study was small and 39 of 60 possible participants took part in the follow-up. The results are based on those who provided complete data, making the estimates less certain. The findings are promising but need to be confirmed in larger studies.
-
The Halifax follow-up: improvements after ISTDP were maintained at 18 months
Efficacy and cost-effectiveness of intensive short-term dynamic psychotherapy for treatment resistant depression: 18-month follow-up of the Halifax depression trial
The study followed adults whose depression had not improved sufficiently after at least one course of antidepressant medication. Participants were randomly assigned to up to 20 individual sessions of ISTDP or continued care from specialist psychiatric outpatient teams. The comparison group received, among other things, medication, clinical follow-up and, in many cases, talking therapy.
Analyses across six measurement points showed a greater reduction on the two depression measures after ISTDP than after the comparison treatment. Post-treatment improvements had largely stabilised and were maintained at 18 months. Among those assessed at that point, 8 of 20 in the ISTDP group, 40%, and 5 of 17 in the comparison group, 29%, were in remission according to the study’s observer-rated measure. In the ISTDP group, 7 of 30 had stopped antidepressants as part of their medical care. The health economic analysis also gave a cautiously positive signal, suggesting that ISTDP may be cost-effective compared with the specialist psychiatric care used in the study.
LimitationsThe study was small and conducted at a single centre. Depression measures were available for 37 of 60 participants at 18 months, although the longitudinal analysis used all available measurements. The health economic analysis was preliminary and affected by how missing data were handled. The findings are promising but need to be confirmed in larger studies.
-
The Halifax study: greater improvement with ISTDP than with treatment as usual in specialist psychiatric care
A randomised controlled trial of Intensive Short-Term Dynamic Psychotherapy for treatment resistant depression: the Halifax Depression Study
The study included adults whose depression had not improved sufficiently after at least one course of antidepressant medication. Participants were randomly assigned to up to 20 individual sessions of ISTDP or treatment as usual in specialist psychiatric care, provided by outpatient teams. The comparison group received, among other things, medication, clinical follow-up and, in most cases, talking therapy.
Over six months, scores on the two depression measures fell more in the ISTDP group than in the comparison group. Differences were moderate to large on both the observer-rated measure, Cohen’s d = 0.75, and the self-report measure, d = 0.85. Among those assessed after six months, 9 of 25 in the ISTDP group, 36%, and 1 of 27 in the comparison group, 3.7%, met the study’s criterion for full remission.
LimitationsThe study was small and conducted at a single centre. At six months, the primary depression measure was available for 52 of 60 participants, although the analysis used all available measurements. The comparison group received varied interventions and the ISTDP group had more frequent sessions and more sessions on average, which may have contributed to the difference. The results need to be tested by more research groups and in other care settings.
No results in this selection.
This research overview is not an individual treatment recommendation. What may be helpful needs to be assessed based on your situation, the problem formulation and how the work develops.