| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ALL ISSUES | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Regular Article |
Received February 14, 2000; revised January 31, 2001; accepted February 1, 2001. From the Département Universitaire de Psychiatrie de l'Adulte, Faculté de Médecine de Lausanne, Switzerland. Address correspondence to Dr. Perry, Director of Research, Institute of Community and Family Psychiatry, Sir Mortimer B. Davis-Jewish General Hospital, 4333 Chemin de la Côte Ste-Catherine, Montreal, Quebec H3T 1E4, Canada.
 |
ABSTRACTS |
|---|
 Top ABSTRACTS INTRODUCTIONMETHODSRESULTSCLINICAL EXAMPLESDISCUSSIONREFERENCES |
|---|
Key Words: Therapeutic Alliance • Defense Mechanisms • Interventions
|
INTRODUCTION |
|---|
|
TopABSTRACTSINTRODUCTIONMETHODSRESULTSCLINICAL EXAMPLESDISCUSSIONREFERENCES |
|---|
How the early and the late alliance emerge in the patient–therapist relationship is a question crucial for learning how to influence alliance development. Beyond measuring the alliance, it is important to understand the patient and therapist factors contributing to the development of early alliance, as well as how these factors interact. Frieswyk et al.5 suggested that the patient's contribution alone influences alliance. By contrast, O'Malley et al.6 suggested that patient involvement is not solely an antecedent patient characteristic, but rather seems to develop over the first few sessions, probably as a result of the therapist's actions. Crits-Christoph et al.7 empirically demonstrated that the abilities to maintain a positive alliance and repair negative elements during treatment are intimately connected to the technical interventions made by the therapist. However, the nature of the relationship between therapist interventions and patient characteristics still has to be clarified.8
Patients' defensive functioning may affect alliance development. Vaillant9 suggested that intermediate-level defenses, such as rationalization, reaction-formation, and intellectualization, may interfere with an individual's capacity to engage in self-exploration. If such defenses play a fundamental role in affecting a therapeutic alliance, any change in them may depend on the frequency with which the therapist addresses or interprets these defenses. For instance, Foreman and Marmar10 found that the therapist actions that occurred most frequently in cases with improved alliance and good outcome were 1) addressing the subject's defenses, 2) addressing the subject's guilt and expectation of punishment, 3) addressing the subject's problematic feelings in relation to the therapist, and 4) linking the problematic feelings in relation to the therapist with the subject's defenses. Following Winston et al.,11 but using different methods, we wished to determine how the adjustment of the therapist's interventions to the patient's defense level predicts outcome. We hypothesized that when addressing defense mechanisms, it is important that the psychotherapist adjust his or her interventions to fit the patient's level of personality organization and ego strength.12–14 From this point of view, the level of defensive functioning is one measurable facet of the level of personality organization.12,14
From the perspective of dynamic psychotherapy, two broad types of intervention can be distinguished by their aims: interventions that offer support, and those that foster psychological exploration and understanding (interpretations). The appropriate proportions of these two types of intervention depend mainly on the patient's characteristics, including symptoms, functional level, and personality features such as underlying personality structure, defensive level, and ego strength.12,13 Gaston and Marmar15 suggested that when patients have difficulty in establishing an alliance, a greater proportion of supportive interventions will be more beneficial. We believe that the therapist has to choose the appropriate proportion of supportive to exploratory interpretations, giving consideration to the level of defensive functioning as an immediate indicator of personality organization and ego strength in a given session. We therefore hypothesized that the therapist will influence the early development of a positive alliance when there is a positive adjustment of therapeutic interventions to the patient's defensive functioning, defined as follows: whenever the patient's defensive functioning is low the therapist must be more supportive, whereas whenever the patient's defensive functioning is high the therapist should use more interpretive techniques.
This study considers two questions. The first is whether the patient's defensive level affects early alliance-building—specifically, is lower defensive functioning associated with poorer alliance development? The second is whether the development of alliance is influenced by the level of adjustment of therapist interventions to the patient's level of defensive functioning—specifically, does the optimal adjustment of intervention types differ with different levels of defensive functioning? This latter is a specific case of the more general proposition that alliance will improve when the therapist optimally adjusts technique based on the patient's level of functioning. This is a preliminary report of a larger project examining the influence of intervention adequacy on alliance development in patient–therapist dyads engaged in Brief Psychodynamic Investigation (BPI).16
|
METHODS |
|---|
|
TopABSTRACTSINTRODUCTIONMETHODSRESULTSCLINICAL EXAMPLESDISCUSSIONREFERENCES |
|---|
The five participating psychotherapists completed their training in psychiatry and psychotherapy in accordance with the rules of the Swiss Medical Federation and the Swiss Society of Psychiatry and Psychotherapy. All had extensive backgrounds in both long-term and short-term psychoanalytic psychotherapies (mean=19 years of experience, range 8–38 years).
Treatment
Considering how rapidly alliance develops, we chose to study it as early as possible. Developed in Lausanne, the BPI (previously called SPI) is a formalized 4-session technique of assessment for psychodynamic psychotherapy that focuses on the reasons for the consultation and the early interaction between the patient and therapist.16 The main objectives of the BPI are to establish a diagnosis, further the development of the early alliance, and develop an optimal plan to resolve any crisis situation with which the patient presents. The therapist helps the patient view his or her difficulties in a deeper way to facilitate choosing the best treatment relative to his or her disorder and expectations. After the first or at the beginning of the second interview, the consultant offers the patient an initial interpretation built on a hypothesis about the crisis. The therapist proposes three more sessions to discuss it, followed by an evaluation after the fourth interview. The interviews are nondirective, encouraging free association. Training for BPI was based on a manual and involved continuous participation of therapists in a 4-year supervision seminar with the author of the method, which employed videotaped sessions. Training and supervision was finished before the research began.
Sample
In the present study sample (N=12), the mean age was 28.5 years (SD=12.5); 11 patients were women, 9 were single, 2 were married, and 1 was divorced; 9 were college educated. After the BPI, 10 patients decided to continue with a psychotherapy and 2 decided not to pursue further treatment.
Alliance Profiles
The Helping Alliance Questionnaire, HAQ-1,18 was used for measuring the alliance. The patient completed it at the end of each interview. Examination of the first 21 patients from the total sample of 60 patients in the larger study showed that the mean alliance improved at each successive BPI session and that values at each session were strongly intercorrelated (Pearson's r, range 0.65 to 0.85). We defined the level of the alliance for each subject at each session as either low or high based on its being below or above the mean for the whole sample combining all four sessions. This resulted in three clearly different patterns of the evolution of the alliance: pattern I: alliance is high and stable (n=7); pattern II: alliance is continuously improving during BPI (n=10); pattern III: alliance is low and stable (n=4). Although theoretically expected, no decreasing or unstable profile of evolution has been observed to date. The three patterns were subsequently confirmed by visual inspection as well as by cluster analysis using the Ward method19 on the 21 subjects. We then selected as our sample (N=12) the first 4 subjects in order of admission who showed each profile, in order to study the contribution of defense mechanisms and therapist interventions for each pattern group.
Psychodynamic Instruments
The original version of the Defense Mechanism Rating Scales (DMRS) is American.20 We used the DMRS quantitative scoring described in the manual, with calculations being carried out for Overall Defensive Functioning scores (ODF), Defense Level Scores (DLS), and the proportion of specific defense mechanisms for each session. However, because of the small sample size, examination of the specific defenses will be deferred until a larger study sample is available. Studies have supported the validity and interrater reliability of the method.20–22 Studies using the DMRS on assessment interview data have yielded acceptable reliability for the summary and ODF scores.23 More recent applications to psychotherapy sessions have yielded intraclass R>0.70 interrater reliability on individual defense scores per session, and intraclass R=0.89 for the ODF score.22 The French version yielded comparably high reliabilities for the defense level and ODF scores, with both attaining intraclass R=0.80.
The Psychodynamic Intervention Rating Scales (PIRS) identifies nine types of interventions based on psychodynamic psychotherapy and broadly divided into three categories: interpretive, supportive, and therapy-defining.24 Each intervention is identified throughout an interview transcript, and the raw count is expressed as a proportion of all interventions. In a study of brief psychotherapy, the interrater reliabilities of the categories varied from kappas of 0.83 to 0.99.25 Some support for the construct validity of the PIRS was obtained by finding relationships between class of therapist intervention offered and the patient's initial level of distress, and between subsequent therapist elaboration and patient outcome. The present study, using a French version, yielded comparable interrater reliability for each category, with all kappas greater than 0.75.
We devised the following method to measure the adjustment of therapist interventions to the level of patient defensive functioning. The patient's level of defense functioning is summarized by the ODF, ranging from 1 (corresponding to low-level action defenses) to 7 (corresponding to high-level adaptive defenses), as shown in Figure 1. In an analogous manner, we constructed an Expressive-Supportive Intervention Level (ESIL) summary score for the PIRS by rank-ordering the intervention scores from the most supportive (1) to the most exploratory (7) and taking a weighted average (Figure 1). Acknowledgments (e.g., "Uh-hunh") by the therapist are omitted because they are considered neutral interventions. In a schematic way, one can then represent the adaptation of the therapist's interventions (ESIL) to a given patient's level of defensive functioning (ODF) in a session as a ratio (ESIL/ODF), which we designate the Adjustment Ratio (AR). We transform AR into z-scores for further data analyses. We can then examine the Adjustment Ratio for each session as a predictor of some outcome, such as the pattern of alliance development.
|
|
RESULTS |
|---|
|
TopABSTRACTSINTRODUCTIONMETHODSRESULTSCLINICAL EXAMPLESDISCUSSIONREFERENCES |
|---|
Early Alliance Building
The mean alliance score was successively higher for each session (session 2: 14.17±9.48; session 3: 16.42±12.64; session 4: 19.18±11.14). Comparing the first and fourth session yields a within-condition effect size of 0.58 for improvement in the alliance. Within this overall finding, three distinct profiles were evident (Figure 2). In the first and second sessions, Profile II (improving alliance) was not statistically different from Profile III (low, stable alliance), whereas by the third and fourth sessions, Profile II was statistically different from Profile III, but not from Profile I (high, stable alliance).
|
|
Adjustment:
In simple linear regression using all subjects, Adjustment Ratio at session 1 predicted alliance at sessions 3 (F=8.06, df=1,11, P=0.02) and 4 (F=7.88, df=1, 11, P=0.02), as well as subject's mean alliance across all four sessions (F=10.43, df=1, 11, P=0.009). The AR was significantly different across alliance profile groups at the first session (see Table 1), in which Profiles I and II did not differ from one another, whereas both differed from Profile III. The low mean AR for the Profile III group (low, stable alliance) indicates that for a given level of ODF, those therapists offered a greater level of support than did those in the other profile groups.
|
CLINICAL EXAMPLES |
|---|
|
TopABSTRACTSINTRODUCTIONMETHODSRESULTSCLINICAL EXAMPLESDISCUSSIONREFERENCES |
|---|
Case 416. Miss A. was a 25-year-old student in mathematics who sought help because of complaints of difficulties in her relationships with men. She had previously tried to practice relaxation on her own for her anxiety, but without success. She reported that her mother was not supportive enough to her and her father was involved with many extramarital affairs. She appeared intelligent, well dressed, and somewhat masculine. During the initial interview she tried to avoid crying, without success. On Axis I she had both major depression and social phobia; on Axis II she had avoidant personality disorder. Her defensive functioning was characterized chiefly by minor image-distorting (narcissistic) defenses, with some obsessional and hysterical defenses.In the initial session the therapist first tried to help Miss A. express her feelings, but did so by using too supportive an approach (associations, support strategies, and reflections). The therapist did not interpret the conflict between her wish to be able to control her emotions independently and her antithetical wish to be helped passively without having to ask for it. Not addressing her wish to remain in control and instead relying on a highly supportive approach actually triggered a counterdependent reaction. The alliance remained at a low level, and there was a small decrease in distress reflected in the GSI score of the SCL-90-R (as shown in Table 2).
View this table:
[in this window]
[in a new window]
TABLE 2. Scores at the initial session for three cases and subsequent alliance pattern and change in SCL-90 scores
Case 417. Miss B. was a 20-year-old student in sociology who complained of conflicts with her boyfriend. She blamed her mother for being incapable of keeping the patient's father from leaving the family and for subsequently becoming somewhat exclusive in her relationship to her second husband. Miss B. presented with depressed mood, which did not mask her intelligence and subtlety. Although potentially attractive, she appeared a bit untidy. On Axis I she suffered from dysthymia, panic disorder without agoraphobia, and specific phobia (claustrophobia); she also had depressive personality disorder. Her defensive functioning was quite similar to Miss A.'s in case 416, with predominantly minor image-distorting (narcissistic) defenses and some obsessional and hysterical defenses.In the initial session the therapist often confronted the patient, interpreting to her how she tried to maintain her independence in order to avoid her fear of being rejected whenever she adopted a more passive role. The patient agreed and made an effort to understand her internal conflicts better throughout the sessions. The moderately high degree of questioning, clarification, and interpretation of her conflicts was well adjusted to her level of defensive functioning. The patient's alliance continuously improved over the four sessions, and there was a large decrease in distress (see Table 2).
Case 419. Miss C. was a 25-year-old secretary. She presented with strong feelings of jealousy toward her sister, who was pregnant, and with depressed feelings following a series of unsuccessful affairs. She was not satisfied at her job and described some overly ambitious professional projects. She reported she had been treated poorly by both her mother and father, but had only recently recognized this. She was tall and dressed without any attempt to appear attractive. Her affect was depressed, and she cried a lot initially. The therapist felt she acted as if she were much younger than her stated age. On Axis I she reported having a major depressive episode and specific phobia, while on Axis II she had both borderline and depressive personality disorders.Her defensive functioning was characterized heavily by lower-level action and disavowal defenses. The therapist reflected her depressed feelings back to her, offering supportive associations and suggestions, as well as linking her distress to her recent attempts to confront the painful memories that she had previously long denied in order to protect her self-esteem. This highly supportive approach communicated a sense of empathy for the patient's distress while allowing the therapist to maintain an interpretive-exploratory attitude that was well adjusted to the patient's relatively low level of defensive functioning. The alliance continuously improved across the sessions, and there was a large decrease in distress (see Table 2).
Discussion of Clinical Examples:
As Table 2 shows, the patients in two of the cases (416 and 417) had similar levels of defensive functioning but differed in the level of interventions (ESIL) offered by the therapist, thus yielding different Adjustment Ratios. On the one hand, there was little difference in the proportions of transference and defense interpretations (see Table 3). But the therapist used more questions, clarifications, and work-enhancing strategies and correlatively fewer reflections, support strategies, contractual arrangements, and associations with case 417 than with case 416. Overall, the general approach with the patient in case 417 was more exploratory and dynamic and less supportive than in case 416. Given that the two patients had similar defensive functioning (as shown in Table 4), the higher exploratory-interpretive mix in case 417 led to a growing alliance. The patient in the third case (419) had a markedly lower level of defensive functioning, but the therapist offered a much more supportive level of interventions (lower ESIL), which yielded a good Adjustment Ratio. The patient's heavy reliance on action defenses would have been very challenging; the therapist still offered some interpretation and exploration, but heavily buffered by a high proportion of supportive interventions similar to those offered in case 416. This mix led to an improved alliance (Profile II). Overall, the patients in cases 417 and 419, who differed widely in defensive functioning but had similar Adjustment Ratios (around 1), both developed an improved alliance (Profile II) along with a substantial drop in SCL-90-R scores.
|
|
|
DISCUSSION |
|---|
|
TopABSTRACTSINTRODUCTIONMETHODSRESULTSCLINICAL EXAMPLESDISCUSSIONREFERENCES |
|---|
Another limitation is that the design was naturalistic and observational. Patients were not randomized to different levels of interventions at given levels of defensive functioning. In practice, such an experiment might be quite hard to construct and carry out without adversely distorting the treatment. Nevertheless, the interpretation of our findings requires caution, in that other intervening factors may have influenced the results. In viewing this limitation from a different perspective, our longitudinal design considers each patient's four-session evaluation as a naturalistic experiment. The results, when theoretically predicted, have some confirmatory value for the overall theory that therapists' interventions must be adapted to the patient's level of functioning to produce a desirable effect. Naturalistic experiments serve similar confirmatory roles in any area of science.
In this study, patients' alliances differentiated very rapidly over the four-session Brief Psychodynamic Investigation into three patterns. The finding that the alliance patterns were differentiated by the third session concurs with the observation by O'Malley et al.6 that alliance was formed by the third session in short-term psychotherapy. A high, stable alliance (Profile I) would be considered the most desirable, reflecting a therapy that starts with a good therapeutic alliance and stays on course, presumably to yield whatever positive outcomes are associated with a good alliance. However, to improve the way BPI is conducted, it is more informative to understand the factors predicting the differentiation of the other two alliance profiles. Patients with a low alliance at the first session diverged by the third session into either those with an improving alliance (Profile II) or those with a continuing low alliance (Profile III). To date we have not detected two other potentially problematic profiles: a high initial alliance followed by deterioration, and a vacillating or unstable alliance. The latter may show up only in treatments of greater length. The remaining analyses, then, addressed the factors that differentiated the two profiles that diverged from an initial low alliance.
The patient's initial overall level of defensive functioning alone did not differentiate these alliance profiles, suggesting that even patients with low initial ODF can develop a good alliance. Defense mechanisms can be considered an aspect of the subject's personality organization, and it is widely believed that poorer personality organization is associated with more difficulty establishing a positive alliance. However, our results suggest that the defensive functioning level per se is probably less significant for alliance formation than how the therapist chooses to intervene given the patient's presenting level of defensive functioning.
We found an association between Expressive-Supportive Intervention Level and patient alliance at the first session only. In the first session, a relatively higher proportion of exploratory interventions was associated with high alliance and a relatively higher proportion of supportive interventions with lower alliance. This may reflect the therapist's general style or his or her initial judgment as to what mix of interventions will be best for the patient. However, this initial mix of interventions considered alone did not predict subsequent alliance-building. This result is in line with a conclusion in the review by Orlinsky and Howard.25 To determine the effects of psychotherapists' techniques for any model of psychotherapy requires examining an intervention in the context of how it is adapted to a specific patient—that is, in relationship to specific patient characteristics.
To be supportive, in general or in response to a patient's level of defensive functioning, may be necessary, but supportiveness alone is not sufficient to contribute to alliance building. Rather, at each level of a patient's defensive functioning there appears to be some specific range of more exploratory interventions that will be optimal to facilitate growth of the alliance. This pattern was exemplified in our findings. The patients in the Profile II (improving) and III (low, stable) groups received about the same proportions of exploratory interventions, but because of differences in defensive functioning, the Adjustment Ratio of Profile II was far closer to that of Profile I (high, stable) than to the significantly lower AR for Profile III. Thus, groups I and II received more optimal proportions of exploration for their given levels of defensive functioning, leading to the development of equally good alliances. By contrast, the Profile III group received relatively fewer exploratory interventions, despite a slightly higher given level of defensive functioning, resulting in no improvement in alliance. These findings support our hypothesis that development of a strong alliance requires that the therapeutic interventions be adapted to the patient's defenses.
Together, our empirical findings are consistent, in part, with Winnicott's26 warning that supportive techniques are no substitute for interpretive techniques, even with regressed patients:
The adaptive technique that must meet a patient's regression is often classified (wrongly, I am sure) as reassurance. We assume that reassurance is not part of the psycho-analytical technique. The patient comes into the analytic setting and goes out of it, and within that setting there is no more than interpretation, correct and penetrating and well-timed.
Most clinicians and researchers do consider supportive interventions (of which reassurance is one element) to be important therapeutic techniques. However, supportive interventions alone do not appear to have the power to improve the alliance, which a more optimal mixture of supportive and exploratory interventions does have. Furthermore, the present research suggests that at higher levels of defensive functioning, a proportional increase in exploratory interventions by the therapist will yield better alliance development.
Apart from its influence on the alliance, Kernberg14 has described the interpretation of lower-level defenses, such as splitting, as a major tool of supportive-expressive psychotherapy from a clinical point of view. Support has to be understood in psychoanalytic psychotherapy as aiming to support ego functioning, which includes defensive functioning, in order to minimize regression and reinforce ego boundaries.
We also expected that a low alliance might be associated with too high a proportion of exploratory interventions for a given patient, yielding too high an AR in the Profile II group (low and stable). We did not find this. It is possible that the extensive training of our participating therapists reduced the likelihood of the overzealous use of exploration. A larger sample might yet yield such examples, and other studies of BPI or psychotherapy may find such cases. It is also possible that extensive interpretation occurs only when the patient has given enough material to the therapist, which may require a longer treatment time frame or an already-developed good alliance.
The training of clinicians to conduct psychotherapy (or a dynamic investigation) may differ greatly depending on whether one views the development of a good alliance as an explicit, primary goal, or as a secondary indicator of the state of the therapeutic process. The latter would reflect how well the therapist has adjusted the level of supportive-expressive interventions to the patient's level of defensive functioning or to some other salient patient characteristic. Further confirmation of the present results might suggest that training manuals should give greater consideration to the latter position, that of adjusting intervention levels to fit the patient.
Although the therapeutic alliance is a conception common to many forms of psychotherapy, our operationalization of adjustment is specifically psychodynamic. Nonetheless, this concept could be applied to any type of patient or treatment. It would also be instructive to examine other treatment models to determine if they employ a similar concept of adapting the treatment to a given patient—whether implicitly as general advice, or explicitly in relation to a measured patient characteristic. Are there any treatment models in which adjustment of interpretations to the patient's defensive level is not an important predictor? Perhaps our psychodynamic version of adjustment is a special case of a more general theory of adaptation of technique that operates across treatment models. What initially appears to us as a finding specifically important for psychodynamic treatment may have some aspects attributable to so-called common factors across treatments. This should be studied. As Lambert and colleagues27 have said, "Those factors that are common to most therapies should not be viewed as theoretically inert nor as trivial." We would add that even though common factors have an important role to play, it is possible that only specific therapy models can produce or generate them in an effective way.
|
ACKNOWLEDGEMENTS |
|---|
|
REFERENCES |
|---|
|
TopABSTRACTSINTRODUCTIONMETHODSRESULTSCLINICAL EXAMPLESDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  C. A. Roy, C. J. Perry, L. Luborsky, and E. Banon Changes in Defensive Functioning in Completed Psychoanalyses: the Penn Psychoanalytic Treatment Collection J Am Psychoanal Assoc, April 1, 2009; 57(2): 399 - 415. [Abstract] [PDF]
|
|
|
|
 |
|
 M. Drapeau Research on the Processes Involved in Treating Sexual Offenders Sexual Abuse: A Journal of Research and Treatment, April 1, 2005; 17(2): 117 - 125. [Abstract] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ALL ISSUES | SEARCH | TABLE OF CONTENTS |
