Don't give up! Giving psychotherapy to those for whom it hasn't worked before
This week I summarise a review from Gloster and colleagues (2020) of randomized controlled trials of psychological therapies for clients who had previously failed to respond to a psychological therapy. According to Gloster et al., the rate of non-response to psychological treatments for anxiety and depressive disorders is consistently between 30-40%. Treatment non-responders tend to have poorer quality of life with increased suicide attempts and completions.
Their review included 18 studies with 1734 participants who had any anxiety or mood disorder, were of any age, and who still met diagnostic criteria after receiving psychotherapy or pharmacotherapy, provided that the psychotherapy for treatment non-response was different to the psychotherapy that produced the non-response in the first place. Ten different “brands” of psychotherapy were used to treat non-response among the 18 studies, including CBT, Cognitive Behavioural Analysis System of Psychotherapy (CBASP), behaviour therapy, MBCT, ACT, DBT, short-term dynamic psychotherapy, long-term psychoanalytic therapy, interpersonal psychotherapy (ITP) and care management. In 7 studies the active psychotherapy was combined with medication. Fourteen of the studies used treatment as usual (TAU) as the comparison group, 3 used wait-list control and one used pharmacotherapy. Some version of the Beck Depression Inventory (BDI) was the most typical outcome measure used (8/14 studies). Half the studies included some measure of quality of life or functioning (e.g., a version of the Short-Form [SF] or WHOQOL scales).
The initial estimated effect size for psychotherapy of treatment non-response was SMD = 0.82 [95%CI 0.56, 1.08]. In clinical terms, this means those who received the active psychotherapy for non-response would, for example, on average, experience a further reduction in BDI scores of 6-7 points compared to the control conditions. However, a test for publication bias revealed a significant bias toward positive findings. Adjusting for publication bias, the estimated effect was SMD = 0.45 [0.16, 0.75]; in clinical terms, a 3-5 point greater reduction in BDI scores for those receiving the active psychotherapy.
Effect sizes did not differ depending on whether participants’ primary diagnosis was an anxiety or mood disorder. Effect sizes did not differ depending on the type of treatment participants had previously failed to respond to (i.e., medication, psychotherapy or their combination). Effect sizes did not differ according to the type of “new” psychotherapy received or on the type of control group used. Effect sizes were related to study quality in the usual way: the better the study quality, the lower the effect size. Interestingly, the longer the period of the previous unsuccessful treatment, the larger the improvement with the new psychotherapy.
Psychotherapy for non-response also produced greater improvements in quality of life, SMD = 0.41 [0.18, 0.65] . Quality of life improvements were smaller when the primary diagnosis was a mood disorder and when the control group was TAU rather than wait list. Importantly, study quality did not affect effect sizes for quality of life improvement.
The conclusions of the review are mainly limited by the state of the literature. There are fairly few studies of psychotherapy for treatment non-response. The effect size estimates were highly heterogeneous and the authors were unable to explain much of this variability within features they were able to describe about the studies. At this stage, the literature on treating non-response consists of a diverse range of strategies, applied to a diverse range of participants with a diverse range of problems, over a diverse range of treatment lengths, varying in whether pharmacotherapy is presented in combination…perhaps it’s not surprising effect sizes are heterogeneous.
On the upside, at this stage there is no need to be too prescriptive about how to work with people who have failed to respond to previous pharmacotherapy or psychotherapy. The main rule would seem to be to not provide what they tried before! Jeromy Frank famously attributed the primary benefit of psychotherapy to remoralization of clients whom had become demoralised. The main ingredients to achieve this are a plausible rationale (what Frank called a myth), credibility and a plausible treatment plan (what Frank called a ritual). There appears to be ample scope within structured traditional CBT, third-wave CBT and psychodynamic approaches to remoralise those who have been demoralised by previous therapy experiences. So, clinician: remoralize thyself, then remoralise thy clients!
Take home: Clinical Implications
Formal psychotherapy is more effective than no treatment, or unstructured “treatment-as-usual”, in helping those who have not previously achieved sufficient benefit from psychotherapy or medication, to no longer meet diagnostic criteria for an anxiety or mood disorder.
Providers have a range of choices for systems of therapy to offer these clients: Traditional CBT, “third wave” CBT, and psychodynamic approaches have all been able to achieve improvements where a previous course of psychotherapy was unsuccessful. The main proviso is that it not be the system of therapy that the client failed to respond to.
These conclusions are limited to cases of primary anxiety or depressive disorder, and structured forms of psychotherapy: those that are manualised, organised, have a plausible rationale, and are seen as credible and acceptable to clients.
Additional Recommendation
Routine outcome measurement and formal diagnostic interviewing should be part of the client review process to ensure that cases of non-response are identified. These are usually evident within 3-5 treatment sessions, but would certainly be evident within the 10 sessions of the standard care provided in the Australian Better Access program. By non-response, I mean that there is no evidence of reliable improvement in symptom measures or goal attainment. At this point, a second opinion on primary diagnosis is warranted, and a new treatment program recommended.
Go to https://www.sciencedirect.com/science/article/pii/S0272735819303228 for the original article.