Blog

Cuijpers, P., Quero, S., Noma, H., Ciharova, M., Miguel, C., Karyotaki, E., Cipriani, A., Cristea, I.A., Furukawa, T.O. (2021). Psychotherapies for depression: A network meta-analysis covering efficacy, acceptability and long-term outcomes of all main treatment types. World Psychiatry, 20, 283-293.

Depressive disorders are common, and they have an important negative impact on quality of life and on mortality. For that reason, the treatment of depression is critical. The most commonly tested psychotherapy is CBT but others like interpersonal psychotherapy (IPT), psychodynamic therapy (PDT), and behavioral activation (BA) have also been tested. In this network meta analysis, Cuijpers and colleagues simultaneously test the effects of different psychotherapies for depression. Network meta-analysis, fundamentally, works by the transitivity assumption: if treatment A = treatment B, and treatment B = treatment C, then treatment A = treatment C even if Treatments A and C were never tested against each other in the same study. This procedure is not without controversy: what if the studies of treatment A vs B are all higher quality (thus resulting in lower effects) than studies of treatments B vs C? Is it fair to equate the studies by comparing treatments A and C when we know study quality impacts effect sizes? Nevertheless, network meta-analyses are used by some to aggregate many studies and to estimate relative outcomes across treatment types. Cuijpers included 331 studies (representing over 34,000 patients) in their network meta-analysis. CBT was tested in over 63% of trials, but other therapies (PDT, IPT, BA) were tested as well. All psychotherapies were more efficacious than care-as-usual and wait list controls with almost no significant differences between therapies for treating depression, except non-directive therapy was less efficacious than other therapies. (Non-directive therapy was often treated as a placebo control condition in studies, and so it may have been delivered in a way that limited its efficacy). CBT, IPT, PDT and BA all were more efficacious than care as usual at 12 months follow up.

Practice Implications

Overall, this network meta-analysis of psychotherapies for depression echoes the findings of many meta-analyses that preceded it. All psychotherapies that were examined, except for non-directive therapy, were equally efficacious for treating depression. When initiating therapy, it may be more important for therapists to be responsive to patient characteristics than to focus on which brand of therapy to deliver. For example, patients with internalizing coping styles may do better with insight oriented therapies, those with high levels of resistance/reactance may require a therapist that is less directive, and patients from marginalized race and ethnic communities may do better with a therapist who is multiculturally competent.

De Smet, M. M., Meganck, R., De Geest, R., Norman, U. A., Truijens, F., & Desmet, M. (2020). What “good outcome” means to patients: Understanding recovery and improvement in psychotherapy for major depression from a mixed-methods perspective. Journal of Counseling Psychology, 67(1), 25–39.

Many researchers consider the randomized controlled trial (RCT) as the best research design for testing medical and psychological treatments. However, critics of the design point to its limitations. For example, in order to collect homogenous samples of patients, researchers may exclude those with complex comorbidities. As a result, patient samples in RCTs may not represent patients one might see in real clinical practice. Also, researchers, and not patients, tend to define the meaning of what is a “good outcome” in these studies. It is possible that researchers and patients may not share the same definition of what it means to have a good outcome from psychotherapy. One key statistical and measurement method that researchers use to define outcomes is the reliable change index, which calculates the degree of change on a symptom scale from pre-treatment to post-treatment relative to the unreliability of the measurement. Using this method, researchers classify patients as “recovered” (reliably changed and passing a clinical cut-off score), “improved” (reliably changed but remaining in the clinical range), “not improved”, or “deteriorated”. However, this commonly used approach does not indicate whether the changes are actually meaningful to the patients. In this study, De Smet and colleagues interviewed patients from a randomized controlled trial of time-limited psychotherapy (16 sessions of CBT vs psychodynamic therapy) for depression who were classified as “recovered” or “improved” at post-treatment based on the reliable change index of a commonly used depression self-report scale. The authors asked how the patients experienced their depression symptom outcome, and what changes the patients valued since the start of therapy. In the original treatment trial of 100 patients, 28 were categorized as “recovered” and 19 patients were categorized as “improved”. During the post-therapy interview, the “recovered” and “improved” patients typically reported a certain degree of improvement in their symptoms. However, the patients categorized as “improved” reported that their gains were unstable from day to day, some reported having relapsed, and half did not feel that they improved at all. None of the “recovered” patients indicated that they felt “cured” of depression. Patients identified three domains of change that they experienced and valued. First, they felt empowered – that is, they had increased self-confidence, greater independence, and new coping skills. Second, they found a personal balance – that is, they had better relationships with loved ones, felt calmer, and had greater insight into their problems. Third, patients tended to identify ongoing struggles despite positive changes in the other domains – that is, certain key problems remained unresolved. “Improved” patients, and even some in the “recovered” group, indicated that their core difficulties had not been altered by the therapy.

Practice Implications

Although measurement of symptom change can give a clinician a general sense of how the patient is doing with regard to their symptoms and whether the patient is on track, such measurement may not capture the complexity of patients’ experiences of the therapy and any broader changes they may value. Patients in this trial, especially those classified as “improved”, had varied experiences. Aside from symptom reduction, clinicians should assess what their patients may value, such as: better relationships, greater self-understanding, more self-confidence, and feeling calmer. Most patients, including some who “recovered”, felt that they were engaged in an ongoing struggle, even after therapy. These findings suggest that addressing some of the core difficulties patients face may require longer term psychotherapy.

Cuijpers, P., Noma, H., Karyotaki, E.,  Vinkers, C.H., Cipriani, A., & Furukawa, T.A. (2020). A network meta‐analysis of the effects of psychotherapies, pharmacotherapies and their combination in the treatment of adult depression. World Psychiatry, 19, 92-107.

Mental disorders represent a significant health burden worldwide, with over 350 million people affected. Depression is the second leading cause of disease burden. There is ample evidence that psychotherapies and pharmacotherapies are effective in the treatment of depression. There is also evidence for the efficacy of different types of psychotherapy (CBT, IPT, PDT), and for different types of antidepressant medications. Some research suggests that combining psychotherapy and medications is better than either intervention alone, but the evidence is inconclusive. Existing meta analyses only compare two existing treatments directly to each other at a time: psychotherapy vs medications, psychotherapy vs combined treatments, medications vs combined treatments. In this meta-analysis, Cuijpers and colleagues use a method called “network meta-analysis” to study the relative impact of medications, psychotherapy, or their combination. Network meta-analysis is controversial because it relies on indirect comparisons to estimate effects. For example, let’s say one study compared medications (A) to psychotherapy (B), and another study compared medication (A) to combination treatment (C), then a network meta-analysis would estimate the effects of psychotherapy vs combination treatment by using the transitive principle (if A = B, and B = C, then A = C). This logic relies on everything being equivalent across studies. However, in treatment trials one cannot assume that the different studies comparing A, B, and C are equivalent in terms of quality and bias (in fact, we know they are not). In any case, Cuijpers and colleagues found that combined treatment was superior to ei­ther psychotherapy alone or pharmacotherapy alone in terms of standardized effect sizes (0.30, 95% CI: 0.14-0.45 and 0.33, 95% CI: 0.20-0.47). No significant difference was found between psychotherapy alone and pharmacotherapy alone (0.04, 95% CI: –0.09 to 0.16). Interestingly, acceptability (defined as lower patient drop-out rate and better patient adherence to the treatment) was significantly better for combined treatment compared with pharmacotherapy (RR=1.23, 95% CI:

1.05-1.45), as well as for psychotherapy compared with pharmacotherapy (RR=1.17, 95% CI: 1.02-1.32). In other words, pharmacotherapy alone was less acceptable to patients than another treatment approach that included psychotherapy.

Practice Implications

This network meta-analysis by a renowned researcher and in a prestigious journal adds to the controversy around the relative efficacy of psychotherapy vs medications vs their combination. What is clear is that patients find medication alone to be less acceptable as a treatment option, and previous research shows that patients are 4 times more likely to prefer psychotherapy over medications. Unfortunately, most people with depression receive medications without psychotherapy.

Courtois, C. A. & Brown, L. S. (2019). Guideline orthodoxy and resulting limitations of the American Psychological Association’s Clinical Practice Guideline for the Treatment of PTSD in Adults. Psychotherapy, 56(3), 329-339.

Recently the American Psychological Association (APA) published clinical practice guidelines for the treatment of post-traumatic stress disorder (PTSD). The reaction from the clinical community that treats those with PTSD, client groups, and from many academic and research quarters was swift and negative. APA received almost 900 comments in their public consultations from many who felt the document was overly prescriptive, overly symptom-focused, and narrow in its recommendations. In this interesting inside look at the process, the Chair of the PTSD Practice Guidelines Committee (Christine Courtois) and a senior member of the Committee (Laura Brown) wrote a scathing commentary of the process imposed on them by APA that constrained the Committee’s access to information which affected their decisions. The Committee was bound by APA’s use of the Institute of Medicine (IOM) rules for developing practice guidelines. In other words, a psychological organization (APA) used a biomedical model to define what is relevant research, how to define treatment, what is an appropriate outcome, and how to decide on recommendations. As a result, the APA Committee reached several conclusions/decisions that were biased or premature. First, they defined PTSD only by its symptom presentation and not for the complex disorder that it is. In other words, PTSD was viewed almost exclusively from within a framework that defined it as only a fear-based response to a stressor. Such an approach downplays any developmental or attachment-related factors in the genesis or maintenance of PTSD. Second, the Committee was instructed to ignore a vast array of research on therapist factors, relationship factors, and client factors in psychotherapy. This runs counter to many clinicians’ views that one cannot engage in technical interventions related to PTSD symptoms without the patient experiencing a heightened sense of security in their relationship with the therapist. This also meant that the Committee largely ignored cultural and diversity factors. Third, the treatment recommendations focused on time-limited exposure-based interventions – which is a natural outcome of the first two decisions (i.e., seeing PTSD as only fear-based, ignoring issues of development, and ignoring relational factors in the treatment context). The authors were also disappointed that the APA ignored its own policy on evidence-based practice that puts equal weight on research, clinician expertise, and client factors when making clinical decisions. In the end the authors clearly were not confident in the narrow focus of the Clinical Practice Guideline, and they were concerned that clinicians, researchers, policy makers, and third party funders could misuse the Guideline to limit research, theory, and funding.

Practice Implications

In this extraordinary piece, the Chair and a senior committee member of the PTSD Practice Guideline Committee were highly critical of the process and outcome of APA’s effort to develop clinical practice guidelines for PTSD. The authors did not diminish the importance of exposure-based interventions for PTSD, however they did argue that these interventions must be offered only after clinicians take a sufficient amount of time to create a clinical context characterized by clients experiencing heightened safety in the therapeutic relationship, and to into account client preferences and culture. Further, clinicians should be highly sensitive to attachment-related insecurities and developmental traumas that may lengthen the treatment and that may have a complicating impact on the therapeutic relationship.

Norcross, J. C., & Wampold, B. E. (2019). Relationships and responsiveness in the psychological treatment of trauma: The tragedy of the APA Clinical Practice Guideline. Psychotherapy, 56(3), 391-399.

The American Psychological Association’s (APA) Clinical Practice Guideline for Posttraumatic Stress Disorder in Adults published in 2017 was met with a great deal of concern and criticism by the community of scholars and practitioners working with patients with PTSD. A key concern was that the APA used a biomedical model and not a psychological or contextual model in guiding their understanding of PTSD, their approach to what constitutes evidence, and to decisions about recommended treatments. In particular, the biomedical approach focuses almost exclusively on treatment methods, and down-plays the context of treatment (i.e., the relationship, patient factors, and therapist responsiveness). In this critique, Norcross and Wampold highlight the flaws in the APA Clinical Practice Guideline for PTSD, and the authors focus specifically on those variables that are known to predict patient outcomes but that were ignored by the Guideline. Norcross and Wampold highlighted that there exists numerous meta analyses that demonstrate that all bona fide psychotherapies work about equally well for trauma, and that the particular treatment method has little impact on PTSD outcomes. Yet, the restrictive review process undertaken by APA all but ignored this well-established finding. Also ignored was the research on the importance of the therapeutic relationship in the treatment of trauma. One review outlined nineteen studies that found that the therapeutic alliance was associated with or predicted reduction in PTSD symptoms. This is consistent with the general psychotherapy research literature, in which the alliance is the most researched and most reliable factor related to patient outcomes. Also missing from the PTSD Guideline was reference to a large body of research on therapist responsiveness to patient characteristics. Patients are more likely to improve if their therapists can adapt to the patient’s coping style, culture, preferences, level of resistance, and stage of change. In one study of cognitive-processing therapy (CPT; a treatment recommended by the APA Guideline), there were substantial differences between therapists in their patient’s PTSD symptom outcomes. That is, some therapists reliably were more effective than others, even though all therapists were trained in and supervised in providing the same manualized evidence-based treatment. Among the identified skills of the most effective CPT therapists were: a flexible interpersonal style, and an ability to develop and maintain a good therapeutic alliance across patients.

Practice Implications

There is growing consensus that the APA Clinical Practice Guideline for PTSD are based on dubious methodology and are of limited use to therapists and their patients with PTSD. Psychotherapists should practice a bona-fide therapy for PTSD, but should do so by taking into account the treatment context. In other words, more effective therapists are good at developing, maintaining, and repairing the therapeutic alliance across a range of patients. Effective therapists can also respond and adapt to patient characteristics such as level of resistance, coping style, culture, and stage of change. And so, even when providing a treatment based on the APA Guideline, therapists should nurture trust in the therapeutic relationship and be adaptive to their patients’ characteristics.

Merz, J., Schwarzer, G., & Gerger, H. (2019). Comparative efficacy and acceptability of pharmacological, psychotherapeutic, and combination treatments in adults with posttraumatic stress disorder: A network meta-analysis. JAMA Psychiatry, 76, 904-91.

Posttraumatic stress disorder (PTSD) is a highly debilitating disorder characterized by re-experiencing trauma, avoidance of situations related to the trauma, negative mood and cognitions, and hyperarousal. The lifetime prevalence of PTSD in the population is about 8%, and PTSD is associated with a great deal of medical problems, and social and economic burden. Difference between a variety of psychological treatment approaches for PTSD are small and not statistically significant. Some treatment guidelines tend to recommend both psychotherapy and pharmacotherapy to treat PTSD, but other guidelines indicate only psychotherapy as the first-line treatment. Merz and colleagues conducted a meta-analysis to examine comparative outcomes and acceptability of psychotherapy and pharmacotherapy and their combination in adults with PTSD. The authors focused on randomized controlled trials because these designs tend to produce the most reliable evidence. The authors identified 12 published studies with a total of 922 participants. Six of the studies included data on long term outcomes. The meta-analytic procedures that the authors used in this study included network meta-analyses (which some have argued may produce unreliable results) and direct comparison meta-analysis (which is more reliable, but resulted in fewer studies being included here). I report in this blog only results that were consistent between the network and direct comparison analyses. Pharmacological and psychotherapeutic treatments and their combinations were not significantly different in their effectiveness immediately post-treatment. However, at long-term follow-up psychotherapy was significantly more beneficial than pharmacotherapy (SMD, −0.63; 95% CI, −1.18 to −0.09). Combined psychotherapy plus pharmacotherapy was not significantly more effective that pharmacotherapy alone (SMD, −1.02; 95% CI, −2.77 to 0.72), and combined treatment was not more effective that psychotherapy alone (SMD, 0.06; 95% CI, −0.31 to 0.42). There were also no statistically significant differences between psychotherapy, pharmacotherapy, or their combination in the acceptability of treatments to participants as defined by differing rates of dropping out from the studies.

Practice Implications

This meta-analysis of a small number of studies suggests that psychotherapy produces better long-term outcomes than pharmacotherapy for PTSD. There is also a suggestion that combining psychotherapy and pharmacotherapy does not improve outcomes compared to either treatment alone. This research area seems to be new and not well developed, but so far, the results seem to favor psychotherapy for longer term outcomes. These findings are similar to those from a larger meta-analysis for depression. In that study, the authors suggested that the long-term benefit of psychotherapy was due to participants learning coping and interpersonal skills that were not gained from receiving pharmacological intervention alone.