|Home | About | Journals | Submit | Contact Us | Français|
As a substudy of a manual-based outcome study of the Beth Israel Brief Psychotherapy Program, the authors studied the efficacy of supportive psychotherapy in personality change, with particular attention to changes that outlast the period of treatment. They examined results from the Inventory of Interpersonal Problems (IIP) at intake, 40th-session termination, and 6-month follow-up in the first 20 subjects randomized to the supportive group. Eight subjects (40%) dropped out, but their initial IIP scores did not differ from those of follow-up completers. Six of 10 subjects with complete 6-month follow-up data showed significant improvement in interpersonal problems (4 cases P < 0.001; 2 cases P < 0.05). In a case method design, using the IIP mapped to an interpersonal circumplex model, the authors graphically demonstrate lasting positive changes in interpersonal functioning in subjects treated with supportive psychotherapy. (The Journal of Psychotherapy Practice and Research 1999; 8:55–63)
Supportive psychotherapy is widely practiced but seldom studied. The decision-making process about whom to treat or not treat with supportive therapy has been based on clinical traditions embodying assumptions that have not been experimentally validated. Because supportive psychotherapy has been traditionally seen as the treatment for those not suitable for expressive therapy, supportive treatment has been generally confined to patients with whom expressive techniques are expected to fail or who are difficult to treat. In an assessment of patient difficulty in relation to therapist performance of Interpersonal Psychotherapy, Foley et al.1 found that greater patient difficulty predicted lower therapist performance. There has been selection bias in that more difficult patients are routinely shunted to supportive psychotherapy and the better candidates for therapy are selected for expressive treatments, a process that reinforces the notion that supportive psychotherapy is not indicated for what Murphy2 called YAVIE (young, attractive, verbal, intelligent, and educated) patients. Thus, by the circular process of assigning very sick and very difficult patients, the image of supportive therapy as a low-potency intervention is perpetuated.
These self-fulfilling clinical traditions, reinforced by models of change that have been based on classical or even postclassical dynamic views (i.e., that supportive therapy builds needed structure), have limited the application of supportive treatment primarily to patients who have primitive impulses and defenses, poor object relations, poor ego strength, or who are difficult to manage.3–7 Thus, there has been inadequate attention to the use of supportive psychotherapy in higher functioning patients, and as a result there are almost no outcome data from clinical trials. We have devised a model8,9 (and manual10) of supportive psychotherapy that is not based on a formal dynamic theory, and that, within a technology model of psychotherapy research, lends itself to internal and external validation.11–14
We have defined supportive psychotherapy as a dyadic treatment characterized by use of direct measures to ameliorate symptoms and to maintain, restore, or improve self-esteem, adaptive skills, and psychological (ego) function. To the extent necessary to accomplish these objectives, treatment may examine relationships (real or transferential) and both current and past patterns of emotional or behavioral response. The techniques are not limited to use with severely impaired patients, as might be assumed in a treatment derived from the supportive end of the supportive–expressive continuum.15 Conversely, it appears that several types of patients, such as those with schizophrenia16 or substance use disorders,17 who have been traditionally prescribed supportive psychotherapy, may make use of more expressive treatments, within certain constraints, than previously predicted from dynamic theory.
The inclusion of the terms restore and improve in the definition implies that some type of adaptive behavioral change over the functioning at initial presentation (such as clinically significant improvement) is a potential outcome of supportive treatment. With traditional supportive psychotherapy of the low-functioning patient in mind, one tends to conceive of a long-term maintenance-oriented treatment where stabilization is dependent on the continued supportive relationship with the therapist. However, we have treated higher functioning patients in supportive psychotherapy who improved clinically. These higher functioning patients now have been shown to have outcomes comparable to those of patients treated with therapies associated with intrapsychic and interpersonal change.18
On the basis of clinical experience, we believe that supportive psychotherapy often leads to improvement in adaptive and interpersonal function, even to what may be called “structural” change. This is change described as alteration in specific intrapsychic configurations, such as modification in patterning of defenses, thought and affect organization, anxiety tolerance, and ego strength.19 In order for improvement to be structural, or at least internalized in the patient, it should outlast the immediate relationship with the therapist.20
Questions thus arise for research validation. First, does change occur as a result of supportive psychotherapy, and if so, at what level of integration (behavioral skills, defensive structure, self-organization) does it occur? Second, if change in interpersonal functioning occurs because of supportive psychotherapy, is it mirrored in the constructs of intrapsychic function (that is, does behavior imply character)?
This article addresses the first of these questions. We hypothesized that supportive psychotherapy would lead to interpersonal change as measured by reductions in the average Inventory of Interpersonal Problems scores between intake and termination and between intake and 6-month follow-up.
Subjects were 20 patients consecutively assigned to the Brief Supportive Psychotherapy cell of the Brief Psychotherapy Research Program of Beth Israel Medical Center, NY, as part of a randomization procedure. After giving informed consent, all subjects were screened prior to randomization into the Supportive Psychotherapy cell, using the SCID semistructured interview,21 Inventory of Interpersonal Problems (IIP),22 Wisconsin Personality Disorders Inventory (WISPI),23 Global Assessment of Functioning (GAF),24 and Target Complaints,25,26 by research assistants trained to commonly accepted levels of interrater reliability. All subjects met criteria for DSM-III-R cluster C personality disorders. Subjects were excluded if they were deemed to be poorly related; acutely suicidal; suffering from psychotic disorders, borderline or antisocial personality disorders, or current substance use disorders; or acutely in need of medication. All treatments in the overall study are manualized and videotaped. Ratings of videotaped sessions by trained raters blind to the treatment type are used to ensure therapist adherence and fidelity to the specific treatment in each cell. All therapists were senior clinicians at Beth Israel Medical Center, each with greater than 10 years' psychotherapy experience.
Brief Supportive Psychotherapy is a conversation-based dyadic treatment whose focus is the maintenance or increase of patients' self-esteem, adaptive skills, or psychological function by direct techniques. Because the process is modeled on conversation, the therapist tends to respond more frequently in supportive psychotherapy than in typical expressive therapies. In addition to directly supportive techniques (data-based praise, advice, appropriate reassurance, and anticipatory guidance), clarification, confrontation, and, more rarely, interpretation are useful, but with no requirement that the unconscious be made conscious or that full linkage of impulses or affects be made with early important figures. Defenses are not confronted unless they are grossly maladaptive (e.g., primitive projection or splitting) or threaten the frame of treatment.
In contrast to the neutral, anxiety-provoking stance of the therapist in expressive treatments aimed at evoking transference phenomena, supportive psychotherapy emphasizes the real relationship, based on the interest and helpful responsiveness of the therapist and reflected in the therapeutic alliance. The therapeutic alliance is actively supported through the use of accurate empathic responses and validation of feeling states. The development of a transference neurosis is avoided.
The concept of change in supportive psychotherapy is that it occurs through a collaborative effort of patient and therapist to understand interpersonal and other patterns in order to discover what can be changed—not to discover reasons for the existence of the behavior or feelings that should be changed. Self-understanding is not central to the treatment, and it is pursued only to the extent that it supports the accomplishment of patient goals and therapist objectives.
Interpersonal behavior can be measured based on the circumplex model,27–29 which was conceptualized by Leary30 and further developed by Benjamin,31 Wiggins,32 and Kiesler.33 The circumplex model arranges interpersonal behavior and attitudes on a surface defined by two axes: a horizontal axis reflecting affiliation, with poles of friendliness and hostility, and a vertical axis reflecting autonomy and control, with poles of dominance and submission. The regions between the axes may be subdivided in various ways, but the most psychometrically desirable way usually involves eight regions.34,35 In the case of this analysis via the IIP, these regions are labeled Domineering, Vindictive, Overly Cold, Avoidant, Nonassertive, Exploitable, Overly Nurturant, and Intrusive.36
The circumplex model has been used both as a measure of personality and as a measure of interpersonal process. The IIP, an instrument scaled according to the circumplex model, has been used to measure maladjustment in terms of the rigidity and restriction of interpersonal behavior. The more constricted the behavior, the easier it is to define it along the two axes of the circumplex model. As such, the model lends itself to identifying the rigid and restricted nature of the interpersonal behavior characteristic of personality disorders.
By most accounts, individuals are considered to have personality disorders when their construal style and interpersonal behavior are particularly rigid and restricted. Accordingly, their characteristic style of construing and interacting fosters vicious cycles that perpetuate and intensify their difficulties.37
Responses on the 64-item IIP at intake were compared within subjects to responses at 40th-session termination and 6-month follow up by one-tailed, paired-samples t-test.
Statistically significant change was also established on a case-by case basis by the formula for reliable change (RC) as calculated by Jacobson and Truax.38 The standard error of the difference score from an overall IIP mean from 118 concurrent Brief Psychotherapy Research Program patients was 0.11. The significance criteria for RC coefficients are 1.96, P < 0.05; 2.53, P < 0.01; and 3.40, P < 0.001. Pooled average intake IIP scores of dropouts were compared with pooled average scores of subjects who completed to termination or follow-up. So that we could better visualize modification of interpersonal problems in subjects with significant change, the scores at intake, termination, and follow-up points on the eight circumplex subscales from the 64-item version of the IIP were calculated, plotted graphically on a circumplex octant model,36 and compared.
Eight subjects (40%) dropped out. Average initial IIP scores of subjects who dropped out (mean ±SD = 1.06 ± 0.43) were not significantly lower (t = 1.79, df = 18, P = –0.089) than initial scores of subjects who completed treatment and/or follow-up (1.56 ± 0.70). Paired-samples t-test for 11 subjects demonstrated a significant reduction in average IIP scores (t = 3.58, df = 10, P < 0.005) between intake (1.59 ± 0.73) and termination (1.28 ± 0.84). This is a moderate to large effect (r2 = 0.562). Paired-samples t-test for 10 subjects demonstrated a significant reduction in average IIP scores (t = 1.98, df = 9, P < 0.039) between intake (1.61 ± 0.80) and follow-up (1.28 ± 0.89), a moderate effect (r2 = 0.304). Six of 10 subjects (60%) who had completed 6-month follow-up data showed statistically significant improvement in interpersonal problems by reliable change methods38 (Table 1).
It is useful to relate the numerical data on the IIP subscales with a graphical representation of change mapped to the circumplex model. Four cases are presented briefly in order to support the graphical changes with historical and clinical data.
Case 1 (Figure 1) is a 38-year-old divorced man whose target complaints consisted of being greatly bothered by his silence and withdrawal in dealing with people, lack of confidence about decisions, and low self-esteem. On the SCID-I/SCID-II, he had threshold criteria consistent with a diagnosis of primary dysthymic disorder and avoidant personality disorder, and subthreshold for self-defeating personality disorder. As a child, he had been physically abused by his alcoholic father. He also had little interest in sex. He often felt inferior to those around him, and he had always been sensitive to rejection and had difficulty tolerating criticism. As a result, he developed an obsequious style, had few friends, and was generally aloof in his relationships. On the subscales of the IIP at intake, he had prominent concentration of problems in the avoidant, nonassertive, exploitable, and overly nurturant octants, consistent with the initial clinical descriptors and diagnoses. By the follow-up at 6 months after the termination of therapy, he was bothered only a little by social withdrawal and low confidence in dealing with people. He had become more assertive, less avoidant, and less focused on having to please others. He was bothered very little by self-esteem problems. Although the concentration of interpersonal problems is still mostly in the submissive hemisphere, the entire set is markedly reduced in intensity. In addition, there is a relative shift toward the friendly side.
Case 2 (Figure 2) is a 43-year-old single female musician and teacher with a history of binge eating and low moods with lethargy and oversleeping. She complained of being bothered very much by her difficulty controlling her weight, failure to follow through on ideas and plans, and tendency to be critical of herself and others. She attributed the failure to follow through on a fear of exposing too much of herself, since she is sensitive to criticism. However, she attributed her present overly critical behavior to having had parents who were overly critical of each other and of her and to having studied music with a critical teacher. She was extremely concerned that she worried too much about her family's reactions to her. She thought that she was not competitive enough with her peers, subordinating her needs in order to maintain relationships. She fit SCID/DSM-III-R criteria for simple phobia and personality disorder not otherwise specified. She had been grieving over a former lover who had died the year before her treatment began. Her father is described as distant and unable to communicate verbally on a deep level with her. She describes herself as warm and gregarious. She believes that her musical career started while she was learning the flute in junior high school and had a crush on her music teacher. At intake, she demonstrated a concentration of problems on the IIP subscales in the nonassertive, exploitable, and overly nurturant areas. She had difficulty expressing anger, holding to her position, and competing in a straightforward way. At termination, there was a clear decrease in most areas of interpersonal difficulty, including a reduction in her critical behavior. She was better able to compete at work and to make reasonable demands of those around her both at work and in her personal life. Brief Supportive Psychotherapy had a clear impact at follow-up on her overall position with respect to interpersonal space. Although still somewhat reticent and critical, she had become less overly friendly at the expense of autonomy and had become more assertive.
Case 3 (Figure 3) is a 53-year-old single man who is a college-educated art director. He was diagnosed as HIV- positive 1 year prior to intake. In his youth he was sickly and physically slight and had been repeatedly verbally and physically attacked by neighborhood “toughs.” His mother, whom he described as warm and supportive, died of multiple sclerosis when he was in his late twenties. His father was described as a strict, demanding, hot-tempered, but hard-working man with poor self-esteem, impatient and not understanding of his son. The father committed suicide after the death of the mother. The patient had a multitude of physical problems as a youth, including rheumatic fever, asthma, and enuresis. He suffers currently from ulcerative colitis. As an adult, he has enjoyed long-term monogamous relationships. His main complaints at intake were of being bothered very much by anxiety and depressed mood with regard to his medical condition, his job situation, and his feelings of overdependency in his current relationship. He had always been overly dependent, rejection-sensitive, and sensitive to criticism. In his current relationship he would passively tolerate the rejections by his lover without taking a stand. He met SCID/DSM-III-R criteria for anxiety disorder not otherwise specified and obsessive-compulsive personality disorder. There was a concentration of IIP problems in the nonassertive and exploitable octants of the circumplex. Although he remained very distressed by his anxiety and depression regarding his physical condition at follow-up, there was a clear reduction in the impact of anxiety and depression regarding his job and relationship. In his relationship with his lover, he was able to take a less submissive and humiliated position. Correspondingly, there is a marked decrease in problem severity in the friendly-submissive quadrant.
Case 4 (Figure 4) is a 38-year-old single male graduate student whose target complaints were depressed mood, chronic negativism expressed as negative and pessimistic thoughts, and dysphoria characterized as anxious loneliness. He described having close relationships with his parents and best friend while growing up, and he was close as an adult with his mother, whom he described as an anxious worrier. His father died from a myocardial infarction before his eyes, which the patient described as a severe loss. He described his feelings of loneliness as lifelong. His last relationship with a woman was 3½ years ago. He does not make friends easily. He often answered questions in a vague style. He met SCID criteria for current moderate major depression, and for paranoid personality disorder of mild severity. He admitted to being self-absorbed at times, passive and withholding of his thoughts in his personal relationships, and prone to losing his temper, feeling out of control, and saying the wrong thing. He had problems with self-worth. On follow-up, he still had problems with negativistic thinking and severe anxious loneliness and was somewhat less depressed. He still had major problems socializing and feeling close to others.
From an interpersonal point of view, maladjusted individuals operate within a narrow scope of relationships that serves to reinforce their rigid models of self and other. At the broadest, an interpersonal model of change would demonstrate an elaboration of self, generating a more differentiated and adaptive self from a more stereotypic and less evolved one.39 A dynamic interpersonal model of behavior places neurotic problems as maladaptive interpersonal strategies that attempt to maintain a connection to an earlier attachment figure.40 Therefore, one attempts to treat what is more classically viewed as neurotic behavior by fostering a change in the maladaptive interactional patterns seen in the therapeutic relationship, in hopes that the change will generalize into everyday-life encounters. However, temporary decreases in anxiety or depression or increases in self-esteem or optimism due to supportive interventions will not be sustained if the core pathogenic belief that maintains the problem is unaltered.20 Given this frame, the overall outcome data and the cases described above support the premise that patients can develop a more differentiated and adaptive self as a result of the interactions present in supportive psychotherapy. These changes can be measured as reductions in interpersonal problems that are maintained after termination of treatment.
Cases 1–3 demonstrate changes consistent with the types of interpersonal problems measured by the IIP, such as friendly submissiveness, which is the problem most responsive to brief dynamic treatment.40,41 The relative efficacy of supportive psychotherapy was greater in the submissive hemisphere of the control axis. Perhaps, from an experiential view, it is easier for a patient to reclaim control that was divested as an affiliation-maintaining strategy than it is to give up interpersonal control already acquired for the purpose of shoring up security. However, Cases 2 and 3 demonstrate as well that some reductions in problems with hostility and hostile submission may occur with supportive psychotherapy.
Horowitz et al.40 have found that patients who complain of relatively more distress from non-interpersonal problems (such as somatic complaints or inability to work) fare less well and have a higher rate of dropout from brief dynamic therapy than patients whose distress relates more to interpersonal problems. Supportive psychotherapy has traditionally been recommended for people with external issues. The patient presented as Case 3 was able to make use of supportive psychotherapy to reduce depression and anxiety related to relationship and job function while he remained distressed by ulcerative colitis and complications of HIV seropositivity.
In considering the relative efficacy of psychotherapy, and specifically supportive psychotherapy, for problems on the friendly side of the affiliation axis, it makes intuitive sense that the therapeutic alliance, a strong predictor of outcome in all therapies,42,43 is likely to be stronger with patients who are able to bond and work with the therapist. Generally, “friendly” complementary affiliations between patient and therapist early in treatment significantly predict positive outcome, while “hostile” complementarity predicts poor outcome.27,28 Indeed, the general consensus from research regarding single-session ratings early in treatment suggests that the affiliation axis is most pertinent to both therapeutic alliance and treatment outcome. By its very construction, supportive psychotherapy—with its conversational style, primary focus on self-esteem, and avoidance of interpretation of positive transference—particularly and explicitly emphasizes the affiliative dimension of the therapeutic encounter. Despite this emphasis, however, in our experience help-rejecting complainers and liars, patients with typical problems of hostile dominance, do poorly in supportive psychotherapy, as they do in other treatments. Sampson and Weiss44 propose that a cyclical process occurs within psychotherapy wherein this type of patient sets for the therapist a transference test which, if failed, confirms the patient's pathogenic beliefs regarding self and others.
Generalization of the findings in this study is limited by the small sample size. Thus, replication with a larger data set, as well as with other measures of interpersonal functioning, is needed to more definitively establish that supportive psychotherapy can lead to change in personality, and not just symptomatology. These data do, however, provide preliminary experimental evidence for significant and lasting improvement in interpersonal problems in subjects treated with supportive psychotherapy. Analysis of change in the IIP subscales across the different cells of the Beth Israel Medical Center Brief Psychotherapy Research Study might reveal differential efficacy for different types of interpersonal problems. The specific efficacy of supportive psychotherapy for certain types of interpersonal problems versus others, and in comparison with other types of treatment, remains to be demonstrated.
This work was presented in part at the American Psychiatric Association 149th annual meeting, May 4–9, 1996, New York, NY, and the Society for Psychotherapy Research 28th annual meeting, Geilo, Norway, June 25–29, 1997.