Psychodynamic psychotherapy of

borderline personality disorder:

A contemporary approach
Glen O. Gabbard, MD

Recent trends in the economics of mental health care threaten to

undermine the use of psychodynamic psychotherapy for the
treatment of patients with borderline personality disorder. These
trends are driven in part by the assumption that such treatment of
these challenging patients is very expensive. The author highlights
empirical research that supports both the usefulness and the cost-
effectiveness of this treatment approach. He also reviews some
effective clinical strategies with borderline patients. (Bulletin of the
Menninger Clinic, 65[1], 41–57)

Psychodynamic psychotherapy of borderline personality disorder has

a rich tradition (Adler, 1985; Boyer, 1977; Gunderson, 1984; Kern-
berg, 1975; Meissner, 1984; Rinsley, 1989; Waldinger & Gunderson,
1987). However, the clinical wisdom derived from this tradition has
recently been placed in substantial jeopardy by the politics of mental
health care economics. Some managed care companies have estab-
lished policies that prohibit reimbursement for disorders on Axis II of
the Diagnostic and Statistical Manual of Mental Disorders, 4th edi-
tion (DSM-IV; American Psychiatric Association, 1994). Other com-
panies will provide a small number of sessions, often 6 to 10, for the
psychotherapy of borderline personality disorder (BPD). Still others
will even assert that there is no evidence that psychotherapy is effec-
tive for borderline patients.
The underlying assumption in this extreme limitation on treatment
of BPD is that extended psychotherapy of such patients will be ex-
traordinarily costly. Many psychotherapists have despaired at their
inability to apply what they have been trained to do with this group
of challenging patients. Similarly, patients with the BPD diagnosis are
often unable to obtain much-needed treatment and sink into suicidal

This article is based on a presentation at the 22nd Annual Menninger Winter

Psychiatry Conference held March 5–10, 2000, at Park City, Utah.
Dr. Gabbard is Bessie Walker Callaway Distinguished Professor of Psychoanalysis
and Education in the Karl Menninger School of Psychiatry and Mental Health
Sciences, The Menninger Clinic. Correspondence may be sent to Dr. Gabbard at The
Menninger Clinic, PO Box 829, Topeka, KS 66601-0829; e-mail:
gabbargo@menninger.edu. (Copyright © 2001 The Menninger Foundation)

Vol. 65, No. 1 (Winter 2001) 41

Gabbard

despair. A growing literature actually demonstrates that patients with

BPD may benefit from a year or more of psychodynamic psychother-
apy and that such treatment may ultimately be highly cost-effective
(Gabbard, 1997; Gabbard, Lazar, Hornberger, & Spiegel, 1997).
In this article I provide a brief overview of the recent empirical
research supporting the usefulness and cost-effectiveness of psycho-
dynamic psychotherapy for BPD. I also review some clinically useful
strategies with these patients and illustrate them with clinical
examples.

Research on psychotherapy outcome and cost-effectiveness

No research has demonstrated that any short-term psychotherapy is

effective for BPD. On the other hand, in studies investigating a year
or more of therapy, the results are encouraging. Hoke (1990) fol-
lowed 58 borderline patients for up to 7 years. These patients were
divided into two distinct groups. The first group, which comprised
about 50% of the sample, had inconsistent or intermittent psy-
chotherapy, while the second group had regular weekly psychother-
apy for at least 2 years. The second group had considerably greater
improvement in mood functioning, decreased impulsiveness, im-
proved Global Assessment Scale (Endicott, Spitzer, Fliess, & Cohen,
1976) scores, and a decreased need for more intensive psychiatric
treatment, such as partial hospital treatment, emergency room visits,
or inpatient services.
Thirty patients with borderline personality disorder were treated in
Sidney, Australia, with twice-weekly outpatient psychotherapy for 12
months by trainee therapists who were closely supervised (Stevenson
& Meares, 1992). The psychotherapy was psychodynamically ori-
ented and specifically influenced by the ideas of Winnicott and Kohut,
with a primary emphasis on self-development. No control group was
used in this study, but patients were rated before and after the psy-
chotherapy in a “pre-post” design. The patients’ ratings 12 months
after the 1-year psychotherapy ended were compared to those for the
12 months preceding the beginning of the psychotherapy. There were
highly significant differences. Among the most striking findings were
the following: (1) the number of hospital admissions decreased by
59% after the psychotherapy; (2) the time spent as an inpatient de-
creased by half; (3) the number of visits to medical professionals
dropped to one seventh of pretreatment rates; (4) the number of self-
harm episodes declined to one fourth of pretreatment rates; and (5) the
time spent away from work was only 1.37 months per year after ther-
apy, compared to 4.47 months per year prior to therapy.

42 Bulletin of the Menninger Clinic

 Psychodynamic psychotherapy of BPD

These substantial improvements were sustained at 5-year follow-

up for the most part (Stevenson & Meares, 1995). In some cases the
improvements were even greater at 5-year follow-up. Only one out-
come measure seemed to worsen with time. Time spent away from
work started increasing during the 5-year outcome, although a reces-
sion in Australia may have contributed to this outcome.
The same group of Australian researchers (Meares, Stevenson, &
Comerford, 1999) later published a comparison of this same cohort
of 30 BPD patients to a waiting-list group of patients with the same
diagnosis. This comparison group was composed of the first 30 pa-
tients on the list who had been waiting for 12 months. During that
interval, the patients had their usual treatments, including crisis inter-
vention, cognitive therapy, and supportive therapy. Of the 30 pa-
tients treated with psychodynamic psychotherapy, 30% no longer
met criteria for the diagnosis of BPD after 12 months of psychother-
apy. The 30 patients on the waiting list for 12 months or longer
showed no change in diagnosis. Although these results suggest that
the substantial gains occurred from the dynamic psychotherapy, con-
clusions must be tentative because randomization was not employed
in establishing the control group.
In the Halliwick Day Unit study in London (Bateman & Fonagy,
1999), 38 borderline patients in a psychoanalytically oriented partial
hospital program were compared to those in a control group. The
major thrust of the partial hospital treatment was psychoanalytic
psychotherapy.* In that treatment cell, the treatment consisted of
once-weekly individual psychoanalytic psychotherapy, three-times-
per-week group psychoanalytic therapy, once-weekly expressive ther-
apy informed by psychodrama techniques, weekly community meet-
ing, medication review by a resident psychiatrist, and meeting with a
case coordinator. By contrast, in the control condition, the treatment
was as follows: medication similar to the treatment group, no psy-
chotherapy, outpatient and community follow-up, inpatient admis-
sion as appropriate, and regular psychiatric review approximately
twice a month by a senior psychiatrist.
At 18-month follow-up, there were substantial differences be-
tween the two groups. Because the investigators employed a random-
ized controlled design, more definitive conclusions about the efficacy
of such treatment can be drawn. After 18 months of treatment, the
partial hospital group had a dramatic reduction in the percentage of

*In this article, the terms psychoanalytic and psychodynamic will be used
interchangeably.

Vol. 65, No. 1 (Winter 2001) 43

Gabbard

the sample who made suicide attempts in the previous 6 months. At

the beginning of the treatment, 95% had made suicide attempts in
the past 6 months, but after 18 months the proportion had decreased
to 5.3%. In addition, the mean length of hospitalization in the con-
trol condition dramatically increased in the last 6 months of the
study, while it remained stable in the treatment group at approxi-
mately 4 days per 6 months. Anxiety symptoms also decreased sub-
stantially in the treatment group, while remaining unchanged in the
control group. Depression scores significantly decreased in the treat-
ment group but not in the control group, and the severity of symp-
toms also decreased significantly in the treatment group in 18
months. The investigators concluded that suicidal acts and psychi-
atric symptoms began to show improvement after 6 months of treat-
ment, but a reduction in frequency of hospital admission and length
of inpatient stay was clear only in the last 6 months, indicating a
need for more extended treatment.
This brief overview of recent studies supports the notion that ex-
tended psychodynamic psychotherapy may provide substantial im-
provements for patients with BPD. In addition, these data also pro-
vide convincing evidence that extended weekly psychotherapy may
be quite cost-effective over time. Borderline psychopathology by its
very nature leads patients to seek help from professionals. If they are
denied access to regular psychotherapy, they often will appear in
emergency rooms after overdoses that require intensive care or inpa-
tient treatment. They also visit the offices of other medical practition-
ers with a variety of somatic complaints. The extensive work disabil-
ity accounts for so-called indirect costs in relation to BPD (Gabbard,
1997). As these studies suggest, psychotherapy may reduce the pa-
tient’s use of other treatments, and particularly the hospital, while
also improving work performance. The Australian study (Stevenson
& Meares, 1999) found that based on the decrease in hospital treat-
ment alone, psychotherapy of borderline patients results in consider-
able savings in health care costs. During the 12 months prior to treat-
ment, in Australian dollars hospital treatment alone cost $684,346,
with a range of $0 to $143,756 per patient. The cost of hospital ad-
missions for the year after treatment was $41,424, with a range of $0
to $12,333 per patient. The average decrease in cost per patient was
$21,431 over 12 months. The average cost of therapy per patient was
$13,000, representing a savings per patient of $8,431.
Hence, preliminary data suggest that a few sessions here and there
of hit-and-miss treatment will not provide the kind of containment
necessary for borderline patients to avoid using other treatments. On
the other hand, regular weekly psychodynamic therapy of 12 months

44 Bulletin of the Menninger Clinic

 Psychodynamic psychotherapy of BPD

or more results in much less use of hospitalization and other special-

ists and may well save money in the long run. One of the worst situa-
tions for patients who have borderline psychopathology is to be in a
psychotherapy of uncertain length, where a managed care reviewer
monitors the treatment week by week and may decide to discontinue
reimbursement at any moment. Because abandonment issues are a
central concern for patients with BPD, the uncertainty of this type of
arrangement may cause them to have overwhelming anxiety based on
the fear that their attachment to their therapist may be disrupted at
the whim of an external reviewer (Gabbard, 1997).

Psychodynamic strategies

Use expressive and supportive approaches flexibly

Much of the controversy in the psychoanalytic psychotherapy litera-
ture regarding BPD has revolved around whether the treatment
should be predominantly expressive or primarily supportive (Hor-
witz et al., 1996). While Kernberg (1975), for example, has advo-
cated early interpretation of transference, others, such as Zetzel
(1971), have argued for a supportive ego-building approach at lower
frequency. To study these different viewpoints, the Menninger Treat-
ment Interventions Project (TRIP) studied process material from
three cases of extended psychoanalytic psychotherapy of borderline
patients at The Menninger Clinic (Gabbard et al., 1988, 1994; Hor-
witz et al., 1996). Two sets of investigators worked from typed tran-
scripts of randomly selected psychotherapy hours that were audio-
taped. One team of clinician researchers rated the therapist
interventions along an expressive-supportive continuum. That con-
tinuum included the following interventions, from the most expres-
sive to the most supportive: interpretation, confrontation, clarifica-
tion, encouragement to elaborate, empathic validation, advice or
praise, and affirmation. These seven interventions were also rated in
terms of whether the therapist had a transference or extratransfer-
ence focus. A separate research team rated the patient’s collaboration
with the therapist. These investigators used upward or downward
shifts in the patient’s collaboration, as measured by making produc-
tive use of the therapist’s contributions or by bringing in significant
content. Collaboration was viewed as a measurable marker of the
therapeutic alliance. The two research teams were trying to determine
whether shifts upward or downward in a patient’s therapeutic al-
liance could be linked to specific styles of interventions by the
therapist.
After reviewing the data, the TRIP researchers concluded that

Vol. 65, No. 1 (Winter 2001) 45

Gabbard

transference interpretations are a “high-risk, high-gain” intervention

in the dynamic psychotherapy of BPD (Gabbard et al., 1994). These
highly expressive interventions had greater impact—both negative
and positive—than the other interventions made by the therapist. In
some cases transference interpretation led to marked deterioration in
the patient’s capacity to collaborate with the therapist, while in other
situations collaboration substantially improved.
The investigators looked at several factors that determined which
interpretations increased collaboration and which led to deteriora-
tion of the therapeutic alliance. Paving the way for transference inter-
pretation with empathic validation of the patient’s internal experi-
ence appeared to be of crucial importance. Just as surgeons require
anesthesia before they can operate, psychotherapists may need to cre-
ate a holding environment through an affirmation of the patient’s ex-
perience before offering an outside perspective on what they see as
happening inside the patient. In many discussions of the psychother-
apy of BPD, supportive and expressive approaches are often artifi-
cially polarized. The TRIP study, however, demonstrated that these
two approaches often work synergistically on the patient’s behalf.
The investigators concluded that there is a spectrum of borderline
psychopathology, and a “one size fits all” strategy is misleading.
Some patients benefit from predominantly expressive interventions,
while others require much more support. The skilled psychotherapist
tailors the psychotherapeutic approach to the patient’s particular
needs.
The etiology and pathogenesis of BPD often involves trauma in the
form of abuse and neglect (Zanarini et al., 1997). In order to form a
strong therapeutic alliance, patients who have experienced a trau-
matic childhood may need a supportive and validating atmosphere
that recognizes the reality of early trauma (Gunderson & Chu, 1993;
Gunderson & Sabbo, 1993; Horwitz et al., 1996). For example, such
patients may experience the therapist’s interpretation of a distortion
in the therapist’s intent as an empathic failure. On the other hand, if
therapists can recognize that patients have good reason to distrust a
therapist because of early experience with authority figures, then the
patient may feel validated and understood.
The TRIP investigators found that borderline patients who have
greater ego strength and greater psychological mindedness will in
general be able to use a more expressively oriented psychotherapy
than those who are closer to the psychotic border. Patients with poor
impulse control, poor tolerance of anxiety, and an excessively con-
crete cognitive style will need a predominantly supportive emphasis.
The typical borderline patient will require a flexible stance by the

46 Bulletin of the Menninger Clinic

 Psychodynamic psychotherapy of BPD

therapist, with shifts between interpretive and noninterpretive ap-

proaches that are related to the patient’s experience of the therapist
at any given moment. Experienced psychodynamic therapists typi-
cally use a trial-and-error approach until they can clearly determine
which interventions are most effective in enhancing the alliance and
deepening the patient’s understanding.
Elsewhere, I (Gabbard, 2000b) have outlined several principles of
technique that apply rather broadly to most patients who have BPD.
Avoid rigidity
Many beginning therapists who are well aware of the literature on
boundary problems with borderline patients (Gabbard & Wilkinson,
1994; Gutheil, 1989) become excessively rigid in their psychothera-
peutic stance. Patients with BPD may experience this unyielding pos-
ture as cold and remote and end up quitting the therapy because they
do not feel understood by or connected to the therapist. A more clin-
ically useful strategy is to strive for a spontaneous and flexible posi-
tion in which boundaries are observed, but the patient is allowed to
actualize certain patterns of internal object relations with the thera-
pist (Gabbard, 1998; Sandler, 1981). As in all human relationships,
the borderline patient tries to impose on others a particular way of
experiencing and responding. Therapists must be sufficiently flexible
to join in this “dance” evoked by the patient so that the characteristic
pattern of relatedness can be observed and understood. For example,
Mr. A began his first several sessions with his new psychotherapist by
sulking and not participating in the psychotherapy. His therapist be-
came increasingly active in trying to coax him to speak. Mr. A resis-
ted all attempts by his therapist to get him to talk, and eventually his
therapist became irritated and acknowledged his irritation to the pa-
tient. Mr. A told him, “You’re just like my mom. She’ll never leave
me alone and let me do what I want.” By getting sucked in to this
pattern with the patient, the therapist had re-created an interaction
that commonly occurred between the patient and his mother. The pa-
tient’s contribution to the re-creation of that interaction could then
be reflected on and understood.
Establish conditions that make psychotherapy viable
The inherent instability of the borderline patient demands that struc-
ture must be imposed from external sources. Before psychotherapy is
started, several points should be thoroughly discussed with the pa-
tient: (1) what therapy is and what it is not, (2) the need to end ses-
sions on time, (3) expectations about payment, (4) regular appoint-
ment times, (5) a missed appointments policy, and (6) the therapist’s
expectation that the patient must be a collaborator in the psychother-

Vol. 65, No. 1 (Winter 2001) 47

Gabbard

apy process and work actively with the therapist in defining goals
and working toward them.
The approach to suicidality in BPD is somewhat controversial.
Some clinicians (Clarkin, Yeomans, & Kernberg, 1999; Kernberg,
Selzer, Koenigsberg, Carr, & Appelbaum, 1989) advocate establish-
ing a “contract” in the pretherapy phase of consultations. Within this
framework therapists should clarify with the patient that their role is
not to get involved in the actions of the patient’s life outside of psy-
chotherapy sessions. They would make clear to the patient that their
availability is limited and that they would not expect to receive
phone calls between sessions. Therapists have different tolerance lev-
els for phone calls between sessions. My own preference is to have a
suicidal borderline patient call me if the patient feels that suicidal im-
pulses are out of control and hospitalization is required. Moreover,
borderline patients with poorly developed object constancy or evoca-
tive memory (Adler, 1985) may feel that their therapist has disap-
peared over the weekend or over a vacation period when they cannot
summon an internal image of the therapist to sustain them through a
stressful time. A brief phone call may reestablish the connection with
the therapist and head off a good deal of self-destructive behavior or
even suicide.
Gunderson (1996) stresses that too rigid a contract may interfere
with the development of a stable attachment to the therapist. He
also suggests that the therapist’s between-session availability should
not be brought up as part of the initial consultation phase. Rather,
the therapist should wait until after the patient asks about the ther-
apist’s availability. Gunderson and I both agree that even within
this model, patients should be told to contact their therapist in a
bona fide emergency.
Some patients experience contract setting as being asked to do the
impossible. They may feel misunderstood and accused, thus starting
off the process in an adversarial relationship with the therapist.
When phone calls do occur between sessions, Gunderson (1996) em-
phasizes that these calls should then be the therapeutic focus. If the
patient is having recurrent reactions of panic because evocative mem-
ory and object constancy are not well established, the therapist can
help him or her understand that fear of aloneness and the incapacity
to internalize a soothing figure. Over time the therapist may be inter-
nalized as a stable representation that will help the patient get
through periods of aloneness. This approach is in keeping with
Adler’s (1985) notion that borderline patients lack a holding-sooth-
ing introject. When phone calls become excessive, clear limits should
be set with the patient. Therapists may wish to explain their own lim-

48 Bulletin of the Menninger Clinic

 Psychodynamic psychotherapy of BPD

its to the patient as well as explore the significance and meanings of

excessive between-session contact.
Allow transformation into the bad object
Containment and management of hatred, sadism, aggression, and
anger are fundamental to the psychotherapy of patients with BPD.
These affects are central to the psychopathology of the patient and
are activated in the therapist’s countertransference. Therapists com-
monly feel that they are being falsely accused (Gabbard, 1991), and
they are frequently tempted to retaliate against the patient as a way
of defending themselves against what they perceive as attacks on
their character. Borderline patients frequently have internalized a
hating self and a hated internal object, either of which they may fran-
tically try to externalize using the defense of projective identification.
In this regard, Rosen (1993) has pointed out that such patients are
searching for a “bad enough object.” Recreating a sadomasochistic
object relationship with the therapist is often experienced by border-
line patients as familiar, predictable, and even soothing because this
relationship paradigm is what they have known since childhood.
Therapists who resist this transformation into the bad object by act-
ing increasingly saintly and empathic may force patients to escalate
their provocativeness and try even more desperately to transform the
therapist (Fonagy, 1998).
To allow the transformation into the bad object role does not
mean that the therapist must lose all professional decorum. Signs of
exasperation and irritation may begin to manifest themselves in a va-
riety of countertransference enactments. Some therapists may force
transference interpretations prematurely in the effort to make pa-
tients take back their own projected hostility. Another variant of en-
actment is for the therapist to withdraw and become silent, essen-
tially disengaging emotionally from the patient. Still other therapists
may feel a sense of despair and have an urge to give up on the ther-
apy all together. Sarcastic and hostile comments may be offered
under the rationalization of necessary confrontation. While all these
enactments are expectable developments in the therapy of borderline
patients, the therapist must nevertheless strive to function as a con-
tainer who can tolerate the hatred and aggression being projected
and contain it until the patient can reown it. The optimal state of
mind for therapists is one in which they “allow themselves to be
‘sucked in’ to the patient’s world while retaining the ability to ob-
serve it happening in front of their eyes. In such a state, therapists are
truly thinking their own thoughts, even though they are under the pa-
tient’s influence to some extent” (Gabbard & Wilkinson, 1994, p.

Vol. 65, No. 1 (Winter 2001) 49

Gabbard

82). In other words, therapists must gradually disengage themselves

from becoming what the patient evokes and finding their way back
to a center deep within themselves where they can think their own
thoughts rather than the patient’s thoughts.
Promote reflective function
Many patients with borderline personality disorder lack the capacity
to reflect on their own internal states and those of others (Fonagy,
1998). Because of early attachment problems, they have a great deal
of difficulty recognizing that their actions and interactions are moti-
vated by internal states and that others operate out of separate and
different internal states. This capacity to mentalize or conceive of in-
ternal states in self and others must be promoted as an integral part
of the psychotherapy process. Interpreting meanings of enactments
may be premature in such patients. A beginning step is to assist the
patient in elaborating on the emotional state that may have led to the
enactment.
Reflective function or mentalization can also be encouraged by
helping the patient observe moment-to-moment changes in feelings
that occur in the here-and-now interactions between therapist and
patient. The eventual goal is for the patient to internalize the thera-
pist’s observations of his or her internal states. Encouraging the ex-
pression of fantasies about the therapist’s internal state may also pro-
mote mentalization. Hence Gunderson (1996) suggests that when a
therapist is called in the middle of the night, a useful question at the
next psychotherapy session might be “How did you think I would
feel about your call?” In this manner the therapist helps the patient
recognize that the therapist has a separate center of autonomy and
subjectivity. Asking the patient to think through consequences of
self-destructive behavior also promotes reflectiveness and may assist
in heading off the patient’s damaging enactments.

Ms. B, a 29-year-old chronically suicidal borderline patient who

started psychotherapy with me, revealed a core feeling of
hopelessness about the treatment. She was a survivor of incest, and
while sitting in the therapy, she had the posttraumatic sensation of
her father’s hands being all over her. She told me that when she cut
her wrist, it was her father’s blood that she saw. When she looked at
her skin, she said she saw her father’s skin. She said that she was
rotten inside because of what he’d done to her. She carried him with
her and said that she wanted to commit suicide as a way of actually
killing him. When she scratched her wrist and saw her blood, she felt
disgusted because she thought the blood was rotten. She was

50 Bulletin of the Menninger Clinic

 Psychodynamic psychotherapy of BPD

completely unable to differentiate herself from her father and felt

that suicide was the only way out of his clutches.
Ms. B would frequently come to therapy and say that she did not
want to talk. She felt it was futile. One day she asked if she could
draw what she was thinking. I told her that I would certainly be
interested in seeing what she would draw. She had a pad of paper
and some colored pencils. She drew a picture of herself in a coffin 6
feet underground. On top of the ground was a tombstone. She
handed me the picture. I told her I thought it was incomplete. She
asked what I meant. I asked her if I could have the pencils. I then
drew a picture of her 5-year-old son standing at the gravesite
looking down at the grave. She angrily pushed the paper away and
told me she did not want to think about that. I replied that if she
were thinking seriously about suicide, she certainly had to consider
all the consequences of her actions. She accused me of laying a “guilt
trip” on her. I responded that I was simply helping her think
realistically about what would happen in the wake of her suicide and
how her son would feel about it. I explained to her that it was
terribly difficult for children to deal with their mother’s suicide and
that I knew she loved her son and would be concerned about his
welfare after she was gone.
A couple of months after this session she was discharged from the
hospital and returned to her home city. About 2 years later I had a
contact with her subsequent psychotherapist in that city. I asked if
Ms. B were still alive, and he said she was. He told me that she
frequently made reference to the drawing we did together, and she
stated angrily that I had made her feel guilty by bringing up the fate
of her son following suicide. The therapist added, “Of course, she’s
still alive.”

Through my simple intervention of drawing her son at the

gravesite, I had promoted Ms. B’s reflective function. I had helped
her realize that others, particularly her own child, might feel differ-
ently about her suicide than what she had imagined. I have occasion-
ally read in the literature on suicide that it is ill-advised to talk pa-
tients out of suicide by encouraging them to go on living for someone
else. These patients often have spent their lives, the argument goes,
living for others. Although I can see some merit in that position, I
think there is a great deal to be gained by emphasizing the meaning-
ful connections to others. Lifton (1979) noted, “However diverse the
meanings of induced death in various cultural and historical situa-
tions, we should not be blinded to the essential feature of all suicide:
its violent statement about human connection, broken and main-

Vol. 65, No. 1 (Winter 2001) 51

Gabbard

tained” (p. 239). Almost all suicidal people are affected by despair,
by a sense of radical absence of meaning and purpose, and of the im-
possibility of human connection.
Part of borderline patients’ psychopathology is an absorption in
their own suffering to the point where the subjectivity of others is
completely disregarded. When a borderline patient lacks reflective
function, the therapist must provide this reflective aspect rather than
simply empathizing with the patient’s point of view. Fonagy and Tar-
get (1996) stress that “in order to move the child from the mode of
psychic equivalence to the mentalizing mode, analytic reflection, of
whatever orientation, cannot just ‘copy’ the child’s internal state, but
has to move beyond it and go a step further, offering a different, yet
experientially appropriate representation” (p. 231). In so doing, the
therapist gradually helps the patient learn that mental experience in-
volves representations that can be played with and ultimately altered.
This perspective implies a theory of therapeutic action that empha-
sizes the therapist as a new, real object as the patient gradually appre-
ciates the therapist’s separate subjectivity.
One implication of this developmental model is that the therapist
must bring a different point of view to bear. Thus when Ms. B said
that her son would cry for a little while but would eventually get over
it, I replied, “No, that’s not actually very likely at all.” I went on to
say that children are often haunted their entire lives by their parent’s
suicide and often blame themselves.

Set limits when necessary

Boundaries are typically tested by patients with BPD. They often ex-
perience the usual professional boundaries as sadistic and cruel depri-
vations by the therapist. Some patients demand and beg for more
overt demonstrations of caring, such as extension of the session be-
yond the time boundary, decreases in the fee, hugs during the ther-
apy, and 24-hour availability (Gabbard & Wilkinson, 1994). Thera-
pists who fall into the trap of trying to gratify these demands soon
come face to face with a profound insatiability in the patient. The de-
mands become endless and tormenting.
On the other hand, therapists cannot maintain an emotional dis-
tance and be unresponsive to the patient’s emotional pleas. A helpful
guideline is the distinction drawn by Casement (1985) between “li-
bidinal demands” and “growth needs.” The former cannot be grati-
fied without gravely jeopardizing the treatment and committing seri-
ous ethical compromises. The latter can be facilitated within
professional boundaries. While consistency is part of creating a hold-
ing environment for the patient, empathic responses to the patient’s

52 Bulletin of the Menninger Clinic

 Psychodynamic psychotherapy of BPD

changing needs is also a critical factor in maintaining the therapeutic

alliance.
Much of the countertransference difficulty is the therapist’s feeling
of being cruel when enforcing reasonable limits on the patient’s en-
actments. Paradoxically, though, many patients who demand greater
freedom get worse when it is granted to them. In a spin-off study of
the Menninger Psychotherapy Research Project, Colson, Lewis, and
Horwitz (1985) examined the cases that had negative outcomes. One
common denominator was the therapist’s failure to set limits on act-
ing-out behavior. Instead, the therapist would simply go on interpret-
ing unconscious motivations for the acting out while the patient’s
condition continued to deteriorate.
Waldinger and Gunderson (1987) recognized that the therapist
cannot be policing the patient’s behavior to the point where limits are
set for any untoward enactments. They suggested that the behaviors
targeted for limit setting should be those that threaten the safety of
the patient or the therapist, or that jeopardize the psychotherapy it-
self. Therapists must always remember that they cannot omnipo-
tently protect patients from suicide by having continuous contact
with them in an outpatient setting. When suicidal impulses are out of
control, the patient must be hospitalized to be protected.

Establish and maintain the therapeutic alliance

As noted earlier, the therapeutic alliance is a challenge to establish
with a patient who has BPD. Because BPD patients use splitting as a
fundamental defensive style, they will frequently view the therapist as
either an adversary or an idealized rescuer. Therapists must con-
stantly bring the patient back to the task at hand. To strengthen the
alliance, the therapist should revisit the commonly held goals of ther-
apy whenever the process becomes particularly difficult and the ther-
apist is experienced as someone other than a therapist. I often remind
the patient that therapy is not coercive. The patient has chosen to
work with me around specific goals that create suffering. Many pa-
tients lose track of these goals, so coming back to them is a way of
reminding patients that the therapist is an ally who is working
collaboratively.

Avoid a split between psychotherapy and pharmacotherapy

Because of advances in the psychopharmacology of BPD, most pa-
tients today will be involved in combined treatment in which both
medication and psychotherapy are prescribed. A systematic ap-
proach to the pharmacotherapy of BPD has been described else-
where (Gabbard, 2000a). If the same psychiatrist is doing both the

Vol. 65, No. 1 (Winter 2001) 53

Gabbard

pharmacotherapy and the psychotherapy with a borderline patient,

the therapist should avoid splitting off the medication as an admin-
istrative matter that requires no exploration. Transference, counter-
transference, and resistance apply to prescribing in the same way
they do to psychotherapy. Waldinger and Frank (1989) surveyed
dynamic therapists who were experienced in treating borderline
patients. They found that nearly half the patients misused the med-
ication, and that transference themes were intimately involved in
that misuse. Similarly, prescribing often involved countertransfer-
ence pessimism or despair rather than a systematic rationale for ap-
proaching target symptoms.
When two different clinicians respectively are conducting the
pharmacotherapy and the psychotherapy, there is even greater dan-
ger of splitting off the medication from the therapy. The two treaters
should think of themselves as part of the same treatment team and
discuss the treatment openly (Gabbard, 2000a). The time for such
discussions is rarely reimbursed by third parties, and there is often lit-
tle financial incentive for the therapist and the prescriber to speak.
However, lack of communication is a setup for splitting. Borderline
patients may idealize the therapist as an empathic listener while de-
valuing the pharmacotherapist as someone who kicks them out of the
office after 15 minutes. On the other hand, the doctor prescribing the
medication may be seen as humane and responsive to symptomatic
complaints while the therapist only listens and does not seem to pro-
vide practical help. This form of splitting can completely disrupt the
treatment.
Help the patient re-own aspects of the self that are disavowed
and/or projected elsewhere
The experience of being incomplete or fragmented is at the core of
borderline psychopathology. Through the defense mechanisms of
splitting and projective identification, BPD patients may disown as-
pects of themselves. The disowned aspects are then projected into
others in their environment. They may feel they need others to make
them whole, and they may have a profound lack of self-continuity
from week to week when the therapist sees them for a psychotherapy
session. Therapists must attempt to help patients with borderline per-
sonality disorder understand that they are unconsciously and auto-
matically projecting aspects of themselves into others as a way of try-
ing to control those disturbing parts of themselves. Interpretation of
a patient’s fear that integrating the bad and the good aspects will lead
to a destruction of all loving aspects by the intense hatred may be an
effective way to help the patient in the task of integration. Psy-

54 Bulletin of the Menninger Clinic

 Psychodynamic psychotherapy of BPD

chotherapists must point out that hate and anger will always be pres-
ent but can be tempered and integrated with love to create a con-
structive balance within.

Concluding comments

A thread that runs through all these techniques is a recognition that

management of countertransference is essential to the psychotherapy
of borderline patients (Gabbard & Wilkinson, 1994). As therapists
systematically examine how they are being transformed by the pa-
tient’s powerful projective identifications, they can help the patient
understand that past patterns of relatedness are being re-created in
the present. By empathizing with the patient’s need to re-create and
interpreting the manner in which it happens, therapists help patients
develop a more integrated view of self and others. Countertransfer-
ence chaos can be overwhelming, and only experienced psychother-
apists who have been well trained in the psychotherapy of border-
line patients should undertake such treatment. Even experienced
therapists are wise to use consultants frequently to help them sort
out their own countertransference and to understand the way the
patient is attempting to transform them into a transference object.
Dynamic psychotherapy of borderline personality disorder has be-
come increasingly refined in recent years. We now have a growing
empirical literature demonstrating that this approach is effective. We
also have data establishing the cost-effectiveness of the treatment.
More research is needed to determine which specific interventions are
central to the therapeutic action of dynamic therapy. In the mean-
time, there is little justification for denying the wide availability of
this treatment.

References
Adler, G. (1985). Borderline psychopathology and its treatment. New York: Aronson.
American Psychiatric Association. (1994). Diagnostic and statistical manual of mental
disorders (4th ed.). Washington, DC: Author.
Bateman, A., & Fonagy, P. (1999). Effectiveness of partial hospitalization in the
treatment of borderline personality disorder: A randomized controlled trial.
American Journal of Psychiatry, 156, 1563–1569.
Boyer, L. B. (1977). Working with a borderline patient. Psychoanalytic Quarterly, 46,
386–424.
Casement, P. J. (1985). On learning from the patient. London: Tavistock.
Clarkin, J. F., Yeomans, F., & Kernberg, O. F. (1999). Psychotherapy of borderline
personality. New York: Wiley.
Colson, D. B., Lewis, L., & Horwitz, L. (1985). Negative outcome in psychotherapy
and psychoanalysis. In D. T. Mays & C. M. Frank (Eds.), Negative outcome in
psychotherapy and what to do about it (pp. 59–75). New York: Springer.

Vol. 65, No. 1 (Winter 2001) 55

Gabbard

Endicott, J., Spitzer, R. L., Fliess, J. L., & Cohen, J. (1976). The Global Assessment
Scale: A procedure for measuring overall severity of psychiatric disturbance.
Archives of General Psychiatry, 33, 766–771.
Fonagy, P. (1998). An attachment theory approach to treatment of the difficult
patient. Bulletin of the Menninger Clinic, 62, 147–169.
Fonagy, P., & Target, M. (1996). Playing with reality, I. International Journal of
Psycho-Analysis, 77, 217–233.
Gabbard, G. O. (1991). Technical approaches to transference hate in the analysis of
borderline patients. International Journal of Psycho-Analysis, 72, 625–637.
Gabbard, G. O. (1997). Borderline personality disorder and rational managed care
policy. Psychoanalytic Inquiry, 17(Suppl.), 17–28.
Gabbard, G. O. (1998). Treatment-resistant borderline personality disorder.
Psychiatric Annals, 28, 561–656.
Gabbard, G. O. (2000a). Combining psychotherapy with pharmacotherapy in the
treatment of personality disorders. In G. O. Gabbard & J. G. Gunderson (Eds.),
American Psychiatric Press review of psychiatry: Volume 19 (pp.
65–94).Washington, DC: American Psychiatric Press.
Gabbard, G. O. (2000b). Psychodynamic psychiatry in clinical practice (3rd ed.).
Washington, DC: American Psychiatric Press.
Gabbard, G. O., Horwitz, L., Allen, J. G., Frieswyk, S. H., Newsom, G., Colson, D.
B., & Coyne, L. (1994). Transference interpretation in the psychotherapy of
borderline patients: A high-risk, high-gain phenomenon. Harvard Review of
Psychiatry, 2, 59–69.
Gabbard, G. O., Horwitz, L., Frieswyk, S. H., Allen, J. G., Colson, D. B., Newsom,
G., E., & Coyne, L. (1988). The effect of therapist interventions on the therapeutic
alliance with borderline patients. Journal of the American Psychoanalytic
Association, 36, 697–727.
Gabbard, G. O., Lazar, S. G., Hornberger, J., & Spiegel, D. (1997). The economic
impact of psychotherapy: A review. American Journal of Psychiatry, 153,
752–758.
Gabbard, G. O., & Wilkinson, S. M. (1994). Management of countertransference
with borderline patients. Washington, DC: American Psychiatric Press.
Gunderson, J. G. (1984). Borderline personality disorder. Washington, DC: American
Psychiatric Press.
Gunderson, J. G. (1996). The borderline patient’s intolerance of aloneness: Insecure
attachments and therapist availability. American Journal of Psychiatry, 153,
752–758.
Gunderson, J. G., & Chu, J. A. (1993). Treatment implications of past trauma in
borderline personality disorder. Harvard Review of Psychiatry, 1, 75–81.
Gunderson, J. G., & Sabbo, A. N. (1993). The phenomenological and conceptual
interface between borderline personality disorder and PTSD. American Journal of
Psychiatry, 150, 19–27.
Gutheil, T. G. (1989). Borderline personality disorder, boundary violations, and
patient-therapist sex: Medicolegal pitfalls. American Journal of Psychiatry, 146,
597–602.
Hoke, L. A. (1990). Longitudinal patterns of behaviors in borderline personality
disorder. Psychiatric Annals, 20, 33–43.
Horwitz, L., Gabbard, G. O., Allen, J. G., Frieswyk, S. H., Colson, D. B.,
Newsom, G. E., & Coyne, L. (1996). Borderline personality disorder: Tailoring
the psychotherapy to the patient. Washington, DC: American Psychiatric
Press.

56 Bulletin of the Menninger Clinic

 Psychodynamic psychotherapy of BPD

Kernberg, O. F. (1975). Borderline conditions and pathological narcissism. New

York: Aronson.
Kernberg, O. F., Selzer, M. A., Koenigsberg, H. W., Carr, A. C., & Appelbaum, A. H.
(1989). Psychodynamic psychotherapy of borderline patients. New York: Basic
Books.
Lifton, R. J. (1979). The broken connection: On death and the continuity of life.
Washington, DC: American Psychiatric Press.
Meares, R., Stevenson, J., & Comerford, A. (1999). Psychotherapy with borderline
patients, I: A comparison between treated and untreated cohorts. Australian and
New Zealand Journal of Psychiatry, 33, 467–472.
Meissner, W. W. (1984). The borderline spectrum: Differential diagnosis and
developmental issues. New York: Aronson.
Rinsley, D. B. (1989). Developmental pathogenesis and psychoanalytic treatment of
borderline and narcissistic personalities. Northvale, NJ: Aronson.
Rosen, I. R. (1993, January 21). Relational masochism: The search for a bad enough
object. Paper presented to Topeka Psychoanalytic Society.
Sandler, J. (1981). Character traits and object relationships. Psychoanalytic
Quarterly, 50, 694–708.
Stevenson, J., & Meares, R. (1992). An outcome study of psychotherapy for patients
with borderline personality disorder. American Journal of Psychiatry, 149,
358–362.
Stevenson, J., & Meares, R. (1995, May). Borderline patients at 5-year follow-up.
Paper presented at the annual Congress of the Royal Australian–New Zealand
College of Psychiatrists, Cairns, Australia.
Stevenson, J., & Meares, R. (1999). Psychodynamic treatment with borderline
patients, II: A preliminary cost-benefit study. Australian and New Zealand Journal
of Psychiatry, 33, 473–477.
Waldinger, R. J., & Frank, A. F. (1989). Clinicians’ experiences in combining
medication and psychotherapy in the treatment of borderline patients. Hospital
and Community Psychiatry, 40, 712–718.
Waldinger, R. J., & Gunderson, J. G. (1987). Effective psychotherapy with borderline
patients: Case studies. New York: Macmillan.
Zanarini, M. C., Williams, A. A., Lewis, R. E., Reich, R. B., Vera, S. C., Marino, M.
F., Levin, A., Yong, L., & Frankenburg, F. R. (1997). Reported pathological
childhood experiences associated with the development of borderline personality
disorder. American Journal of Psychiatry, 154, 1101–1106.
Zetzel, E. R. (1971). A developmental approach to the borderline patient. American
Journal of Psychiatry, 127, 867–871.

Vol. 65, No. 1 (Winter 2001) 57