| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ALL ISSUES | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Regular Article |
Received March 8, 2000; revised June 9, 2000; accepted June 23, 2000. From the Departments of Psychiatry, University of Alberta, Edmonton, and University of British Columbia, Vancouver, Canada. Address correspondence to Dr. Joyce, Department of Psychiatry, University of Alberta, 8440 112 Street, Edmonton, Alberta, Canada T6G 2B7.
 |
Abstract |
|---|
 Top Abstract IntroductionMETHODSRESULTSDISCUSSIONCONCLUSIONSReferences |
|---|
Key Words: Psychotherapy, Short-Term • Therapeutic Alliance
|
Introduction |
|---|
|
TopAbstractIntroductionMETHODSRESULTSDISCUSSIONCONCLUSIONSReferences |
|---|
The patient's expectancies of each participant's role behaviors during therapy would likely develop as a function of several factors, including that patient's 1) motivation for change; 2) understanding of or previous experience with psychotherapy; and 3) anticipation of certain forms of interpersonal response on the part of the therapist.4 This last element would be influenced by the patient's internalized object relations (i.e., past history of relationship gratifications or losses). In like fashion, the therapist's role behavior expectancies are likely to be based in part on his or her 1) past clinical experience; 2) theoretical and experiential working knowledge of the components of treatment associated with successful outcome; and 3) initial impressions of the patient. Once again, this last element would be influenced by the therapist's internal representations of relationships with other people, in particular previous patients.
Greater agreement between patient and therapist on expected role behaviors would lessen the need for an involved negotiation of the therapy relationship, would reduce the potential for conflict regarding patient and therapist responsibilities, and would protect against early disappointment on the patient's part. Thus, the extent of the agreement between the patient's andtherapist's expectancies of how each will behave during therapy may act as a determinant of whether the relationship "gets off on the right foot" and readily develops into a productive working alliance or flounders and has difficulty developing a collaborative focus on the tasks and goals of therapy.5
We predicted that expectancies regarding role behaviors would be more strongly related to the change in the alliance over the course of therapy than to the average level of the alliance based on an aggregation of alliance scores across sessions. Thus, expectancies about role behaviors (i.e., anticipations about what each party would actually do during therapy) were predicted to have a greater bearing on the variable or dynamic characteristics of the alliance than on some summary index of the overall quality of the alliance.
We also postulated that the extent of the agreement between patient and therapist role behavior expectancies would be a stronger determining factor than the expectancies of either patient or therapist alone. Greater disagreement regarding expected behaviors of self and other would require a more complicated negotiation of the therapy relationship. Consequently, we predicted that congruence of patient and therapist expectancies regarding role behaviors in therapy would provide for the strongest prediction of growth in the therapeutic alliance over the course of treatment.
The current study is based on a controlled clinical trial of short-term, time-limited individual (STI) psychotherapy completed in Edmonton, Canada.6 The trial provided evidence that the patient personality variable Quality of Object Relations (QOR) was directly related to the strength of the therapeutic alliance and beneficial treatment outcome.7 That is, a more mature (high) QOR is associated with a higher average level of the alliance and greater benefit from therapy. In contrast, further analyses identified that a progressive increase in the quality of the therapeutic alliance across therapy was predictive of benefit for patients with more primitive (low) QOR.8 The low-QOR patient may approach therapy with distorted interpersonal expectations because of a history of chronically unsatisfying and abusive relationships. Our research suggests that the development and the concrete experience of a gratifying therapeutic relationship may be more beneficial for the low-QOR patient than gaining insight into the workings of past and current unhealthy relationships. Therefore, our third prediction was that the relationships between role behavior expectancies and change in the therapeutic alliance would be more pronounced among the low- QOR patients.
This study is the second based on a comprehensive assessment of expectancies that was conducted as part of the controlled trial. The first report9 considered expectancies regarding the value and difficulty of the "typical session" as predictors of the average level of the therapeutic alliance and treatment outcome. That study provided evidence that expectancies about session benefit are strongly related to the overall strength of the alliance and are also predictive, though less strongly, of treatment benefit in STI therapy. The current study examined the patient's and therapist's role behavior expectancies—that is, each participant's anticipation of how both parties would behave during therapy sessions.10
The current study had two objectives. First, we wanted to identify the underlying dimensions associated with role behavior expectancy ratings provided by the patient and therapist. For this purpose, we considered the dimensions associated with the patient's and therapist's role behavior expectancies for themselves and for their counterparts; we also considered the congruence between patient and therapist ratings5 to identify dimensions of expected behavior associated with the joint view of each participant's role. Second, we wanted to examine the association between role behavior expectancies and the therapeutic alliance for patients of differing QOR (the alliance being expressed both in terms of the average level and the change over the course of treatment).
|
METHODS |
|---|
|
TopAbstractIntroductionMETHODSRESULTSDISCUSSIONCONCLUSIONSReferences |
|---|
Setting, Patients, and Therapists
The setting for the trial was the Psychiatric Walk- In Clinic, Department of Psychiatry, University of Alberta Hospital Site, Edmonton, Canada. Following referral and informed consent procedures, patients were matched in pairs on QOR, age, and gender and were randomly assigned to immediate or delayed therapy and to one of 8 project therapists. During a 3-year period, 86 of 105 patients who began therapy completed the protocol. Sixty-four cases formed a sample that was balanced for QOR, treatment condition (immediate vs. delayed), and therapist.
Diagnoses were made by an intake therapist according to the DSM-III11 after an initial assessment and consultation with a staff psychiatrist. For the sample of 64 patients, 72% received Axis I diagnoses, the most frequent being affective (27%), impulse control (8%), or anxiety (6%) disorder. An Axis II diagnosis was assigned for 27% of the sample, the most frequent being dependent (14%) or avoidant (5%) disorder. The average age of the patients was 32 years (SD=8, range 21– 53 years), and 62% were female. The patients' presenting problems were representative of an outpatient psychotherapy population—namely, difficulties with depression, anxiety, low self-esteem, and interpersonal conflict. Three psychiatrists, one psychologist, and four social workers served as therapists in the study. Their average age was 40 years, and they had practiced dynamically oriented individual therapy for an average of 11.5 years.
Therapy
The short-term, time-limited therapy was psychodynamic in orientation and followed a technical manual that drew on the approaches of Malan12 and Strupp and Binder.2 Interpretation and clarification are emphasized relative to support and direction. Twenty weekly sessions of 50 minutes' duration were contracted. The average number of sessions attended was 18.8. The technical nature of the therapy was verified by a content analysis of therapist interventions for eight sessions (sessions 4, 7, 9, 11, 14, 16, 18, and 20) using the Therapist Intervention Rating System.13 On average, there were 44 interventions, 11 interpretations, and 5 transference interpretations per session, confirming that the therapists had been active, interpretive, and transference-oriented as intended.
Measures
Predictor Variables—Role Behavior Expectancy Ratings:
The expectancy ratings administered to patients and therapists addressed the role behavior of each party during therapy sessions. Patients completed expectancy ratings as part of the initial outcome assessment, prior to meeting their therapists. The first two sessions of STI therapy are commonly used for history-taking and development of rapport. Therapists rated expectancies following these sessions, after formulating an initial impression of the patient.
Each participant rated the expected or anticipated behavior of self and other during therapy sessions. Rating items addressed therapy-related behaviors that would be readily evident to an objective lay (nonclinical) observer of the session. Items were evaluated by the investigators for face validity and adequate coverage of therapy behaviors. Ratings of anticipated attitudes or personality qualities of self and other (commonly assessed in previous expectancy research14,15) were not developed. Twelve role behavior items (e.g., I expect that I/the therapist will talk) were formulated and adapted to each of the four perspectives (patient rating of self and of therapist, therapist rating of self and of patient). Each item was rated on a 6-point Likert-type scale ranging from 1 ("very seldom") to 6 ("very frequently"). Table 1 presents the complete set of 12 role behavior expectancy items.
|
Dependent Variable—Therapeutic Alliance:
The alliance was defined as the nature of the working relationship between patient and therapist. It was rated by the patient and the therapist by means of six 7-point Likert- type items that ranged from 1 ("very little") to 7 ("very much"). Four items were rated after every therapy session (immediate alliance ratings) and two were rated after each one-third of the therapy course; that is, after sessions 7, 14, and 20 (reflective alliance ratings). The immediate alliance items addressed whether the patient 1) had talked about private important material, 2) had felt understood by the therapist, 3) was able to understand and work with the therapist's interventions, and 4) felt that the session enhanced understanding. The two reflective alliance items addressed Luborsky's16 concept of the helping alliance; that is, whether the patient and therapist worked well together (collaboration) and the patient experienced the therapist as helpful (helpfulness).
The six items were averaged across their respective assessments, which enhanced their reliability. Each set of six (patient, therapist) was subjected to a principal components analysis and varimax rotation. The analysis of therapist ratings resulted in two factors accounting, respectively, for 63% and 22% of the common variance. The four items rated after every session (immediate alliance) loaded heavily on the first factor (mean loading=0.87) and were internally consistent (Cronbach's alpha=0.92). The two items rated after each third of therapy (reflective alliance) loaded heavily on the second factor (mean loading=0.92) and were highly correlated (r=0.76, df=62, P<0.001). In the case of the patient ratings, a single factor emerged and accounted for 77% of the common variance. The six items loaded highly on the factor (mean loading=0.88) and were internally consistent (Cronbach's alpha=0.94). Thus, three variables—one patient-rated factor and two therapist-rated (immediate, reflective) factors—served as measures of the therapeutic alliance. Because of the frequency of missed ratings for any given session, alliance factor scores were calculated for each third of therapy (sessions 1–7; 8–14; and 15–20). The scores on the alliance variables thus reflected patient and therapist views of the working relationship across beginning, middle, and termination phases of STI therapy.
Patient Grouping Variable—Quality of Object Relations:
QOR is defined as a person's internal enduring tendency to establish certain types of relationships with others.17 The dimension ranges across five levels of object relations (primitive, searching, controlling, triangular, and mature), encompassing a developmental sequence of maturation. The five levels and anchor points of the QOR dimension are presented in Table 2. In the controlled trial, the assessment of QOR comprised two 1-hour clinical interviews. During each semistructured interview, the patient's lifelong pattern of relationships is examined. Relationships during childhood, adolescence, and adulthood are reviewed. The interviewer considers the relationship patterns in terms of behavioral manifestations, regulation of affect, regulation of self-esteem, and historical antecedents for each of the five levels. Following the interview, the assessor distributes 100 points among the levels and derives a single global score ranging from 1 to 9. At the primitive or low end of the 9-point scale, relations are characterized by inordinate dependence, extreme reactions to real or imagined loss, and destructiveness. At the mature end, relations are characterized by equity and the expression of love, tenderness, and concern.
|
5.0). Each subgroup was composed of 32 patients.
Approach to Analysis
Expectancy Rating Dimensions:
Six principal components analyses were conducted to identify dimensions underlying the ratings of expected role behaviors for each of the four perspectives (patient-of-self, patient-of- therapist, therapist-of-self, therapist-of-patient) and for congruence of expectancies regarding the patient role and the therapist role. The principal components analyses were conducted by using the total sample of 64 therapy cases. Factors identified in each analysis had to fulfill two criteria to be retained as potential predictors of change in the alliance. First, the factor had to account for a substantial proportion (15% or greater) of the variance in the item ratings employed in the components analysis. Second, the items that defined each factor (based on loadings of 0.40 or greater) had to show substantial internal consistency. Cronbach's alpha coefficient was calculated to evaluate the internal consistency of each factor. An alpha coefficient of 0.70 or greater, indicating that the items comprising a factor showed moderate or better internal consistency, was required for retention of the factor as a predictor variable.
Predicting Change in the Alliance:
Change in the therapeutic alliance was defined as the variation in the alliance ratings across the course (in thirds) of therapy within each case. The hierarchical linear modeling (HLM) procedure of Bryk and Raudenbush21 was employed to assess whether 1) variation across cases in the pattern of change of the alliance was significant, and 2) the role behavior expectancy factors derived from the principal components analyses could account for this variation. The HLM procedure involves two levels. The first level (the unconditional HLM analysis) focused on the change in the alliance within cases, represented by the slope of the alliance over time (thirds of therapy) unique to each individual case. A positive slope indicated an increase in the alliance over the thirds of therapy and a negative slope indicated a decrease. Variation of the slopes across patients was calculated and tested for significance with the chi-square goodness-of-fit statistic. If this variation was found to be significantly different from zero, the second level of the HLM analysis was conducted. The second level (the conditional HLM analysis) focused on predictors of the variation in the patterns of change. The individual slopes represented the dependent variable in the conditional analysis and were regressed on the predictor variable (a role behavior expectancy factor). A t-test assessed whether the relationship between role behavior expectancy and variation in the patterns of change in the alliance was significant.
Finally, comparisons of the correlations involving role behavior expectancy variables and either the slope or average level of the alliance were conducted. Separate HLM and correlation analyses were performed for the groups of low-QOR and high-QOR patients (n=32 per group).
|
RESULTS |
|---|
|
TopAbstractIntroductionMETHODSRESULTSDISCUSSIONCONCLUSIONSReferences |
|---|
|
For patient expectancy of therapist behavior, two factors met the retention criteria. The first (Active Therapist) reflected the patients' expectation for a therapist who would ask questions, explore interpersonal relationships, and suggest topics for examination. The second (Therapist Content) reflected the content areas patients expected the therapist to address, including symptoms, sexuality, childhood memories, the here- and-now relationship, and the linking of past experiences with current difficulties.
For therapist expectancy of own behavior, one factor (Therapist Work Focus) met the retention criteria. It reflected therapists' expectation of what the work focus would be during sessions. Therapists anticipated a strategy that involved linking past experiences to current difficulties and that included consideration of childhood events, important interpersonal relationships, and the here-and-now therapy relationship. The therapists' expected in-session behavior therefore corresponded to the approach outlined in the interpretive therapy manual.
For therapist expectancy of patient behavior, two factors met the retention criteria. The first (Patient Work Focus) reflected the therapists' expectancy of patient behavior representative of work in therapy, involving discussions of the therapeutic relationship, patient initiation of discussion of certain topics, and patient expression of personal feelings. The second factor (Patient Resistance) reflected the therapists' expectancy of patient behavior representative of resistance, involving avoidance of difficult issues, a questioning stance, and little attention to links between past and current experience or to potential problem solutions. The factors thus represented the therapists' expectancies of patient behavior that constituted both positive and negative contributions to the therapy process.
For congruence of expectancies regarding patient role, none of the five factors that emerged in the analysis met the two retention criteria.
For congruence of expectancies regarding therapist role, one factor (Supportive Therapist Role) met the retention criteria. It defined a supportive role for the therapist, characterized by a focus on problem resolution, personal self-disclosure, interested inquiry, determination of session topics, and activity.
Examination of the correlations among the seven retained role expectancy factors indicated a considerable degree of independence among the variables. Only three relationships were significant. Patients with higher expectancies of making a positive contribution to the process of therapy (Patient Process Contribution) also had higher expectancies for an Active Therapist (r=0.33, df=62, P<0.01). Greater expectancies in these two areas were also associated with more disagreement between patient and therapist on the importance of a Supportive Therapist Role (r=0.29, df=62, P<0.05, and r=0.46, df=62, P<0.001, respectively).
The Unconditional HLM Analysis:Pattern of Alliance Change
A series of unconditional hierarchical linear modeling analyses was conducted to examine the patterns of change for each of the three alliance variables over the course of therapy across patients. Given that a total of 6 unconditional HLM analyses were performed, a Bonferroni-adjusted significance criterion of P<0.008 (0.05/6) was employed. The analyses showed that there was significant variation in the patterns of change in the alliance across cases within the high-QOR and low- QOR samples. Table 4 presents the significant findings from the unconditional HLM analyses.
|
In low-QOR cases, for the patient-rated alliance, variation of the slopes (reliability=0.14) was not significantly different from zero; differences in the patterns of change for the patient-rated alliance among the low- QOR cases were negligible. For the therapist-rated immediate alliance, variation in the slopes (reliability=0.66) was significant. Variation of the slopes for the therapist-rated reflective alliance (reliability=0.62) was also significant. Among the low- QOR cases, tests of predictor variables were therefore possible only for the variation in the two therapist-rated impressions of the alliance.
The above analyses were based on a linear model. Analyses with a curvilinear model (i.e., testing whether the alliance ratings followed a curvilinear pattern across thirds of therapy) revealed no significant increase in the variation of the slopes. Individual slopes derived from the linear model were thus used in the predictive analyses. In summary, the unconditional HLM analyses indicated that three patterns of change in the therapeutic alliance involved sufficient variation to warrant study of possible predictor variables. For the high-QOR patients, significant variation was evident in the patterns of change for the patient-rated alliance. For the low- QOR patients, significant variation was evident in the patterns of change for the two therapist-rated impressions of the alliance.
The Conditional HLM Analysis:Expected Role Behaviors as Predictors
For each of the three alliance variables that showed significant variability in the patterns of change across cases, seven separate conditional HLM analyses were conducted. Each analysis tested a different role behavior expectancy variable as a predictor of the particular dependent measure (i.e., the slopes of each alliance variable). A Bonferroni adjustment was applied to each set of seven conditional HLM analyses. The adjusted significance criterion for a given predictor variable was thus 0.007 (0.05/7). Two predictor relationships that met the adjusted significance criterion were identified.
For high-QOR cases, the patient's expectancy of being able to contribute to the process of therapy (Patient Process Contribution) was significantly and inversely associated with the pattern of change in the patient- rated alliance (t=–3.77, df=30, P<0.001). For high- QOR patients, the lower the initial expectancy of being able to contribute to the process of therapy, the greater the increase in the patient-rated alliance across thirds of therapy. The predictor relationship accounted for a substantial proportion (49.7%) of the true variance in the slopes of the patient-rated alliance among the high- QOR cases.
For low-QOR cases, greater patient–therapist congruence regarding expectancies of a Supportive Therapist Role was significantly and directly associated with the pattern of change in the therapist-rated immediate impression of the alliance (t=–4.27, df=30, P<0.0001). In other words, greater patient–therapist agreement in anticipating that the therapist would behave in a supportive manner was associated with a progressive increase in the therapist's ratings of the alliance over the course of therapy. The predictor accounted for 53.5% of the true variance in the slopes of the therapist-rated immediate impression of the alliance among the low- QOR cases.
Relationships of Role Expectancy to Slope of Alliance vs. to Average Level of Alliance
The two significant predictive relationships identified in the conditional HLM analyses could be represented by the Pearson correlation between the role expectancy factor scores and the individual slopes of the alliance variable. These relationships could then be contrasted with the association between the expectancy factor and the average level of the alliance across therapy. The average level of the alliance was calculated by aggregating alliance scores across sessions and thirds of therapy. The method provided by Cohen and Cohen22 to assess the significance of the difference between dependent correlation coefficients was employed to determine whether role behavior expectancies had significantly different relationships with the pattern of change versus the average level of the therapeutic alliance. Table 5 presents the relevant correlation coefficients and the results of the tests of significance.
|
For low-QOR cases, the correlation between scores on the congruence expectancy factor Supportive Therapist Role and the slopes of the therapist-rated immediate impression of the alliance was large and inverse. In contrast, the correlation between this factor and the average level of the therapist-rated immediate alliance was smaller and nonsignificant. The difference between the two coefficients was nonsignificant at P=0.08; in absolute value, the expectancy factor–slope relationship was slightly the larger of the two. The two indices of the alliance (slope, average level) were moderately correlated (r=0.36, df=29, P<0.05).
In summary, both comparisons indicated that the expectancy factor had a stronger relationship to change in the alliance than to the average level of the alliance across therapy. Among the high-QOR cases, the expectancy factor–slope relationship also differed in sign (direction) relative to the expectancy factor–average level relationship.
|
DISCUSSION |
|---|
|
TopAbstractIntroductionMETHODSRESULTSDISCUSSIONCONCLUSIONSReferences |
|---|
Among high-QOR patients, the patients' expectancy of their own behavior (Patient Process Contribution) was inversely associated with change in the patient-rated alliance. Among low-QOR patients, the congruence of expectancy between patient and therapist regarding the therapist's supportive behavior (Supportive Therapist Role) was directly associated with change in the therapist-rated immediate impression of the alliance. These findings suggest that the particular role behavior expectancies associated with varying patient QOR have differing influences on the therapy process and may thus require different strategies from the STI therapist.
Implications for Therapeutic Strategy
High-QOR Patients: Patients of high QOR are more likely to present with a history of satisfactory, give-and- take relationships. As a consequence, we believe that these patients are more able to tolerate the demands of the interpretive therapy situation, are less fearful of losing the relationship, and are more receptive to the therapist's interpretive interventions.6,7
These patients commonly have concerns about interpersonal control or about faring successfully in competitive relationship situations. At the time of presentation, circumstances in the patient's relationships may have contributed to feelings of demoralization, failure or defeat, and pain associated with interpersonal conflict. Some high-QOR patients may not believe that they will be able to master the challenges without the assistance and direction provided by a professional helper. This belief may translate into a low expectancy of being able to contribute to the work of therapy and to the resolution of problems. Through the mechanism of the self-fulfilling prophecy, the patient may actually provide a low level of work in the initial stages of therapy. Once therapy has been under way, however, the experience of being understood by a concerned listener and recognizing that problems can be addressed by making use of the therapy relationship may encourage greater patient engagement in the collaborative tasks of treatment. A progressive increase in effective interpretive work and the therapeutic alliance can then follow.
Conversely, other high-QOR patients may begin therapy with a high expectancy of being able to contribute to the therapy process. Given their internal object relations, this may reflect the patients' views that they will control the direction of the treatment or that the therapist represents an adversary to be overcome in the process.25 Again, through the mechanism of the self- fulfilling prophecy, the patient may initially provide considerable clinical material in order to impress the therapist. However, the therapist does his or her best not to engage in an enactment of the patient's controlling or competitive relationship pattern, but instead comments on the relationship dynamics as representing an opportunity for exploration. Developing a focus on these relationship patterns becomes the work of therapy and can eventuate in a strain being placed on the patient's experience of the alliance, resulting in a minimal increase or an actual decrease of the patient's alliance ratings across therapy. However, the average level of the patient-rated alliance may still reflect sufficient strength in the working relationship to promote a positive benefit. In other words, the work of therapy, while difficult and a stress on the therapeutic alliance, may nonetheless be successful in helping the high-QOR patient revise internal object relations in the direction of greater maturity and thus achieve a reasonable outcome.
Considerably high or low patient expectancies of being able to contribute to the therapy process may thus have implications for the development of a working relationship with the high-QOR patient. The key consideration would appear to be the level of the expectancy regarding his or her potential contribution that the high- QOR patient brings to the therapy situation. An overly low expectancy can be augmented by the re-moralizing effects of beginning psychotherapy,26 while an overly high expectancy can be modified by an interpretive focus on the maladaptive object relations implied by the patient's anticipated course of treatment. Discussion of the roles and responsibilities of each party in the relationship early in the course of treatment can assist the therapist in determining how best to address the patient's expectancy of his or her own in-therapy behavior. It is therefore contingent on the therapist to address considerably high or low expectancies on the part of the high- QOR patient early in the treatment process to ensure that an appropriate foundation for the therapeutic alliance can be put into place.
This construction may account for the inverse relationship between the patient's expectancy of being able to contribute to the therapy process (Patient Process Contribution) and the pattern of change in the patient-rated alliance across treatment for the high-QOR patients. In contrast, the patient expectancy factor was directly associated with the average level of the patient- rated alliance. The two measures of the alliance (slope, average level) were independent. This would suggest that, from the perspective of summarizing the quality of the entire therapy relationship, patient expectancies of being able to contribute to the therapy process are indeed important to a strong alliance. Our previous research9,10 indicated that the average level of the patient-rated alliance (versus the change in the alliance) was directly associated with improvement in STI therapy for high-QOR patients. The findings of the current study indicate that the average level of the alliance is directly influenced by the patient's expectancy of being able to contribute to the therapy process. However, if patient expectancies about making a contribution are considerably high or low, they have a substantial potential to influence the pattern of change in the high-QOR patient's perception of the relationship across treatment. These contrasting findings for average level and pattern of change also highlight how different perspectives on the therapeutic alliance have different implications for understanding the therapy process.
Low-QOR Patients: One congruency of expectancy factor was retained for the predictive analyses. The Supportive Therapist Role factor proved to be a strong predictor of the pattern of change in the therapist's immediate impression of the therapeutic alliance in the treatments of low-QOR patients. For these patients, the prediction indicated a relationship between agreement on the expected supportive role of the STI therapist and a progressive increase in the therapeutic alliance across thirds of therapy. The important element here appears to be the degree of shared expectation regarding the supportive therapist role held by the patient and therapist.
We know from our previous work using the HLM method that a progressive increase in the quality of the therapeutic alliance across STI therapy is predictive of treatment benefit for low-QOR patients.10 For these patients, a congruent expectation with the therapist regarding the importance of a supportive therapist role at the outset of therapy is directly related to just this kind of progressive growth in the alliance, as perceived on a session-by-session basis by the therapist. By influencing the development of the alliance, this shared perspective appears to play an important role in the short-term interpretive therapy of low-QOR patients.
Low-QOR patients are characterized by a history of nongratifying or abusive relationships and difficulties regulating affective experience and self-esteem. Independent of their presenting problems, these patients come for therapy seeking nurturance and assistance by an understanding other to resume an arrested developmental trajectory.27 For these patients, a gratifying therapy relationship may be at least as important as the achievement of insight into intrapsychic and interpersonal conflict.10 In the pre-treatment or early therapy situation with the low-QOR patient, there apparently needs to be congruence between what the patient expects from the therapist—commonly referred to in the contemporary literature as "holding" or "mirroring"— and what the therapist expects will be important to promote a good working relationship in the interpretive therapy situation. What role seems best for the therapist may vary widely across patients as a function of their previous history, understanding of the therapy process, and so on. A strictly interpretive and confrontational therapy approach28 that abstains from the provision of support may simply be inappropriate for certain low-QOR patients.
During the pre-therapy preparation or early phase of STI therapy, the therapist should aim to assess the degree of support expected by the low-QOR patient. During these early contacts, an open discussion of the roles each party will play in the upcoming therapy would be one way to achieve this end. The development of an explicit agreement on how the therapist will work to support the low-QOR patient during interpretive treatment would likely promote the patient's effective use of the STI therapy approach.
Limitations of the Study
Limitations of the current study can be noted. Patient and therapist role expectancy ratings in this study addressed a set of simple behaviors that are commonly associated with the interaction during psychotherapy sessions. The rating items were initially generated and developed on a conceptual rather than an empirical basis. An initial attempt to discover the underlying dimensions associated with the rating items was one of the purposes of the current study. Consequently, the findings of this study must be considered with caution until replication is provided. An important objective for future study would be to develop expectancy rating items empirically, by asking patients and therapists about their expectancies, developing and refining items based on those reports, and assessing the reliability and validity of the final versions of the rating items with independent samples. Ultimately, this investment could provide for a more definitive test of the relationships identified in this study.
The principal components analyses conducted on the six sets of expectancy item ratings were based on the sample of 64 therapy cases and thus offered a relatively low ratio of subjects to variables (5:1). Concerns about the reliability and generalizability of the identified factors could be raised. To deal with the limitation posed by the small sample, relatively stringent criteria for retention of factors as predictor variables were imposed. The factors retained had substantial internal consistency and some degree of face validity, suggesting that we could have reasonable confidence that the factors adequately sampled the role behaviors patients and therapists anticipated of themselves and their counterparts in STI therapy. Still, the findings reflect only the characteristics of this sample of therapy cases until the expectancy dimensions are replicated by independent investigators.
Analyses addressing variation in the patterns of alliance change (unconditional HLM analyses) and the ability of expectancy predictor variables to account for this variation (conditional HLM analyses) involved a large number of statistical tests. This likely had an inflationary effect on the probability of Type I error (that is, erroneously identifying a relationship as significant). Adjustments to the significance criterion were implemented in both sets of analyses in order to protect against this inflation while ensuring that important relationships evident in the data could still be identified. Once again, these adjustments were made to retain a reasonable level of confidence in the findings and in no way supplant the importance of independent replication. Finally, the significant relationships between role expectancy and the alliance variables explained only modest amounts of variance in the latter, emphasizing that a large number of other factors (e.g., patient, technical, and relationship characteristics) are likely to be influential on the development and overall quality of the patient–therapist interaction.
|
CONCLUSIONS |
|---|
|
TopAbstractIntroductionMETHODSRESULTSDISCUSSIONCONCLUSIONSReferences |
|---|
|
References |
|---|
|
TopAbstractIntroductionMETHODSRESULTSDISCUSSIONCONCLUSIONSReferences |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ALL ISSUES | SEARCH | TABLE OF CONTENTS |
