Are transdiagnostic mental health interventions the future of treatment?

Anxiety and depression are highly prevalent in both children (see previous Mental Elf blog) and adults (World Health Organization, 2022). They are also associated with significant economic burden and loss to quality of life, not only for patients, but also their families.

Although treatments tackling anxiety and depression exist, they typically treat these conditions separately, overlooking the fact that up to 60% of patients have co-morbid anxiety and depression (Kessler et al., 2015). Further, within primary care it is now more common to have a mixed diagnosis of anxiety and depression rather than purely one or the other (Newby et al., 2014).

It should be no surprise then, that ‘transdiagnostic’ interventions that can be used across diagnoses, have become increasingly popular, especially over the last 10 years. Many earlier meta-analyses exploring transdiagnostic treatments for anxiety and depression were not wide enough in scope, as they only reviewed specific types of transdiagnostic interventions (e.g., internet-based treatments) and/or predominately explored treatments for anxiety, and not depression.

To address these gaps, Pim Cuijpers and colleagues (2023) conducted a meta-analysis (using broad search terms) to capture trials exploring the effectiveness of transdiagnostic treatments for anxiety or depression in comparison to control conditions.

Due to high comorbidity between anxiety and depression, ‘transdiagnostic’ treatments are vital. However, meta-analyses up until now have been limited in scope.

Methods

Four databases were searched to identify randomised controlled trials (RCTs) of psychological interventions for people (18+ years old) with anxiety or depression, compared against a control condition. Trials that exclusively included participants with only an anxiety or depression diagnosis, were excluded. Interventions without human interaction (e.g., unguided self-help) were also excluded.

Study validity was assessed using the risk of bias assessment tool which includes four criteria (Higgins et al., 2011). Of the studies included, 48.9% met all criteria for low bias, with 44.4% meeting two to three criteria, and 6.7% only meeting one criterion.

Random-effects meta-analyses were conducted, with the effects of treatment on anxiety and depression compared across the short and longer term. The Number Needed to Treat (NNT) score and Prediction Interval (PI) were also calculated. The NNT represents the number of patients that need to be treated using this type of intervention, to prevent a bad outcome (such as, worse anxiety), from occurring. The PI can be used to indicate the impact of the intervention, along a continuum from harmful to clinically beneficial (InHout et al. 2016).

Results

45 RCTs (with 51 comparisons between treatment and control groups) were included in the meta-analysis. From these 45 trials, a total of 5,530 (mean age = 46.39; 65.6% female; intervention: n = 2,964 participants; control: n = 2,566) participants were included.

Demographics and characteristics

Participants had a large array of presentations, including Generalised Anxiety Disorder (GAD), Major Depressive Disorder (MDD), Obsessive Compulsive Disorder (OCD), or anxiety or depressive symptoms above a cut-off score without a specific diagnosis.

Participants were recruited from various sources, including clinics and in the community. Studies were predominately conducted within Europe, Australia, and North America, with all studies taking place after 2003.

Interventions described within studies were varied, and included psychodynamic therapy, guided-self-help, cognitive behaviour therapy (CBT), or a mixed format. Most interventions were between six to 12 sessions.

Impact of interventions on anxiety and depression

The overall pooled effect size between the control and treatment groups was g = 0.54 (95% CI [0.40 to 0.69]; NNT = 5.87), meaning the treatment groups improved more than the comparison groups (such as being on a waiting list). High heterogeneity was reported (I² = 78; 95% CI [71 to 83]), with a broad PI (-0.31 to 1.39). This means that the studies include in the analysis were very differently from one another, so pooling them together in a meta-analysis may not give a reliable result.

Even when correcting for potential biases in studies included (such as assessing risk of bias), results remained statistically significant.

A similar pattern of findings was found when looking at anxiety (g = 0.54, 95% CI [0.37 to 0.71]; PI -0.40 to 1.47; NNT = 5.92) and depression (g = 0.61, 95% CI [0.40 to 0.82]; PI -0.39 to 1.61; NNT = 5.11) separately, highlighting that the type of symptoms did not impact results. The impact of different anxiety and depression diagnoses on the results could not be explored due to methodological differences.

Sub-group analyses:

Pre-planned subgroup analyses were also conducted to compare effect sizes according to:

• Recruitment strategies,
• Different treatment formats,
• Target groups,
• Control condition type,
• Cognitive Behaviour Therapy (CBT) as compared to other therapies.

All subgroup analyses reported non-significant differences between subgroups, except for type of control condition. There was a significant difference (p = .003) between the wait list control group (g = 0.79; 95% CI [0.48 to 1.10]), usual care (g = 0.44; 95% CI [0.29 to 0.59]), and ‘other’ control group (g = 0.14; 95% CI [-0.21 to 0.49]).

Longer-term outcomes:

Effects of the psychological treatment were still significant at 6-months, but not at 12-months, indicating that transdiagnostic treatments were effective within the shorter term. However, the number of studies included that assessed longer-term outcomes was small, potentially reducing the ability to find a significant difference at 12-months.

The overall pooled effect of the meta-analysis was g = 0.54, suggesting that transdiagnostic interventions for depression and anxiety may be more effective than comparison groups.

Conclusions

Using a broader search strategy, this meta-analysis explored the impact of transdiagnostic interventions on anxiety or depression. Moderate effects for the intervention were found, which is consistent with previous meta-analyses using more specific search strategies. However, the heterogeneity across the included studies means that we have to interpret these results with caution.

At 6-months follow up, transdiagnostic interventions were more effective, when compared to a control condition. At 12-months, it could not be demonstrated that the intervention remained effective, but this could be due to the small number of studies exploring longer-term effectiveness.

This review suggests that transdiagnostic interventions for anxiety and depression were more effective than comparison groups at 6-month follow-up, but more research is needed to determine the longer-term benefits.

Strengths and limitations

This meta-analysis is a well-conducted and vital addition to the literature exploring the effectiveness of transdiagnostic treatments for anxiety and depression. Some strengths include:

1. Comprehensive search strategy: This paper used a broader search strategy than other meta-analyses exploring the use of transdiagnostic treatments. This allowed for a more comprehensive review and inclusion of studies that might have been missed by other papers.
2. Pre-registration: The protocol was published on the Open Science Framework (Cuijpers et al., 2022a, 2022b; https://osf.io/kyga2), meaning the results can be compared to pre-planned analyses.
3. Recording of NNT (number needed to treat): The reporting of the NNT is a strength as this number gives clinicians, at a glance, a useful and practical measure of how beneficial a treatment is.

However, it has a few limitations, some of which have been acknowledged by the authors themselves:

1. Quality of included studies: As acknowledged by the authors, the quality of included studies was far from perfect, with about half being classified as having a low risk of bias. While this doesn’t necessarily call into question the findings reported, readers should be aware of it.
2. Large variation in included studies (heterogeneity): The type of transdiagnostic intervention varied widely between included studies, as did the conditions in the control groups. Participants also had a wide range of anxiety and depression diagnoses. This made it difficult for the meta-analysis to report on the impact of specific interventions and whether they were more impactful on certain diagnoses.

Limitations exist surrounding the quality of studies included, and the small number of studies using the same type of intervention and sample.

Implications for practice

This meta-analysis is very relevant for those working within mental health services and delivering mental health interventions with clients or community members. It suggests that transdiagnostic mental health interventions should be considered as an effective way of treating anxiety and depression, at least in the short term, as they might have an edge over other treatments, especially when treating co-morbidity or mixed anxiety and depression symptoms.

The findings of this review are also relevant to policy makers and those distributing funding for research into novel approaches to supporting mental health. Policy makers should consider adding the use of transdiagnostic interventions as a recommended treatment strategy for anxiety and depression.

The meta-analysis also draws attention to the fact that more research needs to be done to explore the longer-term effectiveness of transdiagnostic interventions. Without this evidence, the use of these types of interventions as a viable treatment option might be overlooked. It is therefore the responsibility of future research to conduct RCTs to fill this gap in the literature as a way of building the evidence-base. Further, it is also important to extend the review of transdiagnostic interventions beyond their impact on anxiety and depression, to quantify the scope of their effectiveness. For example, can these treatments support the mental health of those experiencing a broad array of mental health challenges, such as psychosis, eating disorders, and perfectionism? As a way of supporting the creation of RCTs, as said above, funding bodies also need to be more open to giving funding to these projects.

From the perspective of the writer of this blog post, I feel that this meta-analysis is a much-needed addition to the literature, as it takes a broader perspective when looking at the impact of transdiagnostic interventions, without restricting the review based on diagnosis, or intervention type. It also highlights the need to conduct more research in this area.

From my own experiences of talking to mental health professionals and those who work with individuals experiencing mental health difficulties, the development of interventions that can be used across diagnoses, which can reach as many people as possible, are very much needed. Mental health challenges especially post-COVID, are becoming more and more common, while the availability of supports cannot meet demand. Transdiagnostic interventions may offer a solution to some of these issues.

There is an increased demand for mental health support, which services are not able to currently meet; transdiagnostic treatments could be an effective method of helping to fill this gap.

Statement of interests

None.

Links

Primary paper

Cuijpers, P., Miguel, C., Ciharova, M., Ebert, D., Harrer, M., & Karyotaki, E. (2023). Transdiagnostic treatment of depression and anxiety: a meta-analysis. Psychological Medicine, 1-12.

Other references

Cuijpers, P., Miguel, C., Ciharova, M., Ebert, D. D., Harrer, M., & Karyotaki, E. (2022a). Transdiagnostic treatment of depression and anxiety: Protocol for a meta-analysis. Open Science Framework.

Cuijpers, P., Miguel, C., Papola, D., Harrer, M., & Karyotaki, E. (2022b). From living systematic reviews to meta-analytical research domains. Evidence-Based Mental Health, 25, 145–147.

Hankey, L. (2023). Online support more helpful for youth anxiety than depression, according to recent review. The Mental Elf.

Higgins, J. P. T., Altman, D. G., Gøtzsche, P. C., Jüni, P., Moher, D., Oxman, A. D., … Sterne, J. A. C. (2011). The Cochrane collaboration’s tool for assessing risk of bias in randomised trials. British Medical Journal, 343, d5928.

IntHout, J., Ioannidis, J. P., Rovers, M. M., & Goeman, J. J. (2016). Plea for routinely presenting prediction intervals in meta-analysis. BMJ Open, 6(7), e010247.

Kessler, R. C., Sampson, N. A., Berglund, P., Gruber, M. J., Al-Hamzawi, A., Andrade, L. … Wilcox, M. A. (2015). Anxious and non-anxious major depressive disorder in the World Health Organization world mental health surveys. Epidemiology and Psychiatric Sciences, 24, 210–226.

Newby, J. M., Mewton, L., Williams, A. D., & Andrews, G. (2014). Effectiveness of transdiagnostic internet cognitive behavioural treatment for mixed anxiety and depression in primary care. Journal of Affective Disorders, 165, 45–52.

World Health Organization (2022). World mental health report; Transforming mental health for all. World Health Organization.

Photo credits