Kraus, D. R., Castonguay, L., Boswell, J. F., Nordberg, S. S., & Hayes, J. A. (2011). Therapist effectiveness: Implications for accountability and patient care. Psychotherapy Research, 21, 267-276.
Some patients benefit from psychotherapy, some do not, and a few get worse. Research has indicated that patient motivation, client-therapist match, and client characteristics might be associated with better or worse client outcomes. What about the contribution of the therapist? Do some therapists consistently have patients with better outcomes or with worse outcomes? Are consistently effective therapists effective for most patient problem areas or only some? Answers to these questions have important public health, funding, continuing education, and training implications. In a large study conducted in the U.S., Kraus and colleagues assessed 12 patient domains (sexual functioning, work functioning, violence, social functioning, anxiety, substance abuse, psychosis, quality of life, sleep, suicidality, depression, and mania) with a standardized reliable measure (the Treatment Outcome Package). The measure was used in a variety of public and private clinics and practices. Almost 700 therapists were sampled (including social workers 43%, mental health counsellors 35%, psychologists 10%, others 12%), with an average of 11 years experience. Ten cases were selected from each therapist caseload, so almost 7000 patients were included that received at least 16 sessions of therapy (16 sessions is an adequate dose for 50% of patients to improve – see my August, 2013 blog). The patients were, for the most part, representative of a typical caseload with regard to age, sex, and problem area as compared to previous national (U.S.) research. The authors used a reliable change index to classify patients as reliably improved, unchanged, or reliably worsened. The reliable change index is a way of assessing if change from session 1 to 16 on average exceeded the scale’s measurement error so that the change was considered reliable (i.e., not due to error). Reliable change for each therapist’s 10 patients was calculated so that a therapist could be classified as “effective” (i.e., on average their patients reliably improved), “ineffective” (i.e., on average their patients did not change), or “harmful” (i.e., on average their patients reliably worsened). The frequency of effective therapists ranged from a low of 29% in treating symptoms of sexual dysfunction to a high of 67% in treating symptoms of depression. Harmful therapists ranged from a low of 3% in treating depressive symptoms to a high of 16% in treating symptoms of substance abuse and violence. When looking at competency areas (i.e., areas of reliable effectiveness), the median number of areas of therapist competence was 5 out of 12 problem areas. Only 1 therapist of the approximately 700 therapists was competent in 11 of 12 domains, and none were competent in all 12 domains. Being effective in one domain was not correlated with effectiveness in another domain. So, one cannot infer that if a therapist was effective in treating depression he or she would also be effective in treating social dysfunction, for example.
There was tremendous variability in therapist skill and areas of competence in this very large sample of therapists. Between 3% and 16% of therapists were classified as reliably harmful to their patients, and between 29% and 67% were reliably effective depending on the problem area they were treating. Therapists who were effective in one domain could be harmful in another. Most therapists had some areas in which they were consistently effective, usually around 5. However, as indicated by previous research, without routine measurement, therapists may not be aware of clients for whom they are consistently helpful or harmful. Routine monitoring of outcomes could guide the matching of client problems to therapists, and could direct therapists to areas for continuing education, training, or personal therapy.
Author email: firstname.lastname@example.org