Blog

The Psychotherapy Practice Research Network (PPRNet) blog began in 2013 in response to psychotherapy clinicians, researchers, and educators who expressed interest in receiving regular information about current practice-oriented psychotherapy research. It offers a monthly summary of two or three published psychotherapy research articles. Each summary is authored by Dr. Tasca and highlights practice implications of selected articles. Past blogs are available in the archives. This content is only available in English.

January 2022

Are Psychological Therapies Efficacious?

Barkham, M. & Lambert, M.J. (2021). The efficacy and effectiveness of psychological therapies. In Barkham, W. Lutz, and L.G. Castonguay (Eds.) Bergin and Garfield’s Handbook of Psychotherapy and Behavior Change (7^th ed.). Wiley. Chapter 5.

Meta-analyses are not perfect, but they are the best way to summarize findings from a research field. In psychotherapy research, for example, meta-analyses aggregate effect sizes across a number of studies by producing a mean effect compared to a control condition. In this part of the chapter, Barkham and Lambert review the research on the efficacy of psychotherapies. The very first meta-analysis of psychotherapies was conducted by Smith and Glass in 1977 in which they aggregated effects of 475 studies that compared treated versus untreated groups. They reported a standard mean effect size of 0.85 in favor of psychotherapy, which is equivalent to a treated person being better off than 80% of untreated people. That sounds impressive, but it turned out to be an over-estimate of the effects of psychotherapy. Later studies took a more conservative approach by removing lower quality studies (small samples, poorly designed), which resulted in an average effect size of 0.67 when psychotherapy was compared to control conditions. Some meta-analyses went even further by excluding studies with wait-list control groups. It turns out that using a wait-list control (people who get no treatment) may inflate the effects of therapy because people waiting for treatment sometimes get worse, which by comparison makes the patients who get therapy look even better. These meta-analyses also controlled for publication bias (the likelihood that some negative or unflattering studies were never published). By taking all these possible sources of bias into account, the overall effect size of psychotherapy drops to about 0.31 (95%CI [0.24, 0.38]). Nevertheless, even when taking such a conservative approach to the research findings, one can conclude with some certainty that psychological therapies are effective to improve mental health conditions.

Practice Implications

The research over the years has demonstrated two important things about psychotherapy. First, psychotherapy is effective for many mental health problems, even if the effects are smaller than we once thought. The success rate for psychotherapy in treated persons is about 60% compared to 40% for untreated people (or to put it another way, therapists must treat 3 to 5 patients for one to recover). And these effects of psychotherapy are as large as one gets from many common medical interventions. Second, the effects of psychotherapy have not changed in the past 50 years. That is, new developments in psychological therapies and technologies have not moved the needle on patient outcomes since the 1970s.
Is Any One Psychological Therapy More Effective Than Another?

Barkham, M. & Lambert, M.J. (2021). The efficacy and effectiveness of psychological therapies. In In Barkham, W. Lutz, and L.G. Castonguay (Eds.) Bergin and Garfield’s Handbook of Psychotherapy and Behavior Change (7^th ed.). Wiley. Chapter 5.

In this part of the chapter, Barkham and Lambert review the debate about which psychotherapy is most effect for a particular disorder. This is a debate that has been ongoing since the middle of the last century. The current status of this debate centers around questions like how big are the differences between treatments, are these differences clinically meaningful, and if the specific interventions of a particular therapy are not the cause of change, then what are? It turns out that the early estimations of differences in therapies favoring CBT were over-blown. This was caused by combining effects across different studies that did not directly compare the treatments in question (the technical explanation is that comparing the effects of CBT vs a wait list control group in one study to the effects of psychodynamic therapy to a control group in another study introduces a whole host of confounds that make interpreting the differences in effects across studies unreliable). When subsequent researchers conducted meta-analyses of the effects of psychological therapies and only included studies in which bona-fide therapeutic approaches were directly compared to each other in the same study, then the differences between therapeutic approaches were small and clinically irrelevant (and even those small effects were be wiped out by controlling for the researchers’ allegiance [i.e., the effect caused by the researchers’ preference of one therapy over another]). Another way to consider the question is to assess the effects of the specific interventions of a therapy. That is, if CBT is effective because of its cognitive or behavioral components, then removing a component of CBT should reduce its effectiveness. These are often referred to dismantling studies. However, reviews of these types of studies indicate that removing components of treatment seem to have little or no impact on patient outcomes. That is, the effect of a therapy seems to have little to do with the specific intervention associated with the therapy (e.g., cognitive restructuring, exposure and response prevention). In fact, by far the largest predictors of patient outcomes are the factors common across most therapies: the therapeutic alliance, therapist empathy, goal consensus, therapist interpersonal skills, cultural adaptation of therapies, and others.

Practice Implications

This research domain is far from perfect, and the debate over which therapy is better unfortunately will likely continue for some time. However, clinically and from a public health perspective, this is not a useful debate and hopefully with time it will fade. Much more useful questions likely center around how we can improve therapist skills in those factors that lead to better patient outcomes. That is, regardless of theoretical orientation, can we train therapists to verbally express emotions and empathy, improve their capacity to develop and maintain the therapeutic alliance, and adapt their interventions to the patient’s culture, characteristics, and preferences?
How Much Therapy is Necessary?

Barkham, M. & Lambert, M.J. (2021). The efficacy and effectiveness of psychological therapies. In Barkham, W. Lutz, and L.G. Castonguay (Eds.) Bergin and Garfield’s Handbook of Psychotherapy and Behavior Change (7^th ed.). Wiley. Chapter 5.

For decades now, researchers have been trying to address the question of how many sessions of therapy is optimal for the average patient. In this part of the chapter, Barkham and Lambert review some of the research related to the question: how much therapy is necessary? Early research noted that it took about 13 sessions for 50% of patients to achieve clinically significant change, and 50 sessions was necessary for 75% of patients to achieve significant change. In other words, the number of sessions need to be doubled for another 25% of patients to improve. (Of note, half of patients did not improve after 13 sessions). Such research suggests a dose-response relationship, such that more therapy sessions result in more patients getting better – but the rate of return slows considerably with more sessions. The problem with this approach to estimating how many sessions are necessary is that it assumes that all patients improve at a consistent rate across sessions, but that may not be the case. Other researchers suggest that patient rate of change is individual, and that when a patient achieves a good enough level of change, they will terminate therapy. The research area is complicated by several factors. First, much of the research was conducted in university counselling centers with therapists in training, student patients, and bounded by the semester system. So, by definition, the therapy is time-limited and patients may be homogenous with regard to symptom severity. Second, some of the data in meta-analyses come from randomized controlled trials that by necessity represent brief treatments. Third, different modes of delivery in different countries appear to have an impact on how many sessions are delivered and what kind of therapy is offered. Fourth, patient severity at baseline may be related to longer treatments and slower rates of change.

Practice Implications

Although the dose-response research has been used in some settings to arbitrarily select treatment length offered to all patients, the results of this research may not be reliable. Predetermined fixed treatment lengths are likely not appropriate for patients. For example, greater patient severity at the start of treatment, and comorbidity and complexity of problems will likely result in such patients requiring a higher number of treatment sessions. A one-size-fits-all approach to determining how many sessions are offered to patients will likely negatively affect the most vulnerable of patients.

December 2021

How Much Psychotherapy is Necessary?

Lutz, W., de Jong, K., Rubel, J.A., & Delgadillo, J. (2021). Measuring, predicting, and tracking change in psychotherapy. In M. Barkham, W. Lutz, and L.G. Castonguay (Eds.) Bergin and Garfield’s Handbook of Psychotherapy and Behavior Change (7^th ed.). Wiley. Chapter 4.

The question of how many psychotherapy sessions are necessary to achieve good patient outcomes, or how frequently sessions should occur has been on the minds of practitioners and researchers for over a century. In this part of the chapter, Lutz and colleagues review some of the research related to how many sessions of psychotherapy is necessary to achieve positive outcomes for patients. A meta-analysis of 70 randomized controlled trials (RCT) of psychotherapy did not demonstrate any correlation between the number of sessions a patient receives and their outcomes. Other research indicates that receiving psychotherapy twice a week is more effective than receiving treatment once a week for depression. The findings of these two lines of research suggest that treatment length may not matter as much as treatment frequency. However, RCTs of psychotherapy tend to test only time limited therapies and they may not reflect exactly what happens in the real world with diverse patients who have complex problems. Perhaps the most relevant research for clinicians may be what is called the dose-response studies. These are studies that indicate how many sessions it takes for patients to get better regardless of treatment length. The dose-response research showed that 50% of patients starting treatment in the dysfunctional range required 21 sessions to achieve clinically significant change. That also means that half of patients did not change meaningfully with 21 sessions of therapy. More than 35 sessions were necessary for 70% of patients to achieve clinically meaningful change (and still, 30% of patients did not benefit). It is likely that some patients get better with a few sessions, but as severity or complexity of problems increase so does the number of required sessions. To add to the complexity, optimal duration of therapy varies according to practice settings. For example, for CBT in controlled studies the average patient needed about 17 sessions to get better, while 35 sessions of CBT was necessary in real world settings for the average patient to improve (again, that means that 50% did not yet improve).

Practice Implications

The findings from this line of research of the optimal number of sessions suggest that it is difficult to translate findings from controlled trials to real world practice. Most RCTs limit therapy to a brief number of sessions whether patients get better or not. Some patients do improve with a few sessions but over half of patients require more than 21 sessions to achieve clinically meaningful change, and about 30% of patients require more than 35 sessions. There is some evidence that more sessions per week leads to better outcomes as well.
Dropping Out From Psychotherapy

Lutz, W., de Jong, K., Rubel, J.A., & Delgadillo, J. (2021). Measuring, predicting, and tracking change in psychotherapy. In M. Barkham, W. Lutz, and L.G. Castonguay (Eds.) Bergin and Garfield’s Handbook of Psychotherapy and Behavior Change (7^th ed.). Wiley. Chapter 4.

In this part of the chapter, Lutz and colleagues review research methods related to patients dropping out of psychotherapy. Drop outs represent an important problem. For the clinician, a patient who drops out may represent loss of income due to missed appointments, extra work, administrative costs, and a lower sense of professional self-efficacy. Not all patients who terminate therapy early have a poorer outcome. But the research indicates that overall, patients who drop out of treatment do have poorer outcomes, higher hospitalization rates, lower work productivity, and higher social costs than patients who complete treatment. Hence, when a patient drops out of therapy it should be defined as a failure of the treatment that could lead to further demoralization of the patient. Defining a drop out is tricky in that some studies indicate that if a patient does not attend a minimum number of sessions, then they have dropped out. However, a more realistic definition might be that if a patient unilaterally decides to end therapy against a therapist’s advice, then the patient can be considered to have dropped out. Estimates of patient drop out from therapy vary widely depending on the treatment context and patient characteristics. For example, highly controlled studies report dropout rates of about 19.7%, but less controlled studies that might be closer to real world practice report average dropout rates of 26%. But the range of dropout rates across studies was very wide from 0% to 74.2%. Patient characteristics that led to higher dropout rates included higher initial impairment, younger age, lower level of education, a personality disorder diagnosis, and negative expectations about treatment. Therapists had a significant impact on dropping out as well. Therapist effects accounted for 12.6% of the variance in dropping out – that is a moderate but important effect. That is, some therapists have higher dropout rates than others, and this is likely independent of patient characteristics. This is like therapist effects on patient outcomes, in which it is estimated that about 10.1% of patient deterioration is predicted by the therapist’s effect.

Practice Implications

Patients dropping out from psychotherapy is an important problem that negatively affects the patient, the therapist, and that has broader social, health, and economic consequences as well. Aligning the patient’s and therapist goals for the therapy, coming to a collaborative agreement on how therapy will work, and developing an emotional and empathic bond with the patient may be ways of reducing the number of dropouts from therapy. These are all elements of the therapeutic alliance that must be negotiated very early in therapy to forestall a negative outcome such as the patient dropping out.
Routine Outcome Monitoring

Lutz, W., de Jong, K., Rubel, J.A., & Delgadillo, J. (2021). Measuring, predicting, and tracking change in psychotherapy. In M. Barkham, W. Lutz, and L.G. Castonguay (Eds.) Bergin and Garfield’s Handbook of Psychotherapy and Behavior Change (7^th ed.). Wiley. Chapter 4.

Routine outcome monitoring is also known as progress monitoring and feedback. This involves regularly assessing patients with a psychometrically reliable scale before a therapy session and providing the therapist with feedback on the patient’s progress. The feedback includes how the patient is doing relative to the average patient, and how the patient is doing in this session relative to their own scores in previous sessions. By doing this, therapists can get regular and reliable information about their patient’s progress and be alerted to when the patient is not on track to improve or is getting worse. Decisions about patient improvement on a session-to-session basis are complex because they involve knowing how the patient is doing regarding symptoms, quality of life, and relationship functioning relative to other patients and relative to their own functioning in the past. No wonder therapists often mis-judge when a patient is getting worse. Routine outcome monitoring involves decision-making tools to enhance a clinician’s decisions – like the way a physician uses a blood test or x-ray to enhance their clinical observations. In this part of the chapter, Lutz and colleagues review the research over the past 50 years on outcome monitoring and feedback. The effects of psychotherapy with feedback compared to psychotherapy without feedback ranges from small (g = .07) to medium (d = .40) in size. These effects seem small, but the authors remind us that feedback is a relatively simple clinical tool provided in addition to psychotherapy, and so these positive effects occur are over and above the general effectiveness of psychotherapy. Highest effect sizes are achieved for clients who are not on track (likely to get worse) such that feedback compared to no feedback in these not-on-track patients result in effect sizes ranging from 0.36 to 0.53, indicating a moderate to large effect. Further, when feedback was provided, patient dropout was reduced by 20% compared to when feedback was not provided.

Practice Implications

Although routine outcome monitoring is relatively easy to use, there are barriers to their implementation. Organizational cultures are difficult to change, and resources must be assigned to implement these strategies. Clinicians must have the technology, some training, and funds to purchase the psychometric scales which may be a challenge for some. And attitudinal barriers are a problem if managers or clinicians do not value outcome measurement. Nevertheless, patient reported outcomes with psychometrically valid scales should be central to ensure good patient-centered care. Psychotherapists can benefit from quality information to help their clinical decision making, especially when it comes to identifying patients who might not be benefitting. Such feedback about patients who are at risk of getting worse may help clinicians to adjusting treatment and their interpersonal stances to these patients.

This month...

Type of Research

Topics

January 2022

Practice Implications

Practice Implications

Practice Implications

December 2021

Practice Implications

Practice Implications

Practice Implications

Psychotherapy Practice Research Network (PPRNet)