Blog

Swift, J. K., & Greenberg, R. P. (2014). A treatment by disorder meta-analysis of dropout from psychotherapy. Journal of Psychotherapy Integration, 24(3), 193-207.

In one of my first PPRNet Blogs I reported on a meta analysis by Swift and Greenberg (2012) in which they found that almost 1 in 5 patients in clinical trials dropped out of therapy. There were no differences between therapeutic orientations in the drop out rates. However, the authors did report that those with eating disorders (23.9%) and personality disorders (25.6%) dropped out at a higher rate than other disorders. Premature termination from therapy is an important problem in that those who drop out are less satisfied and have poorer outcomes than treatment completers. In this follow up to their meta analysis, Swift and Greenberg ask the interesting question of whether premature termination differs across therapy orientations for any of the specific disorders. They compared the drop out rates of different treatment approaches for each of 12 separate disorders.  The studies defined drop out in various ways, including: unilateral termination, not attending a set number of sessions, not achieving clinically significant change, etc. Treatment orientations, included: behavior therapy, cognitive–behavioral therapies, dialectical behavior therapy (DBT), psychodynamic psychotherapies, solution-focused therapy, interpersonal psychotherapy, humanistic/existential/supportive psychotherapies, and integrative approaches. Primary diagnoses included: depression, eating disorders, borderline personality disorder, other personality disorder, somatoform disorder, bereavement, obsessive compulsive disorder, generalized anxiety disorder, panic disorder, post-traumatic stress disorder (PTSD), psychotic disorders, and social phobia. The authors conducted 12 meta analyses, one for each disorder to compare the therapy approaches. Overall, they included 587 studies. There were no differences in drop out rates among therapy approaches for 9 of the 12 disorders. For depression, integrative therapy had significantly lower drop out rates than other approaches (10.9% vs 19.2%), and for PTSD integrative therapy also had the lowest drop out rate compared to other treatments (8.8% vs 21.0%). Also, for PTSD, exposure based interventions had the highest drop out rates (up to 28.5%). For eating disorders, DBT had the lowest drop out rates compared to other approaches (5.9% vs 24.2%), but this was largely explained by older patient samples and shorter duration of treatment in DBT.

Practice Implications

There were no differences between treatments in drop out rates for 9 of 12 disorders. Swift and Greenberg argued that for these disorders, other factors (e.g., therapeutic alliance, client expectations) rather than specific techniques were enough to keep clients in therapy. For depression and PTSD, integrative treatments resulted in the lowest drop out rates. This suggests that therapists might consider incorporating techniques from other orientations that increase the acceptability of therapy for their clients with depression and PTSD. Use of exposure based interventions for PTSD may require a significant amount of work to prepare clients in order to reduce higher drop out rates.

Lindheim, O., Bennett, C.B., Trentacosta, C.J., & McLear, C. (2014). Client preferences affect treatment satisfaction, completion, and clinical outcome: A meta-analysis. Clinical Psychology Review, 34, 506-517.

Giving clients a choice about treatments or to receive their preferred treatment might improve treatment outcomes. Preference usually means clients passively receiving the treatment they prefer. Choice involves clients actively making a decision about which treatment option to receive. Clients may also make informed or uninformed preferences and choices. Informed preferences and choices refer to providing clients with information or education about treatment options. Having a choice or getting one’s preference between two or more efficacious treatments might have several beneficial effects. For example, some research shows that treatment preferences positively affect therapeutic alliance, possibly because clients may enter treatment with a more positive outlook about what intervention they receive. Patients receiving a preferred treatment may also have better overall communication with their providers which may lead to better outcomes. In their meta-analysis, Lindheim and colleagues were interested in the effects of client preference or choice on treatment satisfaction, completion, and clinical outcomes. The meta-analysis included 34 different studies. Client preference or choice of treatment was modestly but significantly and consistently related to satisfaction, completion rates, and to client outcomes. Clients who were involved in shared decision making, who chose a treatment condition, or who received their preference had higher satisfaction, increased completion rates, and better clinical outcomes compared to clients who were not involved in the decision, who did not choose, or who did not receive their preference. Setting (inpatient vs outpatient) or diagnosis did not have an effect on these findings.

Practice Implications

The findings highlight the clinical benefits of assessing client preferences and providing treatment options when two or more efficacious options are available. Increasingly, two or more efficacious options are available for common mental disorders like depression and anxiety. Many times, patients prefer psychotherapy over medications, for example. However, whereas medication prescriptions for mental disorders like depression rose dramatically in the past decades, rates of psychotherapy use remained stable or slightly declined. For those disorders for which two or more treatment options have comparable efficacy, client preference should be the deciding factor.

Werbarta, A., Andersson, H., & Sandell, R. (2014). Dropout revisited: Patient- and therapist-initiated discontinuation of psychotherapy as a function of organizational instability. Psychotherapy Research, Online first publication: DOI: 10.1080/10503307.2014.883087.

Premature termination of psychotherapy in mental health care is a problem both in terms of patient outcomes and in terms of financial consequences for providers. Drop out rates for psychotherapy in general range from 20% to 75% with an average of 50%. In my April, 2013 blog I reported on a meta analysis by Swift and Greenberg (2012) in which they reported an overall drop out rate of 20% in randomized control trials; but the average drop out rate could be up to 38% in randomized trials depending on how premature termination was defined (failure to complete a treatment, attending less than half of sessions, stopping attending, or therapist judgment). Drop outs are commonly believed to represent therapeutic failures. Much of the research to predict psychotherapy non-completion has focused on patient variables like age, gender, symptom severity and others. This implicitly puts the responsibility for dropping out on the patient. Swift and Greenberg (2012) found that on average young, male, single patients with a personality disorder diagnosis were more likely to drop out. Therapist variables are less frequently studied, and the only therapist variable related to lower drop out was greater experience. Therapeutic orientations were not related to more or less dropping out. Very few studies have examined work conditions or organizational variables as predictors of premature terminations. Werbata and colleagues (2014) conducted a large study in 8 clinics in Sweden with 750 patients treated by 140 therapists. The clinics were three psychiatry outpatient units, three specialized psychotherapy units, one young adult psychotherapy unit, and one primary care setting that provided psychotherapy. Drop out was defined as unilateral termination in which either the patient or therapist discontinued the treatment. Of the patients who started therapy, 66% completed treatment and 34% terminated prematurely (19.7% of patients terminated the therapy, 14.3% were terminated by therapists). On average, clients were in their mid-30s, and most had a psychiatric diagnosis. The most common therapy was psychodynamic (59.1%) followed by integrative (19.0%), and cognitive behavioral (17.1%). The authors looked at patient variables (e.g., symptom severity), therapist variables (e.g., age, gender, etc.), and organizational stability. Ratings of organizational stability of the clinic were based on: the transparency of the clinic structure, the suitability of the organization to provide psychotherapy, the clarity of rules and decision-making policies regarding providing psychotherapy, and the clinic’s financial stability. Client variables such as: older age, greater level of psychopathology, and tendency to act out were moderately predictive of dropping out. Receiving treatment at a less stable clinic made it almost four times more likely for patients to initiate dropping out than to remain in therapy. Organizational instability was more important than patient factors in accounting for premature termination.

Practice Implications

Drop outs were almost four times higher in unstable clinics. Instability in organizations can create anxiety, cynicism, and disengagement in staff, which may have consequences for patient care. Financial and political problems within a clinic or institution, internal conflict related to treatment policy or disruptive administrative routines may affect the therapeutic relationship, which is generally more intimate and more important than in other health care contexts. Organizational instability can result in shortened or interrupted treatment, change in therapists, or therapists who are not fully engaged due to clinic stresses. For patients, these terminations may resemble earlier life losses or neglect that may have precipitated their need for therapy in the first place.

Imel, Z. E., Laska, K., Jakupcak, M., & Simpson, T. L. (2013). Meta-analysis of dropout in treatments for posttraumatic stress disorder. Journal of Consulting and Clinical Psychology, 81, 394–404.

There are now a number of psychotherapies that the Society of Clinical Psychology list as effective psychotherapies available for posttraumatic stress disorder (PTSD). Approaches include prolonged exposure (PE), and cognitive processing therapy (CPT) among others (click here for examples). Therapies for PTSD also vary in how much they focus on retelling the trauma. Some treatments like trauma-focused CBT place a higher level of focus on retelling the trauma event, whereas Present Centred Therapy (PCT), which was originally conceived as a control condition, largely avoids the trauma. Patients may begin a treatment and find some aspect of it distressing resulting in discontinuation. There is ongoing debate regarding the belief that exposure-based treatments, which require the patient to retell traumatic events in detail to his or her therapist, are especially unacceptable or poorly tolerated by patients. Drop out rate is a common metric used to assess tolerability of a treatment. In the April 2013 blog I reported on a meta analysis that found that the average drop out rate in randomized controlled trials of adult psychotherapy was 19.7%. However drop out rates for PTSD in the community can be as high as 56%. Imel and colleagues conducted a meta analysis of drop out rates in randomized controlled trials of treatments for PTSD. They also assessed if drop out rates differed by the amount the therapy focused on retelling the trauma. In the meta analysis, 42 studies were included representing 1,850 patients; 17 of the studies directly compared two or more treatments. The aggregated drop out rates across all studies was 18.28%, which is not different from the rate in randomized trials of adult psychotherapy in general, but is much lower than reported in regular clinical practice. Group treatment was associated with a 12% increase in drop outs compared to individual treatment. In general, an increase in trauma focus was not associated with greater drop out rates. However, when trauma focused treatments were directly compared to PCT (a trauma avoidant intervention) in the same study, trauma-specific treatments were associated with a twofold increase in the odds of dropping out.

Practice Implications

Many have been concerned that exposure-based therapies can lead to symptom exacerbation and result in dropout. The findings of Imel and colleagues’ meta analysis suggest that dropout rates are not significantly different among active treatments. However, PCT may be an exception to this general pattern of no differences among active treatments. Perhaps PCT should be considered a first line treatment for those who do not prefer a trauma focused treatment. In addition, providing treatment for PTSD in groups was associated with greater drop out rates possibly due to shame related to public disclosure of the trauma. The authors suggest mimicking research trial procedures in community practice in order to reduce drop out rates, such as: providing therapist training, support, and supervision; careful patient screening; regular assessment of patient progress; and ongoing contact with assistants that may promote session attendance.

Author email: zac.imel@utah.edu

Swift, J. K., Callahan, J. L., Ivanovic, M., & Kominiak, N. (2013, March 11). Further examination of the psychotherapy preference effect: A meta-regression analysis. Journal of Psychotherapy Integration. Advance online publication. doi: 10.1037/a0031423

Client preferences consist of preferences regarding the type of treatment offered (e.g., preference for psychotherapy or medication, preference for a behavioral approach to treatment or an insight oriented one), desires for a certain type of therapist or provider (e.g., preference for an older therapist, a female provider, or a therapist who has a nurturing personality style), and preferences about what roles and behaviors will take place in session (e.g., preference for the therapist to take a listening role or an advice giving role). In a previously published meta analysis Swift and colleagues (2011) reviewed data from 35 studies that compared preference-matched and non-matched clients. A small but significant preference outcome effect was found, indicating that preference-matched clients show greater improvements over the course of therapy, and that clients whose preferences were not matched were almost twice as likely to discontinue treatment prematurely. In this follow up meta regression study, Swift and colleagues assessed whether preference accommodation is more or less important for types of disorders, types of treatments, or different demographics like sex or age. (Meta regression involves accumulating data from across many studies to assess predictors [e.g., sex, age, diagnosis, treatment type, etc.] of the preference effect). For example, some research has indicated that men prefer therapists with more feminine traits and that men prefer pharmacological interventions. But does accommodating these preferences affect outcomes and drop out rates? Is matching preferences essential for younger clients? Is matching preferences more important for women or ethnic minorities? The authors analysed data from 33 studies representing 6,058 clients to address some of these questions. The only variable that predicted the influence that preferences have on rates of premature termination was the length of the intervention. That is, it may be more important to accommodate client preferences for briefer therapies. Perhaps, as clients continue in therapy for longer durations, other variables such as the therapeutic alliance play a bigger role in determining whether or not one drops out prematurely. It is also possible that as treatment continues, clients may experience a shift in preferences to more closely match the treatment conditions that they were given. Once this shift in preferences has occurred, preferences are no longer mismatched, and the risk of dropping out may be diminished.

Practice Implications

This study provides evidence that incorporating client preferences may be important for all types of clients. Generally, when client preferences are accommodated, clients show greater improvements while in treatment and are less likely to discontinue the intervention prematurely. As much as is practical, practitioners might collaboratively work with clients to identify what preferences they hold for treatment, and to discuss those preferences in the context of what is the most effective treatment that is available. This is particularly important for psychotherapies of shorter duration..

Author email: joshua.keith.swift@gmail.com

Swift, J.K. & Greenberg, R.P. (2012). Premature discontinuation in adult psychotherapy: A meta analysis. Journal of Consulting and Clinical Psychology, 80, 557-589.

Premature termination or drop out from psychotherapy has long been a significant problem for the practice of psychotherapy. Drop out can be variously defined as: not completing the course of treatment, unilateral termination on the patient’s part without therapist input, and not attending a specified number of sessions, among others. One of the largest meta analyses on the topic was done in 1993, and at that time the average drop out rate was 46.86%. This is a serious problem for a number of reasons. First, the average patient needs approximately 18 sessions to improve, and so early dropping out means that these patients do not benefit. Second, therapists can become demoralized at such drop out rates, and therapists who are not confident in their procedures are less likely to be effective. Third, the agency or practice loses important sources of funding or revenue. And fourth, society in general continues to manage the burden of a significant portion of its population not being at their best. The meta analysis by Swift and Greenberg (2012) is the largest of its kind, comprising 669 studies representing 83, 834 patients. The average drop out rate, largely defined as not completing treatment and unilateral termination without therapist input, was 19.79%. This appears to be a substantial drop from the previous 1993 number of 46.86%, but still represents one in five psychotherapy patients. Swift and Greenberg suggest that perhaps the more recent focus on evidence based treatments and short term treatments, and more systematic and consistent reporting of drop outs from studies may account for the lower numbers. No differences were found in drop out rates between treatment orientations (e.g., CBT vs others) and no differences in treatment format (e.g., individual vs group). Time-limited (20.7%) and manualized (18.3%) treatments tended to have lower drop out rates than non-time-limited (29%) and non-manualized (28.3%) treatments. Patients with eating disorders (29.3%) and personality disorders (25.6%) had the highest drop out rates. As did patients who were younger and less educated. Drop out rates in effectiveness studies (26%), that are more similar to everyday clinical practice, was higher than highly controlled randomized trials (17%). Trainee therapists (26.6%) tended to have higher drop out rates than experienced therapists (17.2%).

Practice Implications

At least one in five clients are likely to drop out of psychotherapy. Clinicians should particularly work on retention with younger clients and those with a personality or eating disorder diagnosis. Extra efforts to prevent dropout should also be emphasized for trainees and in university-based clinic settings. A number of strategies for reducing premature discontinuation in therapy have been identified, including discussing expectations regarding therapy roles and behaviors, providing education about adequate treatment duration, addressing motivation, repairing alliance ruptures, using therapist feedback, addressing client preferences, providing time-limited interventions, and increasing perspective convergence in the psychotherapy dyad. A number of these are described in greater detail in the following blog entry.

Author email: Joshua.Keith.Swift@gmail.com