Reich, C.M. & Berman, J.S. (2018). Are psychotherapies with more dropouts less effective? Psychotherapy Research. Online first publication.
The client dropout rate in clinical studies is about 20%, and almost double that in regular clinical practice. A dropout is a unilateral decision on the part of the patient to discontinue treatment after the first session. This is often viewed as a treatment failure, but it is possible that some patients feel better enough to not continue with therapy, and others may have practical constraints like transportation or childcare difficulties. In three meta analyses, Reich and Berman ask: (1) do those who drop out experience more distress than completers to begin with?, (2) do completers have better outcomes than drop outs?, and (3) do less effective treatments also have more drop outs? In general, the studies included different types of psychotherapy but most were CBT (~75%), most therapists had a masters or doctoral degree (~33%) but many studies also included student therapists (~25%), client problems included depression (~17%), anxiety (25%), or other disorders, and most studies were randomized controlled trials (~61%). To answer the first question the authors conducted a meta analysis of 76 studies. Clients who dropped out of therapy were in more distress prior to beginning treatment than individuals who completed the treatment (d= − 0.14, 95% CI [− 0.08, − 0.20], p < .001). The effect was small but significant. Younger and male clients tended to be in more distress at pre-treatment. To answer the second question, the authors conducted a meta analysis of 43 studies. Clients who dropped out of therapy were significantly more distressed following therapy than individuals who completed treatment (d = .0.56, 95% CI [.0.37, 0.70] p < .001). This was a moderately large and significant effect. To answer the third question, the authors completed a meta regression of data in 34 studies. Overall, treatments with more drop outs also had completers with worse outcomes at post treatment, β = -.37, SE = 0.17, p < .05. Also, when treatments were shorter in length, greater overall dropout was associated with even worse outcomes for treatment completers, β = − 1.28, SE = 0.35, p < .001.
These meta analyses support the notion that on average those who drop out do so because they do not find the treatment to be helpful. Patients who drop out tend to be more distressed to begin with, and are more likely to be young and male. An intriguing finding was that those treatments with more drop outs also tended to be less effective for those who completed the therapy. In other words, effective treatments also tended to maintain more patients. Previously, writers suggested clinically useful methods to reduce premature termination from psychotherapy. These include: providing patients with information about duration of therapy and how change occurs, educating patients about therapist and patient roles, taking into account patient preferences when deciding on treatment methods and therapist stances, strengthening early hope, fostering the therapeutic alliance, and assessing and discussing treatment progress on an ongoing basis.