 |
 |
 |
|
|
The nature of change is at the heart of research on psychotherapy. Researchers have demonstrated that psychotherapy is effective and have begun to address how change unfolds during the therapy process. Eventually, these efforts should clarify how characteristics of the treatment, the patient, the therapeutic relationship, and the therapist’s technique singly and jointly influence change during psychotherapy (1). In this study, we evaluated a contemporary conceptual model of the change process (2), using data from a comparative trial of 2 forms of short-term, time-limited dynamic therapy (3). The phase model of psychotherapy change, developed by Ken Howard and colleagues, is an extension of the dose-response model of treatment response (4). The phase model outlines a progressive, 3-stage sequence of change. In the first phase, the patient experiences a restoration of subjectively experienced well-being. The second phase focuses on the resolution of the patient’s symptoms. This phase occupies a greater number of sessions (generally, between the 5th and the 15th session of therapy. In the third and most protracted phase, changes in maladaptive forms of life functioning emerge more gradually. An important principle of the model is that change in one phase (well-being) constitutes a necessary precondition for change in a subsequent phase (current symptoms). Howard and others present evidence for the causal sequence outlined by the phase model, based on a naturalistic sample of 473 outpatients at the point of initial assessment (2). Patients provided ratings of well-being, current symptoms, and current life dysfunction at sessions 2, 4, and 17. In line with the open-ended therapy format, there was a high degree of attrition from the sample: over one-half of the patients had discontinued therapy by session 4. Two predictions were tested: 1) improvement in well-being is a precondition for improvement in symptoms, and 2) improvement in symptoms is a precondition for improvement in life dysfunction. Tests of the phase model were based on 2-by-2 cross-classification tables, representing the frequency of cases at each assessment that did or did not show clinically significant improvement on the dimensions of interest. (The method section provides details on the analysis approach.) The phase model is conceptually sound and intuitively appealing to clinicians. Other investigations have provided support for the model. Barkham and others reported a dose-effect relationship in a comparative trial of 8- and 16-session therapies for patients with major depression; that is, patients in the longer therapies showed greater improvement (5). They also noted changes in line with the phase model; specifically, well-being and acute symptoms showed the earliest and most rapid improvement. Hilsenroth, Ackerman, and Blagys described early changes in well-being and symptoms that, in turn, predicted changes in interpersonal dysfunction shown by the 9th session of short-term psychodynamic psychotherapy (6). Similar findings have emerged from studies involving cases selected to reflect specific types of emotional problems (7,8). Change in these studies was evaluated in terms of aggregate scores for the entire sample. In contrast, the Howard and others’ cross-classification method attends to the changes shown by individual patients (2). A direct test of the phase model using the cross-classification method is important for replication purposes and should produce findings similar to those of Howard and others. Thus, we attempted to provide a replication that duplicated their findings by using the same measure and methods, based on data from our comparative trial of 2 forms of brief psychotherapy (2). The comparative study used a randomized clinical trial design to investigate the efficacy of interpretive and supportive forms of short-term, time-limited individual (STI) therapy, with a diagnostically mixed and clinically representative sample of 144 outpatients (3). Attendance for both therapies was high; essentially the same patient cohort provided ratings of well-being, current symptoms, and current life dysfunction across the repeated assessments. At posttherapy, patients in both forms of therapy demonstrated statistically and clinically significant improvements of equivalent magnitude. Follow-up assessments at 6 and 12 months indicated equivalent maintenance of gains or further improvement (9). During therapy, we examined patient ratings of well-being, current symptoms, and current life dysfunction to test the sequence of change predicted by the phase model. MethodSetting and Procedure Details of the methods of the comparative trial are provided in our report of the posttherapy outcome findings (3). Patients were referred from the Psychiatric Treatment Clinic, at the University of Alberta Hospital Site, in Edmonton, Alberta. The clinic serves a catchment area that extends across Central and Northern Alberta. After obtaining informed consent, patients participated in interview and questionnaire assessments of predictor, demographic, diagnostic, and outcome variables. Patients were matched in pairs on personality variables, use of medication and, when possible, sex and age. Matched pairs were then assigned randomly to interpretive therapy or supportive therapy and to 1 of 8 therapists. Seventy-two matched patients completed each form of STI therapy (n = 144). Patients and Therapists The computer-administered Structured Clinical Interview for DSM-III-R determined the patient diagnoses (10,11). Axis I diagnoses were validated by an independent clinical diagnosis, assigned jointly by the intake assessor and a staff psychiatrist, both of whom saw the patient on the day of presentation. Nearly three-quarters (72.9%) of the patients received an Axis I diagnosis. The most frequent diagnoses were major depression (48.6%), dysthymia (26.4%), anxiety disorder (7.6%), and adjustment disorder (7.6%). A total of 60.4% of the patients received an Axis II diagnosis. The most frequent personality disorders were avoidant (29.2%), obsessive–compulsive (24.3%), borderline (22.2%), and paranoid (21.5%). Slightly less than one-half of the sample (46.5%) received both Axis I and Axis II diagnoses. Patients with primary problems related to psychosis, substance abuse, or antisocial behaviour were excluded. Pretherapy scores on the following 3 familiar measures of psychiatric disturbance exceeded outpatient norms, confirming the clinical representativeness of the sample: Beck Depression Inventory (mean 18.83) (12); Spielberger Trait Anxiety Scale (mean 52.64) (13); and Global Severity Index of the Symptom Checklist (mean 1.15) (14). The average age of the patients was 34 years (SD 9.6, range 18 to 62 years), and 61% were women. Of the patients, 42% were living with a partner, 21% were separated or divorced, and 37% had never been married. Two-thirds were educated beyond high school, and 71% were employed. Three-quarters of the patients reported receiving previous psychiatric treatment, but few (8%) had a history of psychiatric hospitalization. A total of 8 therapists (3 psychologists, 2 social workers, 2 occupational therapists, and 1 psychiatrist) each treated 18 patients, 9 in each form of therapy. There were 5 women, with an average age of 44 years (SD 6.1, range 37 to 52 years), and their average experience practising individual therapy was 11.8 years (SD 4.9, range 3 to 19 years).
|
||||||