Differential Effects of Interventions on the Therapeutic Alliance With Patients With Personality Disorders

J Psychother Pract Res. 1998 Fall; 7(4): 301–318.

PMCID: PMC3330513

Differential Effects of Interventions on the Therapeutic Alliance With Patients With Personality Disorders

Michael Bond, M.D., Elisabeth Banon, M.D., and Marcella Grenier, Ph.D.

Received March 13, 1998; revised April 21, 1998; accepted April 27, 1998. From the Departments of Psychiatry at McGill University and The Sir Mortimer B. Davis-Jewish General Hospital, Montréal, Québec, Canada. Address correspondence to Dr. Bond, Institute of Community and Family Psychiatry, 4333 C"te Ste-Catherine Road, Montréal, Québec, H3T 1E4, Canada.

This article has been cited by other articles in PMC.

Abstract

The goal of this study was to examine the relationship between clearly defined therapist interventions and the therapeutic alliance with personality-disordered patients. Transcripts of one psychotherapy session for each of 5 subjects taking part in a long-term psychotherapy research project were rated for therapist interventions and therapeutic alliance to determine if specific interventions were followed by enhanced or diminished therapeutic work. Transference interpretations were followed by a deterioration in the therapeutic alliance when the alliance was weak, but by enhanced work when the alliance was solid. In patients with both strong and weak alliances, defense interpretations and supportive interventions enhanced therapeutic work without increasing defensiveness. Supportive interventions seemed to prepare the way for exploration and to repair ruptured alliances. (The Journal of Psychotherapy Practice and Research 1998; 7:301–318)

In a review of research on the role of the alliance in psychotherapy, Horvath and Luborsky¹ caution that “research is unlikely to provide guidance to clinical practice unless the relations between clearly defined therapist actions in specific contexts and the effect of these interventions on process or outcome can be demonstrated” (p. 568).

The goal of this article is to further our understanding of the effects of therapist interventions on therapeutic alliance and outcome in various contexts of psychotherapy, especially with patients with borderline or narcissistic personality disorders. Transference interpretations, defense interpretations, and supportive interventions can enhance or diminish the working alliance depending on such factors as the patient's ego strength, the state of the patient's readiness for self-exploration and elaboration, the current state of the alliance, the phase of the therapy, and the timing within a session.

The issue of establishing and maintaining a therapeutic alliance, especially with borderline and narcissistic patients, has been a focus of study for many authors.²^–⁹ Empirical studies have ranged from Gabbard and colleagues'¹⁰ intensive examination of the effects of transference interpretation on alliance in 1 and 3 patients in analytic therapy to Piper and colleagues'¹¹ investigation of the relationships among transference interpretations, therapeutic alliance, and outcome in 64 patients in “short-term” psychotherapy.

The general consensus in the literature has been that with more severely personality-disordered patients, transference interpretations are very high risk interventions.² Kernberg,⁷ however, maintains that early interpretative work with the negative transference is necessary to prevent premature termination and therapeutic stalemates. Although Kernberg is clear about his recommendations for intervening with patients who are angrily attacking or manipulating the frame, he does not address the impact of transference interpretations that focus mainly on the patient's anxiety.

Piper et al.¹¹ found an inverse relationship between the proportion of transference interpretations (TI) and both therapeutic alliance and favorable therapy outcome for patients with a high quality of object relations. It is possible, however, that there was an optimal level of TIs and the results were skewed by the cases where the number of TIs was excessive. The therapies of these 64 subjects were about 20 sessions long. Buckley et al.¹² found that outcome of psychotherapy was significantly better when therapists rated their interventions as more supportive than interpretive. The subjects were a group of 21 medical students whose therapy lasted a mean of 43 ± 33.4 sessions over 8.8 ± 3.0 months. Thus, support seems to be important for a wide range of patients, not just for those diagnosed with severe personality disorders. In addition, transference interpretations may be disruptive for healthier patients, not only for those in the borderline range, in certain contexts.

Recent studies have shown that the factor most predictive of positive outcome for both long-term and short-term therapies is the quality of the therapeutic alliance.^2,10^,¹³^,¹⁴ Neither the exploratory nor the supportive aspect of therapies was correlated to positive outcome when examined alone; instead, outcome appeared to be related to the quality of the alliance. When the alliance is strong, both exploratory and supportive interventions are more effective. The maintenance and caretaking of the therapeutic alliance seems therefore to be the priority, especially in the initial phase, of any therapy with all patients. It has further been shown that the quality of the alliance is often apparent and is established early in the treatment.¹⁵ Longitudinal analyses of the levels of the alliance in psychotherapy appear to confirm a rupture–repair cycle in successful therapies.¹⁶

Transference interpretations seem to be linked to the most significant movements in the alliance, both toward its strengthening and toward its deterioration. Consensus in recent studies indicates that interpretive work generally, and work in the transference specifically, benefits those patients who are capable of forming good-enough alliances and who possess higher levels of ego strength. Patients who are more fragile and labile, however, particularly those diagnosed as borderline, have greater difficulty establishing an alliance and maintaining it during periods of intense anxiety. With these patients, interventions that are exploratory and transference-based may lead to deterioration in the already fragile alliance and do not facilitate further work. Consequently, more supportive techniques have been recommended for these more fragile and labile patients, especially during the early phases of therapy until a firm alliance has been established.¹⁰^,¹⁷

Gabbard and his colleagues¹⁰ have further challenged the dichotomy between exploratory and supportive interventions, preferring to view these two approaches as working in tandem rather than being polarized opposites. They point out that effective interpretations often follow a series of supportive interventions that pave the way for interpretive work.

In our study, we examine the relationships among therapeutic alliance, interventions, and in-session outcome during the initial phase of four open-ended psychotherapies and the middle phase of a fifth. Our hypothesis is that the stronger the alliance, the more the patient can tolerate and benefit from transference interpretation. On the other hand, fragile patients will tend to form fragile alliances that can be shaken by early transference interpretation, and supportive interventions will be crucial for these latter patients. In addition, we hypothesize that empathically moving back and forth between supportive and interpretive interventions is crucial to maintaining a good working alliance, which allows therapy to proceed with patients of all levels of ego strength. We examine transcripts to explore therapeutic interventions and their associated in-session consequences, especially for the state of the alliance.

Methods

Subjects were chosen from an ongoing long-term individual psychotherapy project at the Sir Mortimer B. Davis-Jewish General Hospital in Montreal, Canada. At the time of selection, the sample consisted of 40 subjects and was still growing. We selected subjects whose session transcripts illustrated a range of interventions and in-session changes and who had personality disorders. All therapy sessions were audiotaped. Selection criteria included age over 18 years; a diagnosis of mood disorder, anxiety disorder, and/or personality disorder; and absence of organicity, substance abuse, or psychosis. All patients gave informed consent to participate in the project and to have their sessions audiotaped.

Transcripts of sessions were used to score all therapist interventions according to the Psychodynamic Interventions Rating Scale (PIRS; S. Cooper, M. Bond, unpublished manuscript, University of California–San Francisco, 1992) by raters trained in the use of the scale. Consensus ratings were used to arrive at the final categorization of interventions. (See Appendix A, p. 318, for definitions.) Construct validity for this scale was shown in a sequential analysis of therapist interventions and patient elaboration in a study by C. Milbrath et al. (unpublished manuscript, 1997). Diagnostic information from the Structured Clinical Interview for DSM-III-R (SCID II)¹⁸ and therapeutic alliance scores from the California Psychotherapy Alliance Scale (CALPAS)¹⁹ were available for all subjects.

We examined selected transcripts of therapy sessions to look for “turning points” in the sessions. Four of the sessions were from the early phase of therapy (sessions 2–6), and one session was the 69th session of a therapy. Turning points are indicated by a change in the therapeutic alliance, changes in the degree of emotionally significant elaboration,²⁰ or changes in the level of defensiveness of the patient. Each of the three investigators independently noted turning points in the transcripts. When all three investigators agreed, those segments were selected. Turning points tended to occur when there was either a shift to more interpretive interventions or a shift away from interpretation; this permitted a high level of agreement. We used these selected segments with turning points to investigate the complex interactions among interventions, alliance, and ego strength or diagnosis. We then reached conclusions based on a qualitative rather than a quantitative analysis of the segments.

Although we have relied heavily on the written transcripts of sessions to illustrate aspects of the alliance, certain authors have suggested that the critical elements in forming the alliance seem to come not from what is said but more from felt emotional involvement of the therapist and patient.¹⁵ Examination of certain paralinguistic devices indicative of such involvement shows that in strong alliances, therapists use less reflection when the patient brings in factual material, generally respond neutrally to patient irritation, and engage and empathize more obviously with patient tension. Further, therapists who are engaged in strong alliances frequently shorten their silences and increase the amount of validation and affirmation during periods of intense anxiety and agitation.¹⁰^,¹⁵ A high level of therapist talk may be associated with support and containment, particularly with fragile patients who need frequent reassurance. Some of these subtle but important interpersonal processes may not be assessed accurately by using our transcript methodology.

The following excerpts are offered to illustrate the impact of interpretive interventions, the use of support techniques to bolster and repair damaged alliances, and their differential effects on the alliances with fragile, intermediate, and neurotic patients. The therapist interventions are coded at the left margin of the transcript.

The codes are as follows (see Appendix A for definitions): defense interpretation (DI), transference interpretation (TI), acknowledgment (A), clarification (Cl), question (Q), therapist association (Ass), reflection (R), work-enhancing strategy (WES), support strategy (SS), and contractual arrangement (CA). It should be noted that transference interpretations are considered to be any comment on the patient–therapist relationship, even if no link to past relationships is included.

Clinical Data and Interpretation of Results

Patient A.

Ms. A. is a fragile borderline patient with a weak alliance (total CALPAS score = 4.0; this project's sample mean = 5.58; median = 5.67). The CALPAS component indicating patient commitment was particularly low (2.3).

Excerpt 1:

This excerpt occurs very early in the first session of therapy after a brief period of superficial talk during which Ms. A., a female patient, asks the therapist many questions.

WES

Th So, for this therapy, basically you shape the session by talking about things that are on your mind rather than me asking you a lot of questions.

How do you feel about that?

Pt.: I don't know. I was thinking this morning, well, what am I going to ask, what . . . (pause) Um. I don't know, I guess just one big issue with me is, um, I don't know if it is a fear of being alone or I don't know what it is, but I don't—I don't expect very much from people. Like I am thinking specifically my boyfriend. Uh, he doesn't treat me very well, and I don't have the strength to break up with him. And I mean, it's so—it affects so much of my life. It's not just like he's part of me, it affects everything and it's just different. . . .

Th.: Okay, can you elaborate on what you mean by that, when you say it affects so much of your life?

Here the therapist has tentatively introduced the therapeutic dyad in what seem like what Gaston and Ring²¹ have termed “work-enhancing strategies.” However, the mere mention of patient and therapist in one sentence may provoke anxiety in a borderline patient. Being linked as a team with the therapist is followed by associations of wanting to break up with her boyfriend and how the relationship affects everything. The patient moves immediately away, and the therapist follows her. However, later in the session, after she talks about her mistrust and fear of men, he again intervenes:

Pt.: I'm afraid of men. (pause) I'm very anti-men.

Th.: I'm a man.

Pt.: Right! (laughs) You're different. Just the topic of men makes me angry and scared at the same time. My grandfather molested me, my father beat me, boyfriends have beat me, one left me for dead, got me into prostitution, got me into drugs. Another one had a relationship with another girlfriend for two years while I was with him.

Again the therapeutic couple has moved out of the transference as the therapist encouraged the patient to talk about her history with men. Finally, toward the end of the session, he intervenes again and redirects her toward the transference:

Th.: So I wonder if you have some anxieties about me?

Pt.: Yes, but no. You're sitting far away . . . (laughs) like I wouldn't sit like a mute, you know. And I just think it was you that asked me at the beginning do I have a preference, man or woman, and I am not comfortable with women, either, so I don't want to make my fear worse. You know, like avoiding men would just make it harder for me to not be afraid, so I am not going to generalize. It's got to start somewhere. Besides, you're different. (laughs)

Th.: What do you mean by that?

Pt.: I don't know, I mean, I'm not uncomfortable, 'cause sometimes I want to talk, but it just—I just remember that, well, at nine o'clock I'll walk out of here and it's just—it's not like a friend that really cares about me and, you know, this is your job. But at the same time I feel safe here and you don't intimidate me. You don't scare me, so. . . . I don't know, I just—I'm not comfortable, but I am, you know what I mean?

Th.: The boundaries are fairly clear here.

Pt.: Yeah.

Th.: And that makes you feel more comfortable.

Pt.: Uh-huh.

Th What makes you feel somewhat uncomfortable is wondering how much I will really care about you.

Pt Yeah.

Th That, you know, actually at ten to nine the session ends and then you wonder, you might leave feeling, like, does he really care about me?

Pt Yeah, I mean, this is your job and, like, I just think sometimes—and I can't believe I'm saying this; I've never said this to any therapist I've ever had—just that this is what you do for a living and I don't know how many years you've been doing this and you listen to people's stories all day, every day. And, like, I would imagine at some point you just get like tired of hearing people's stories and like I just wonder, like, does he think I'm feeling sorry for myself or is she pathetic or, oh my God, is she sick, or you know what I mean? I just wonder what you're really thinking and are you ever bored? Are you ever, like, just tuning me out? But I just—I just hope that you really love what you do and you really care and I'll get the most out of this.

Th.:

Well, those concerns make a lot of sense.

Although this intervention appears to have been accepted and tolerated by the patient, immediately after there occurred an “acting in” that left both patient and therapist feeling bad. The patient continued by talking about a friend of hers who had committed suicide and the intense feelings of guilt and terror she had felt, which had led to her abandoning her friend just before the suicide. This occurred just as the session was ending, leaving the therapist feeling helplessly entrapped in a situation where he also had to “abandon” the patient as she was feeling bad and the patient feeling perhaps that she had been opened up only to be abruptly cast aside. This type of projective identification is typical of a borderline patient. This particular patient returned for only one more session before she terminated treatment.

Excerpt 2:

This excerpt occurs during the second session with Ms. A., which followed a 7-week delay and a no-show after the first session. It occurs about one-third of the way into the session after a gentle exploration of the patient's feeling depressed and victimized by her surrounding environment and unable to count on anyone.

Pt . . . I went back, and, I don't know, I—I try to talk to the therapist [a different one] because, like I told him afterwards, I feel, most of the time after I leave, more upset than before I come to the group, because I talk, I share my input, and then it just brings things up and it just goes nowhere. I mean he says nothing. I said, “Well, do you have any suggestions, like. . . . ” “No,” he says, “nothing.” He says, “Well, you are doing a lot, don't you think?” So, what, I should just [do] things like. . . . (pause)

Th So you like more suggestions.

Pt Yeah.

Th You probably find, say, me frustrating?

Pt Yeah.

Th Because I don't make any suggestions? It sounds like you need very, um—a clear indication of somebody caring about you by getting more actively involved and, you know, the type of therapy that I am doing with you is an analytic type of therapy, where I try to help you understand things at the time. Most of the time that doesn't feel to you like I care, like I am not actually doing anything.

Pt Mmm . . . (pause) I don't know. I'm just—I'm confused.

Ass

Th I think in the short term it might be gratifying to receive suggestions from people, but suggestions are a double-edged sword, as I am sure you are aware. They can be good or you can feel controlled—or you are not being given good advice, or somebody isn't taking the whole picture into account.

Pt Some of the suggestions I get . . . it's just like (pause) I don't know, like the individual therapist I sometimes see at . . . just says things that leave me thinking a lot, you know. I wouldn't want suggestions, I guess because, like you said, it's too controlling and I don't want—I would probably take it as being told what to do. But, like, somebody said something last night in the group that left me thinking, and I guess that is why I keep going back. Um, (pause) they were talking about anger. (begins to cry) I don't know why I am crying. . . . Just saying that . . . um, something to the effect that it is easier to be angry than sad.

Th Uh-huh.

Pt And it makes a lot of sense because, I mean, I have a lot (pause) I know I have a lot to deal with and, um, I have often found myself very angry and never really understood it. Like I will just be like, oh my God and be so angry all the time. I'm just—I'm like consumed with anger. And, um, she was just saying that it's a lot harder to get in touch with sadness and (pause) I identify with that. I. . . . (crying; pause)

Th Well, sadness can be frightening because you could feel helpless. When you're angry, you feel you are stronger; and helplessness, while it is still not far away, generally you feel a little more active—a little stronger.

Pt.:

I don't know how to deal with sadness. Like sometimes I think, like, my God, I have so much to be sad about. Then I just feel like I am feeling sorry for myself or it's just that I'm so pathetic, like everybody has [unclear], just deal with it and move on, but (pause) I don't know, like I just wonder am I really dealing with anything. I am at school full time, all year round, even in the summer. I am working two jobs. I just—like I have no time. (pause; continues to cry)

Following this, the therapist's interventions become increasingly active and supportive, helping the patient in a more instructive way by offering comments on how to manage her life. The therapist moves away from the transference and any further exploratory work in an effort to shore up the patient, who is in danger of breaking down.

Th Well, I think that the activity does keep you from getting too sad. But you're wondering if in some way you're not dealing with the roots of what's making you sad.

Pt Yeah.

Th And you're trying to figure out what is an appropriate extent to allow yourself to stay sad, when to move on, when to count your blessings, when to accept things, when to fight things, and I think people struggle with that all their lives. And certain individuals develop their own style.

Ass

And nobody can dictate what is the best way or the right way for anybody to find the balance of having certain expectations. You know, the more you have expectations, the more you can be disappointed and hurt, sad, angry. But if you lower your expectations too much you feel an emptiness because you sort of expect nothing from anybody, and then you're sort of like dragging yourself through life, saying, “Well, no, I don't expect anything,” and not getting anything, and it sort of is a drag. So I guess every now and then you allow yourself to have some expectations, take some risks, and sometimes it works out and sometimes it doesn't.

Pt.:

Speaking of expectations, I've got two of my three marks for the semester, and I guess they're pretty good, but. . . .

These two excerpts illustrate certain characteristics that develop when the patient is fragile, the alliance is not solid, and the transference is addressed in an exploratory fashion. Both excerpts show how the focus on the transference is followed by a poorly modulated intense affective state in the patient.

In the first illustration, the patient stayed with an exploration of the transference relationship, but this appeared to provoke memories of terror and guilt as well as sadness, followed by a traumatic end to the session that probably damaged the alliance.

In the second vignette, the same patient reacts with confusion to the therapist's intervention, and her wall of defensive anger crumbles very quickly. She becomes overwhelmed by her sadness, and the remainder of the session is spent trying to shore her up by moving away from the transference. The therapist's quick, supportive interventions might illustrate that he too was surprised by how rapidly she became overwhelmed. His attempts to repair the rupture are not successful enough to keep the patient from discontinuing therapy.

This patient was not able to use the transference interpretation to elaborate and explore it further and to relate it to other experiences of significant relationship. In fact, the abrupt turning away from any mention of the transference probably indicates how frightening this topic is for her. A borderline patient with fragile defenses and poorly defined ego boundaries can feel very threatened by discussion of the here-and-now relationship with the therapist. The intense anxiety does not allow for abstract connections to be made. The patient cannot experience the relationship as an “as-if” model for exploring other relationships. Instead, it becomes “real” and threatening. The patient responded well to gentle questioning and exploration of relationships outside of the therapy.

Patient B.

Ms. B. is a borderline personality disorder patient (CALPAS = 4.96; mean = 5.58, median = 5.67).

This case illustrates how the strength of the therapeutic alliance changes depending on the match between the type of therapist intervention and the readiness of the patient to respond to the intervention. Three excerpts from the same session, early on in the therapy with a young borderline patient, demonstrate that empathic defense interpretations can enhance the alliance but transference interpretations may provoke resistance.

Excerpt 1:

Pt Uh, (laughs) I don't know. This always happens. I don't know what to say. Um, everything's been okay, though. I haven't had any problems and, you know. . . . On the way here I was thinking about, like, what to talk about and I (laughs) I always have the same problem, I don't know what to talk about.

Th Uh-huh, did that worry you?

Pt Yeah, very much. Very, very much. It—it really—it—it bothers me, 'cause I—like, I don't—you know, I feel like, you know, I don't want to waste, you know, her time and I don't want to waste my time, but, like, I don't know what to talk about. Like, I don't know. I want to concentrate—I don't know. I want to try and, like, uh, concentrate on, uh, where the problem might be and I feel that it's not in what happens every day.

Th Uh-huh.

Pt You know what I mean? Like, I feel that, uh, my problems started a long time ago and now I feel like—I feel like I'm okay now, you know. Basically it's just, I want to know, you know, what happened or whatever. I don't know. 'Cause it seems like, you know, everything's fine now.

Th Uh-huh.

Pt I don't know.

WES

Th I suppose what I told you, though, that, um, just that—that it's okay. Um, I don't expect you to always have all kinds of new things, we would go on. . . .

Pt Well, what is that we should con—like what is it that we're—like, what are we trying to do here? Like I don't understand. Like, what—by me talking, what does that do?

Th You're not sure . . .

Pt No.

Th . . . it could help?

Pt No, I—no, I want to know what is—like, what is—what does it do? It's just to, like, to get all my feelings out in the open? Is that what the—is that what it is?

WES

Th Well, that's one part, yes, for sure. To help me understand you better.

Pt.:

Okay.

Here, the therapist has handled the ambivalence (resistance) toward treatment, as manifested by lateness to the session, and the fragile alliance, as illustrated by the patient's uncertainty about the usefulness of therapy, by taking a more active and supportive stance. She is instructive as to the workings of psychotherapy, using work-enhancing strategies. This maintains the status quo, but we see improvement in the therapeutic alliance and the elaboration of further material on the part of the patient only when the therapist empathizes with the patient's anxiety and defensiveness with the DI in the next sequence.

Ass

Th I feel I sort of know you. You've told me quite a few things, but I'm sure I don't know you deeply . . .

Pt Right.

Th . . . at this point. And I can understand it's sort of difficult to come for a session and just start opening up. You say you don't know where to start.

Pt Yeah, you don't know where to start.

Th Uh-huh.

Pt Like, I know I haven't told you a lot—I've told you a lot about, like, the history of my family, like my relationship with them, but I don't think I've told you much about, like, me, I think. I don't know. Like I told you about me now, but the me then, it's like, I don't know. I don't know if we've talked about it. (laughs)

Th Uh-huh.

Pt.:

That's the—that's what I'm trying to say. Like, the problem with me has always been, like, my attitude, I guess you would say. Like ever since I was old enough to—you know, to be an individual person from my parents, everybody always complained that, you know, “You're so—you have this attitude that pushes people away.” And I was able to be rude to my parents and I was able to—to do things, you know, like, that—that other people weren't doing. I was always the bad—you know, the bad person when it came to school and when it came to boys and everything. I was always, like, more advanced. But not in a good way, you know.

However, when the therapist does not help the patient to continue elaborating, but instead stays with giving guidelines about how therapy works, the alliance deteriorates, with the patient repeating, “I don't know” and “Okay, but. . . . ” This illustrates how the gains from an empathic interpretation can be reversed when the therapist resorts to empty explanations rather than continuing in the interpretive mode that seems well received.

Th Uh-huh.

Pt So, it's like I want, like, I want to—like, I want to know why I was like that, you know. Like I told you that already.

Th Uh-huh, and. . . .

Pt And I don't know how to—I don't know how to figure that out.

WES

Th Uh-huh, and my, uh, what I offered is that, um, if you can tell me everything, as it comes to your mind, um, I'll try and. . . .

Pt All the things that I've done?

WES

Th Well, anything you're thinking about when you're here with [unclear]. . . .

Pt Well, you see, I—I think about those things. I don't know if, like, that's one of the things that, like, I—I know I'm supposed to tell you everything, but am I supposed to tell you about all the things that I've done? I don't know.

WES

Th As it comes to your mind and as you are telling me things, I'll, if I can figure out something, I'll—I'll tell you as it comes to me.

Pt.:

Okay, but. . . .

Excerpt 2:

Here, an attempt to interpret the potential for a negative transference is met with conscious denial on behalf of the patient. The therapist senses the patient's reticence about exploring her ambivalence toward her at this time and moves away from the transference toward a more supportive and empathic interpretation of the patient's harsh superego. The patient settles after this shift and moves toward material about less threatening objects (her boyfriend).

Pt . . . You know, it makes me feel, like, disgusting, and, you know, how could someone else feel comfortable with me.

Th Uh-huh. In that sense I think your—this fear about what you're going to find, um, you won't be able to avoid this fear when you're here with me. Slowly, but gradually, you might be able to overcome it, but, like, as we're starting now, we're going to have to find ways of dealing with this fear. On one hand you want to find out, but on the other hand you're scared, so we have to take that into account.

Pt But—but here I'm not scared. When I'm here, I'm not scared, because, I mean, this is why I'm here, you know, is to find out. I mean, it's when I—it's when I talk to someone that I know will judge me that I get scared, you know. I know that you're not here to judge me. I know you're not here to tell me, “You're bad. You're good. That's disgusting. That was wrong. How could you do that?” You know? Whereas if I'm sitting with a friend of mine, she could say that and that's what scares me.

Th I see.

Pt You know. I don't—that's not what I need, 'cause that's only going to make me feel worse, you know.

Th Uh-huh.

Pt But there's just so much to talk about that you don't know where to start.

Th I was thinking you seem to judge yourself . . .

Pt Yeah.

Th . . . very harshly at times.

Pt Yes.

Th I suppose there is only so much of that you can take.

Pt It bothers me, and, very honestly, I mean, I look at—I take it day by day and I—I feel that, like—I feel that, honestly, like the last couple of months or, you know, well, say the last eight months or so, I can definitely—I definitely feel that I've become a better person, and I think that it has to do with two things: one, the fact that I'm—I'm—you know, I'm getting a little bit older and I'm starting to realize things, and a lot of it, honestly, has to do with, um, my boyfriend because, um, I—I have to thank him for a lot. He—he—he is a very tough guy and there's a lot of things that he doesn't tolerate, and that's forced me to change. 'Cause the type of person when I—that I was when I met him is the same person that I was, you know, throughout the years. I was easy, I was very flirtatious, I was, uh—you know, when I would go to a club I would dance all over the place, I would dance crazy. I mean, you know, you would go to a club and I'm the type of girl you would find up there on the speaker going crazy. You know, that's the kind of person I—you know, I was when I met him.

Excerpt 3:

A little later on in the same session, a supportive intervention is followed by a transference interpretation. The patient's reaction to this was viewed differently by different raters. One rater saw the patient as elaborating on her frustration toward the therapist in a displaced fashion: “I would have killed . . . my boyfriend.” This rater thought that this allowed her to contain her rage and to maintain a good therapeutic alliance, as manifested by her anticipation of the therapy sessions. The other two raters saw the therapist's comment, “I won't be able to help you figure out answers like that,” as abruptly frustrating the wish for an idealizing transference. They saw the patient's reaction as an outpouring of raw rage that was too much, too early. Thus, once again, a TI led to a threat to the alliance.

Th Uh-huh. Obviously, you've been thinking a lot, trying to figure it out. . .

Pt Yeah.

Th . . . by yourself. I have to think that if it was easy to figure out, you would have figured it out already.

Pt Right.

Th Um, there is something more complex, more. . .

Pt Okay.

Th . . . maybe several elements that—to be put together. Um, in that sense, I think that's where therapy might be frustrating, in the sense that it won't—I won't be able to help you figure out answers just like that.

Pt Right. No, I don't expect it to be easy, but you're right, it is sometimes frustrating.

Th Uh-huh.

Pt You know, but I definitely feel better ever since I—you know, I started. You know, it gives me something to look forward to. I know, like, if, uh, like yesterday I was in just one of these moods. I was so angry, you know, but I don't—I didn't know who I was angry at. You know, I was just—I was in a terrible mood. I woke up in a great mood and as the day went by, I just—I got into one of those moods that everything bothered me, you know, and I didn't understand why. 'Cause, like, ever since school stopped, I haven't found a job and I've been staying home all day, you know, so I don't do much. Like I stay home all day and then at night I'll go out. And I didn't know whether it was the fact that I was just fed up of being home or I didn't know whether it was—I was talking to my boyfriend on the phone and I was—I was literally biting his head off. You know, he's like, “What's wrong with you?” I'm like, “I'm just in a bad mood. Leave me alone.” You know, if, like, words could have killed, you know, I would have killed him. I was being so mean. And when I hung up, like, I hung up on him like three times, you know, 'cause I was just so frustrated. And I was just thinking to myself, why am I taking it out on him? Is it him I'm mad at, is it me I'm mad at? Is it just the world, you know? But somehow I found a way to just—I don't know, I'm just probably having a bad day. You know, I'll figure it out tomorrow. (laughs)

Th Uh-huh.

Pt.:

So, it gives me something to look forward to. Like I know, when everything's going really bad, I know that, you know, it's okay, 'cause, like, you know, I have someone to talk to, you know.

Patient C.

Ms. C. is a dysthymic patient with a narcissistic personality disorder (CALPAS = 6.58; mean = 5.58, median = 5.67).

We can turn now to an excerpt from a patient who seemed to respond in an intermediate fashion to exploratory interventions. By intermediate, we mean that this female patient responded productively to interpretations directed at her defensive behavior. No mention was made of the transference, and it seemed that the therapist was responding to the patient's indirect instructions to avoid transference interpretation.

Th So you made comparisons out loud.

Pt Yes.

Th You told your mother, you compared your mother in her hearing.

I would never scream at my mom, or I would never do what I did in the dream. . . . 'cause she couldn't handle it. She would fall to pieces.

Th Uh-huh.

Pt.:

So, you know, I mean, you know, I have tried to sort of give little hints, because, you know, I think that they should know. I don't have to pounce it on them, but, you know, they should know.

Following this interaction at the beginning of the session, which could be seen as a message from the patient to the therapist as to how to tell her things she should know, the therapist sensitively complies by approaching the patient indirectly. Interestingly, this patient opens up with this approach, and the therapeutic alliance is enhanced:

Th . . . the problem gets so difficult that, uh, that's the only thing you can think of doing, is just leaving it.

Pt Uh-huh.

Th Going away, because the stress can get so great. So which is the want and which is the need? The want. . . .

Pt The want is, you know, I want to be able to—to see my family. I want to be able to. . . .

Th . . . study and to . . .

Pt . . . and to have my life and to be me.

Th Uh-huh.

Pt And, but, I feel that I need him [husband]. You know, whenever it's this thing, you know, this—the two rivals here which are within me, which I came to you with, and you know that I still feel, but to a lesser degree. But it's still there, and I can't say that it's not.

A,Cl

Th Uh-huh. It remains still a—a core problem.

Pt That's right.

Th But the need is for the affection, for the nurturing.

Pt Yes.

Th For the—that close. . . .

Pt The relationship.

Th A close relationship.

Pt A close relationship, yes.

Th And so that need is getting in the way of other wants that you have—wants to pursue your own interests, essentially your own.

Pt.:

Uh-huh.

During this therapy, which at her 2-year follow-up evaluation the patient reported as finding very helpful, the transference relationship was almost entirely avoided. Yet it can be seen that the patient was able to make use of insight and do useful work. More important, the therapeutic alliance appears to have been strong and trusting. Therapist and patient almost complete each other's sentences, building off one another's thoughts to go deeper.

This characteristic of two partners in a collaborative process mutually building on each other's comments to deepen material has been mentioned by Gabbard et al.² and is one obvious and tangible evidence of a good alliance—although not the only one. It was completely absent from the sessions with the previous, more fragile patient, whose ability to collaborate was severely compromised by her fears and mistrust. One could hypothesize that Ms. C. may, in time, have progressed to a more intimate exploration of the transference relationship. However, it is clear that she made good use of defense interpretations and was able to benefit from them.

Next, we include two excerpts from sessions with patients who were able to establish strong alliances and whose psychic structure was healthier at the time of these therapy sessions, although they met borderline criteria several years back. The first case, that of Ms. D., is included here to illustrate how different an alliance looks that has developed over 69 sessions compared with the fifth session. Although Ms. D. began therapy with significant borderline pathology, by this point in her therapy she is able to tolerate transference interpretation quite well. She appears to have a strong alliance that allows for warmth, humor, and a sense of familiarity. The session from the second of these two cases, Ms. E., is an early one, but the alliance is strong and the transference interpretation has an obviously reassuring quality. For both of these patients, transference interpretations were made, but only after extensive supportive interventions to pave the way. Interestingly, the transference was addressed well into the session, which may be important in terms of timing, and both patients were able to elaborate their feelings and thoughts and generalize their insights.

Patient D.

Ms. D. is a borderline personality disorder patient (CALPAS not available; session #69).

Ass

Th Um, going back to something you said a couple of minutes ago, you were being, you were relieved that your father is being supportive?

Pt Uh-huh.

Ass

Th And that he said that he'd much rather support you than pay for hospital bills.

Pt Yes.

Ass

Th And that he would—he saw your decision to quit as a mature . . .

Pt Yes.

Th . . . decision. Um, do you feel that, um, that you're getting the same kind of support here from me?

Pt Yes and no. I mean, I (pause) I mean, I do, um, but the only thing, as I said a week ago, is I did feel a little bit of pressure from you to stick with the job, that this would be good for my self-esteem and self-confidence, and this would shape my personality. . . .

Th I thought I recognized a quote there. . . .

Pt (giggles)

Th You said it earlier.

Pt (giggling) That when I heard you say that two weeks ago, it really put a lot of pressure on me to, you know, really seriously think, OK, if, you know, if I quit this job then I, you know, that shows I'm a failure and it's not helping me grow as a person; but on the other hand, if I stick with the job I'd probably end up in hospital, so. . . .

Th.:

OK, well, actually there we addressed this a little last week, and I did have the feeling, when I heard you say that comment a little while ago, that it was still a live issue, so I'm bringing it up again. Um, you're, I can see why you would feel that way, and you're probably accurate, um, unfortunately, I meant something a little differently, um, then. . . . I didn't want you to quit impulsively.

Later in the session, this topic resumes:

Th That's why I, today, I've been pressing to find out if you know you were unhappy with some of my responses or how you felt when I had said, No, stick it out till next week, don't quit, you know, don't be impulsive, you know, 'cause it might stir up some of those negative feelings that . . .

Pt So, are you saying . . .

Th . . . which is OK.

Pt . . . that I, I am sort of seeing you in the same way that I see my dad, that I'm trying to please both of you.

Th A bit.

Pt Yes.

Th Yes.

Pt Well, certainly, yes, that's true, yes, especially when I said that I, you know, I interpreted not as the way you did, that you felt that I shouldn't quit impulsively, but I interpreted as you thought it would be really good for my self-esteem, that it would really help me grow and, you know . . .

Th You mean stick it out . . .

Pt . . . and solidify my personality. . . .

Patient E.

Ms. E. is a borderline personality disorder patient (CALPAS = 5.5; mean = 5.58, median = 5.67). By session 15, CALPAS moved up to 6.6.

This next example occurs almost at the end of a session with a patient whose alliance is average but grows stronger over the next 10 sessions. She has been exploring the point that she can never express anger directly but can only act it out.

Pt . . . I just can't say no. Why can't I say no?

Th Exactly.

Pt I can't say what I want.

Th I think that's the problem, that you're resorting or you have to resort to acting, doing . . .

Pt Instead of talking.

Th . . . instead of talking. And as long as you can't talk you're stuck with having to act out all the time.

Pt So that's why I lose the weight, that's why I act my actions?

Th Yes. And that's why you have affairs and that's why you do all sorts of things. So what we have to look at is how come you can't put this into words.

Pt I can't.

Th After all, you can do it with me quite easily.

Pt Yes, I can, you see.

Th Uh-huh.

Pt But then again, not that you're a stranger, but you're—you—you won't judge me, do you understand? You won't get mad at me and walk off and start saying da da da. You won't do that.

Th I won't attack you.

Pt.:

That's right. My sister will attack me. A. will definitely attack me. . . . But I can't talk, and I have to learn how to talk.

These two excerpts illustrate well the mutual interaction of therapist and patient in a collaborative process indicative of a strong alliance. Both patients can face their feelings toward the therapist without the danger of their defensive structure crumbling like a dam, flooding them with overwhelming feelings. This enables them to explore the transference relationship, integrate their findings, and generalize to other areas of their lives.

Discussion

Transference Interpretations and the Therapeutic Alliance

In our clinical cases, the impact of transference interpretations was influenced by the state of the alliance, the phase of the therapy, and the timing within the session as well as the psychopathology of the patient. In particular, transference interpretations were well tolerated in the context of a strong, positive alliance and were poorly tolerated in the absence of such an alliance.

Four of our five cases were in the early phase (session 5) of therapy—a time that is crucial for the establishment of a therapeutic alliance. In this early phase, a patient's ambivalence about therapy often becomes evident to the therapist, who is then faced with a dilemma. How can a therapist predict if commenting on a patient's anxieties about the therapy or therapist will enhance the alliance by allowing the patient to air concerns or will threaten the alliance by prematurely challenging the patient's defenses? Will a transference interpretation, which includes empathizing with the patient's doubts and fears, implicitly reassure the patient that these fears are based on past experiences and are not realistic in this relationship? Or will the TI trigger overwhelming feelings with which the patient cannot deal? Our data lead us to conclude that early transference interpretation is risky. Furthermore, addressing negative feelings about the therapist can elicit in borderline patients intolerable affect, which might lead to premature termination.

Different authors offer various conclusions about the effects of defense and transference interpretations on alliance and outcome. Foreman and Marmar²² studied 6 patients treated in time-limited dynamic therapy who had initially poor therapeutic alliance scores: 3 patients went on to have improved alliances and good outcomes, and 3 had unimproved alliances and poor outcome. The therapist actions that most strongly differentiated the two groups and occurred more frequently in the cases with improved alliances and good outcomes were 1) addressing the patient's defenses, 2) addressing the patient's guilt and expectation of punishment, 3) addressing the patient's feelings in relation to the therapist, and 4) linking the problematic feelings in relation to the therapist with the patient's defenses. However, their sample comprised bereaved patients with adjustment disorders or posttraumatic stress disorders, not borderline personality disorders.

Masterson⁸ recommends that even with a brittle and fragile alliance with a borderline patient, confronting the aspects of the transference that are destructive to the alliance can build up the therapeutic alliance. However, in the example that Masterson gave in his 1978 article, the transference work was done well into the therapy—10 months into a thrice-weekly therapy. From our perspective, that is very different from a fifth session when a patient's ambivalence is threatening the viability of the therapy. By Masterson's tenth month, an alliance was formed and a commitment to therapy was evident.

In contrast, Gabbard et al.¹⁰ point out the danger of early transference interpretations and offer tentative conclusions about patient characteristics that might predict a negative impact of transference interpretation on alliance. They list such factors as 1) neuropsychologically based cognitive dysfunction, 2) a history of early trauma, 3) patterns of object relations involving interpersonal distance, 4) masochistic tendencies, and 5) anaclitic rather than introjective psychopathology.

Of course, other factors involved in the therapeutic situation are in play. The warmth and supportive nature of the therapist, the adherence to the frame,²³ countertransference factors, the accurate empathy of the therapist, and the congruence between the patient's expectations of therapy and the therapist's style of interventions²⁴ can all influence the patient's response to interpretation. One cannot measure all of these factors accurately for each intervention, but they most certainly affect the variability of responses.

Our sample's Ms. A. was a young borderline patient with a history of early trauma, masochistic tendencies, and an anaclitic style of relating. Transference interpretations were followed by intense affect and a deterioration of the therapeutic alliance, leading to abrupt termination of therapy. Similarly, in Ms. B.'s therapy, transference interpretations seemed to be very risky. An attempt to elicit the patient's fears about therapy was met with denial, the therapist retreated, and stability was restored. The second transference interpretation dealt with the patient's frustration about therapy and confirmed that the therapist would not be able to figure out answers “just like that.” This intervention was followed by the expression of rage, tempered by reassuring comments about looking forward to therapy. This seems like a high-risk intervention with questionable alliance-building impact. Ms. B. was an 18-year-old borderline patient with a strong tendency to act out. She terminated her therapy about 2 months after this session. The apparent deterioration in the alliance following transference interpretation in the cases of both Ms. A. and Ms. B. would tend to validate Gabbard and colleagues' hypothesis.

This pattern can be contrasted with the active exploration of unexpressed transference feelings by Ms. D.'s therapist, which seemed to enhance an already strong positive alliance. The therapist quickly clarified his real intent after eliciting some negative transference feelings, and the patient seemed to respond favorably to the supportive aspect of his initial intervention, thinking that it would be good for her to stick it out in her job. In this instance, the warmth, the strong alliance, and the supportive nature of the therapist allowed the transference interpretation to lead to useful elaboration. Also, the later stage of the therapy (session 69) created a more secure context.

In the same vein, Ms. E. had a good therapeutic alliance, and when the therapist moved from defense interpretations to the transference, it was to highlight the safety of their relationship (“you can do it with me quite easily,” and “I won't attack you”). This is different from commenting on negative feelings in a therapeutic relationship that is tenuous. The interpretation was also made very late in the session, after several defense interpretations and supportive comments had paved the way.

Clearly, a strong therapeutic alliance allows for more safety for transference interpretations. Moreover, it seems that when the alliance is tenuous, transference interpretations are dangerous for the alliance. The risk is great that the patient will feel criticized or that the comment is confirming the reality of negative feelings. Transference interpretations can prematurely disrupt the development of a fantasized merger or an idealizing transference, both of which may be necessary in the initial stages of building an alliance, especially for more fragile patients.⁹

Defense and Conflict Interpretations and the Alliance

Interpretations of defense and conflict are less threatening than transference interpretations, but they still keep the discourse at an affective level. Regardless of the eventual outcome in the therapy of both Ms. A. and Ms. B., it is possible to observe how their therapists used supportive interventions to buffer or repair a damaged alliance. These are used to contain the patient's anxiety while moving away from a transference focus. When Ms. A. starts to cry and expresses helplessness, her therapist moves out of the transference to more distant explanations of expectations and disappointment and defense interpretations. The patient appears to experience this as less threatening, regains emotional control, and talks about getting “pretty good” marks. With Ms. B., as well, the use of defense interpretations leads to the most productive work in the session.

The use of nontransference interpretation, packaged with support, to deal with the initial phase of therapy is illustrated by Ms. C.'s case. Ms. C. was older than both Ms. A. and Ms. B., she had an established business with her husband, and she was no longer engaged in self-destructive behavior. Thus, she appeared healthier and more mature than Ms. A. or Ms. B. She tended to externalize, and her therapist's interventions were directed to her interactions with third parties. The therapist avoided exploring the patient's feelings or anxieties about the therapist, focusing instead on Ms. C.'s internal conflicts and defenses.

The interventions in this case tend to be supportive, including acknowledgments, reflections, clarifications, and validations. When interpretive, they focus on nontransference relationships. Thus, they are nonthreatening. The risk is that the dyad is avoiding the transference and hence the therapy will be bland and defensive. Possible negative transference would remain unexplored and unresolved, as would subtle but crucial problems. Langs²⁵ calls this a misalliance cure. We are not in a position to say whether this is what took place in Ms. C.'s therapy. We can say that the nontransference interpretation approach led to more elaboration of emotionally significant material in the session, and that the patient was very satisfied at the end of therapy and had made gains toward her own goals for her therapy.

These qualitative observations are congruent with the results of quantitative sequential analysis by Milbrath et al. (unpublished, 1997). They found that the context in which therapists most consistently made interpretive interventions was when patients conveyed emotions or made significant connections. In turn, interpretations of defense and conflict were followed by more emotional elaboration and insightful connections. Transference interpretations were followed by significant elaboration, but also by temporarily increased defensiveness that probably was a reaction to the anxiety created by the transference interpretation.

Thus, there is both quantitative and qualitative evidence that transference interpretations are “high-risk, high-gain” interventions, as Gabbard and colleagues¹⁰ note, and that they must be embedded in a context of strong alliance and supportive interventions. This caution applies not only to borderline patients but to the full spectrum of patients in psychotherapy, especially in the early phase of treatment. Investigation of the larger sample of our long-term psychotherapy project will eventually yield information about these interactions in different types of patients. As of now, we cannot generalize to the larger sample.

Limitations

There are numerous methodological limitations to this study. First, our sample is limited by its small size and range of patient diagnoses and by its restriction to female patients. Second, although our evaluation of the state of the alliance during the whole session is done by an accepted instrument, the CALPAS, our microassessments of changes in the alliance during the course of the session is by subjective judgment, albeit using the consensus of three raters. Third, although we can state the sequence of events in a session, any conclusions about causation are only speculative. Fourth, the category of transference interpretation in our intervention measure includes all therapist comments on the patient–therapist relationship but does not discriminate between positive and negative transference or determine whether or not the comments are well timed or empathic. These limitations must be put in the perspective of the complex nature of the process we are studying.

This study responds to Horvath and Luborsky's¹ call for more research into the relationship between specific therapist interventions and process and outcome. Further analysis of a wider range of therapeutic situations will be needed to shed greater light on these complex relationships and to build on our present findings.

Appendix

References

1. Horvath AO, Luborsky L: The role of the therapeutic alliance in psychotherapy. J Consult Clin Psychol 1993; 61:561–573. [PubMed]

2. Gabbard GO, Horwitz L, Frieswyk S, et al: The effect of therapist interventions on the therapeutic alliance with borderline patients. J Am Psychoanal Assoc 1988; 36:697–727. [PubMed]

3. Greenson R: The working alliance and the transference neurosis. Psychoanal Q 1965; 34:155–181. [PubMed]

4. Zetzel ER: A developmental approach to the borderline patient. Am J Psychiatry 1971; 127:861–871. [PubMed]

5. Adler G: Borderline Psychopathology and its Treatment. New York, Jason Aronson, 1985.

6. Gunderson JG, Frank AF, Rothingstam EF, et al: Early discontinuance of borderline patients from psychotherapy. J Nerv Ment Dis 1989; 177:38–42. [PubMed]

7. Kernberg OF: Severe Personality Disorders: Psychotherapeutic Strategies. New Haven, CT, Yale University Press, 1984.

8. Masterson JF: The borderline adult: therapeutic alliance and transference. Am J Psychiatry 1978; 135:437–441. [PubMed]

9. Kohut H: The Restoration of the Self. New York, International Universities Press, 1977.

10. Gabbard GO, Horwitz L, Allen JG, et al: Transference interpretation in the psychotherapy of borderline patients: a high-risk, high-gain phenomenon. Harv Rev Psychiatry 1994; 4:59–69. [PubMed]

11. Piper WE, Azim HF, Joyce AS, et al: Transference interpretations, therapeutic alliance and outcome in short-term individual psychotherapy. Arch Gen Psychiatry 1991; 48:946–953. [PubMed]

12. Buckley P, Conte HR, Plutchik R, et al: Psychodynamic variables as predictors of psychotherapy. Am J Psychiatry 1994; 141:742–747. [PubMed]

13. Gaston L: The concept of the alliance and its role in psychotherapy: theoretical and empirical considerations. Psychotherapy 1990; 27:143–153.

14. Horvath AO, Symonds BD: Relation between working alliance and outcome in psychotherapy. J Consult Clin Psychol 1991; 38:139–149.

15. Sexton HC, Hembrek K, Kvarme G: The interaction of the alliance and therapy microprocess: a sequential analysis. J Consult Clin Psychol 1996; 64:471–480. [PubMed]

16. Horvath AO, Marx RW: The development and decay of the working alliance during time-limited counselling. Canadian Journal of Counselling 1991; 24:240–259.

17. Wallerstein RS: The psychotherapy research project of the Menninger Foundation: an overview. J Consult Clin Psychol 1989; 57:195–205. [PubMed]

18. Williams JBW, Gibbon M, First MP, et al: The Structured Clinical Interview for DSM-III-R (SCID) II: multi-site test-retest reliability. Arch Gen Psychiatry 1992; 49:624–629. [PubMed]

19. Marmar CR, Gaston L, Thompson LW, et al: Alliance and outcome in late life depression. J Nerv Ment Dis 1989; 177:464–472. [PubMed]

20. Horowitz MJ, Milbrath C, Reidbord S, et al: Elaboration and dyselaboration: measures of expression and defense in discourse. Psychotherapy Research 1993; 3:278–293.

21. Gaston L, Ring JM: Preliminary results on the Inventory of Therapeutic Strategies. J Psychother Pract Res 1992; 1:135–146. [PMC free article] [PubMed]

22. Foreman SA, Marmar CR: Therapist actions that address initially poor therapeutic alliances in psychotherapy. Am J Psychiatry 1985; 142:922–926. [PubMed]

23. Langs R: The therapeutic relationship and deviations in technique. International Journal of Psychoanalytic Psychotherapy 1975; 4:106–141. [PubMed]

24. Weiss J: How Psychotherapy Works: Process and Technique. New York, Guilford, 1993.

25. Langs R: Therapeutic misalliances. International Journal of Psychoanalytic Psychotherapy 1975; 4:77–105. [PubMed]

Articles from The Journal of Psychotherapy Practice and Research are provided here courtesy of
American Psychiatric Publishing