Counseling and Mental Health Center

This study is a service evaluation of short-term therapeutic effectiveness and progress in therapy for participants seeking help at 42 different counseling centers across the nation. To examine the dose-effect relationship between the number of sessions (dosage) and the therapeutic outcome, 1,698 subjects were monitored through their therapy process as part of a large, nationwide counseling center research consortium. Participants were selected on the basis of having completed at least one, and no more than ten, sessions of therapy. Patterns of improvement varied depending on the number of sessions clients attended. There was a positive relationship between the outcome of therapy and the number of sessions the clients attended. This study, which relies upon naturalistic data, indicates that it is possible to make an argument for the partial effectiveness of brief psychotherapy. Dose-Effect and Brief Therapy 3 Dose–Effect Relationships in Brief Therapy Based on a Nationwide College Counseling Center Sample Several studies in the 1950s and 60s examined the relationship between the number of sessions clients attended and the amount of gain they achieved, finding that the greater number of sessions a client attended, the greater the amount of gain the client experienced (Seeman, 1954; Standal & Van der Veen, 1957; Johnson, 1965). More recent studies have verified these early findings concluding that length of treatment is positively and reliably associated with improvement and therapeutic benefit (Orlinsky, Grawe & Parks, 1994; Kadera, Lambert, & Andrews, 1996). In recent years, the dose-response metaphor borrowed from pharmacology research has gained prominence in psychotherapy outcome research (see Jones, Bigelow, & Preston 1999 for example). One of the major contributions to this type of research is the influential meta-analysis performed by Howard, Kopta, Krause and Orlinsky (1986) who proposed a dose-effect model linking the dosage of therapy sessions attended to the improvement clients experienced due to attending to each session. The log of the number of sessions became the dose and the normalized probability of improvement was defined as the effect using probits as the unit of analysis. They constructed their dosage model by probit analysis of 15 previous outcome studies, some dating as far back as 1950, and computed the predicted improvement session by session. Their results verified that the longer patients remain in therapy, the greater their gains. They found that 15% of patients improve between intake and the first session (presumably due to spontaneous remission and the ameliorative effects of having sought treatment), while 50% improve after eight sessions, and 75% improve after 26 sessions. Inspired by these results, other researchers began to analyze the dose effect for different diagnoses and levels of acute or chronic distress (Barkham, Rees, Stiles, & Shapiro, 1996; Kopta, Howard, Lowry & Beutler, 1994). Criticisms of this research surfaced after these articles were published. One critic, Phillips (1988), contends that earlier research failed to provide a sufficient data base for determining a valid dose-effect relationship due to vague definitions of reliable improvement, while others predate the use of repeated measure designs thus obscuring session-by-session patterns of improvement. Other critics questioned the previous research that grouped clients by diagnosis, given that patients with a specific diagnosis and similar symptoms do respond differently to treatment (Kadera, Lambert & Andrews, 1996; Kopta, Howard, Lowry, & Beutler, 1994). And other researchers point out that some previous research was well-controlled, with researchers influencing the number of sessions clients could attend, rather than observing what naturally occurred in various treatment settings (Kadera, Lambert, & Andrews, 1996). Observing what improvement naturally occurs in the session-by-session process of therapy has become much easier with the development of brief outcome measures such as the Outcome Questionnaire 45 (OQ45), a measure designed to be used not just at the beginning and end of therapy, but throughout the course of treatment (Lambert, Lunnen, Umphress, Hansen, & Burlingame, 1994). By utilizing measures such as the OQ45 to track the progress of groups of clients session-by-session naturally without setting up experimental controls, researchers can begin to examine questions not just about efficacy, but about the effectiveness of treatment in these natural settings as well (Howard, Moras, Brill, Martinovich, & Lutz, 1996). Dose-Effect and Brief Therapy 4 Researchers may also begin to further examine the effectiveness of counseling under a managedcare model, one utilized by some college counseling centers and not thoroughly researched from a dose-effect paradigm (Johnson & Shaha, 1996). Further examination of the dose-effect relationship in these centers is warranted, given that some critics contend that current managedcare accountability procedures are inappropriate because they emphasize usage rather than effectiveness curves, and that a more appropriate method may involve determining the sessionto-session effect in therapy and setting those curves to build models for treatment (Herron, Eisenstadt, Javier & Primavera, 1994). Developing effectiveness curves in natural settings like counseling centers is particularly important because researchers have not always found a positive relationship between the number of sessions and the amount of improvement in other settings. On the one hand, some researchers found that their studies supported prior dose-effect findings, although they needed qualification when examining acute, chronic and characterological components of depression, for example (Barkham, Rees, Stiles & Shapiro, 1996). On the other hand, some researchers, while examining the dose-effect relationship with children in the context of behavioral health organizations (BHO’s), found no evidence for a general dose-effect relationship (Salzer, Bickman & Lambert, 1999). Many of these studies are controlled, however; and such controls may limit their concurrent and external validity when compared to what is actually going on in treatment settings across the country. One study, for example (Barkham, Rees, Stiles & Shapiro, 1996), did focus on very time-limited therapy (8 or 16 sessions). Unfortunately, that study was so wellcontrolled that it did not focus on what actually happens in a real treatment setting, namely that clients, counselors, or both together often decide how many sessions clients will attend, within the constraints of the agency’s limits of treatment. Some counseling centers have opted for brief therapy models, which are surrounded by controversy. Some research indicates that brief therapies are ineffective and longer-term models are needed, while others argue that long-term therapy is unethical (Whitaker, 1994; Austad, 1996). Others appear to demonstrate both the viability of brief therapy models and some of the disadvantages within the managed-care setting (Cummings, Budman, & Thomas, 1998; Dziegielewski, 1997). This study provides the first analysis of dose-response relationships using a large nationwide sample of university counseling centers. This study examines only brief, time-limited therapy due to sample constraints (ten sessions or fewer). Based on the recommendations of Kadera, Lambert and Andrews (1996), this study tracked the progress of clients who came in for counseling on their own without posing artificial session controls on the therapists’ and clients’ decisions around how many sessions of therapy to attend, thereby overcoming the shortcomings of previous research. Dose-Effect and Brief Therapy 5