Change in Attachment Patterns and Reflective Function in a Randomized
Control Trial of Transference-Focused Psychotherapy for Borderline
Kenneth N. Levy, Kevin B. Meehan,
Kristen M. Kelly, Joseph S. Reynoso, and
City University of New York
John F. Clarkin and Otto F. Kernberg
Joan and Sanford I. Weil Medical College of Cornell University
Changes in attachment organization and reflective function (RF) were assessed as putative mechanisms
of change in 1 of 3 year-long psychotherapy treatments for patients with borderline personality disorder
(BPD). Ninety patients reliably diagnosed with BPD were randomized to transference-focused psycho-
therapy (TFP), dialectical behavior therapy, or a modified psychodynamic supportive psychotherapy.
Attachment organization was assessed with the Adult Attachment Interview and the RF coding scale.
After 12 months of treatment, participants showed a significant increase in the number classified secure
with respect to attachment state of mind for TFP but not for the other 2 treatments. Significant changes
in narrative coherence and RF were found as a function of treatment, with TFP showing increases in both
constructs during treatment. No changes in resolution of loss or trauma were observed across treatments.
Findings suggest that 1 year of intensive TFP can increase patients’ narrative coherence and RF. Future
research should establish the relationship between these 2 constructs and relevant psychopathology,
identify treatment components responsible for effecting these changes, and examine the long-term
outcome of these changes.
Keywords: attachment, reflective function, borderline personality disorder, randomized controlled trial
Attachment theory and research have proven to be a powerful
paradigm for studying development, personality, interpersonal re-
lationships and psychopathology. In recent years, clinical writing
about attachment theory has come full circle, back to Bowlby’s
original interests in clinical intervention, by noting the potential
contributions that attachment theory can make to psychotherapy
(Blatt & Levy, 2003; Diamond et al,. 1999; Eagle, 2003, in press;
Farber, Lippert, & Nevas, 1995; Holmes, 1995, 1996; Levy &
Blatt, 1999; Slade, 1999). There has also been a burgeoning
research literature addressing the clinical implications of attach-
ment theory for psychotherapy (Cryanowski et al., 2002; Dozier,
1990; Dozier, Cue, & Barnett, 1994; Fonagy et al., 1996;
Mallinckrodt, Gantt, & Coble, 1995; Meyer, Pilkonis, Proietti,
Heape, & Egan, 2001; Tyrrell, Dozier, Teague, & Fallot, 1999).
Recently, psychopathology researchers and theorists have begun
to understand fundamental aspects of borderline personality dis-
order (BPD), such as unstable, intense interpersonal relationships,
feelings of emptiness, bursts of rage, chronic fears of abandonment
and intolerance for aloneness, and lack of a stable sense of self as
stemming from impairments in the underlying attachment organi-
Kenneth N. Levy, Department of Psychology, Hunter College, and the
Clinical Psychology Doctoral Program, Graduate School and University
Center, City University of New York; Kevin B. Meehan, Joseph S. Rey-
noso, and Michal Weber, Clinical Psychology Doctoral Program, Graduate
School and University Center, City University of New York; Kristen M.
Kelly, Department of Psychology, Hunter College of the City University of
New York; John F. Clarkin and Otto F. Kernberg, Department of Psychi-
atry, Joan and Sanford I. Weil Medical College of Cornell University.
Kenneth N. Levy is now at the Department of Psychology, Pennsylvania
State University University Park Campus. Kristen M. Kelly is now at the
Department of Psychology, Pennsylvania State University University Park
This research was supported by grants from the National Institute of
Mental Health, International Psychoanalytic Association, and the Kohler
Fund of Munich awarded to Kenneth N. Levy, and a grant from the
Borderline Personality Disorder Research Foundation (BPDRF) awarded
to Otto F. Kernberg.
We thank the BPDRF founder and executive officer, Marco Stoffel, and
the scientific board for their advice and encouragement. We thank Jack
Barchas and Vita Rabinowitz for their respective institutional support. In
addition, we acknowledge the technical assistance of Catherine Eubanks
Carter, Kenneth L. Critchfield, Jill C. Delaney, Pamela E. Foelsch, Simone
Hoermann, Maya Kirschner, and Joel McClough for their help in conduct-
ing assessments, and James Hull for organizing and maintaining the data.
We also acknowledge the consultation of Peter Fonagy, Erik Hesse, Ar-
mand Loranger, Mary Main, and Mary Target to this study in training with
the administration and coding of some of the assessment instruments and
design issues. We also thank Jenifer Clark, Komal Choksi, Diana Gutier-
rez, and Chase Stovall-McClough for coding Adult Attachment Interview
transcripts. We thank members of the Personality Disorders Institute and
the Developmental Psychopathology Lab at Hunter College. We thank the
therapists in the study, and Frank Yeomans, Ann Appelbaum, and Barbara
Stanley for serving as treatment cell leaders and providing supervision.
Finally, we thank the patients for their participation.
Correspondence concerning this article should be addressed to Kenneth
N. Levy, Department of Psychology, 521 Bruce V. Moore Building,
Pennsylvania State University, University Park, PA 16802. E-mail:
Journal of Consulting and Clinical Psychology
2006, Vol. 74, No. 6, 1027–1040
Copyright 2006 by the American Psychological Association
zation (Blatt & Levy, 2003; Diamond et al., 1999; Fonagy, 1991;
Fonagy, Gergely, Jurist, & Target, 2002; Gunderson, 1996; Levy,
2005; Levy & Blatt, 1999). These investigators have noted that the
impulsivity, affective lability, and self-damaging actions that are
the hallmark of borderline personality typically occur in interper-
sonal contexts and are often precipitated by real or imagined
events in relationships. These theorists have suggested that change
in attachment representations, conceptualized as a social-cognitive
and affective construct, may be the primary mechanism by which
patients with BPD improve. Attachment theory offers a cogent
conceptualization of the development and maintenance of the
interpersonal difficulties and adaptations that characterize person-
ality pathology while explaining the concomitant development of
self-concept and the problems of self-definition and self-
To date, however, researchers of BPD have not examined
change in attachment representations as a function of psychother-
apy. The purpose of the present research is to explore changes in
attachment representations and organization as a putative mecha-
nism of change in the psychotherapy treatment of borderline
BPD is a highly prevalent, chronic, and debilitating psychiatric
problem characterized by a pattern of chaotic and self-defeating
interpersonal relationships, emotional lability, poor impulse con-
trol, angry outbursts, frequent suicidality, and self-mutilation
(Skodol et al., 2002). Approximately 1%–2% of the population,
10% of psychiatric outpatients, 20% of inpatients, and 6% of
primary care patients meet the Diagnostic and Statistical Manual
of Mental Disorders (4th ed., text revision; DSM–IV–TR; Ameri-
can Psychiatric Association, 2000) criteria for BPD (Gross et al.,
2002; Lenzenweger, Loranger, Korfine, & Neff, 1997; Torgersen,
Kringlen, & Cramer, 2001), the majority of whom are women.
Individuals diagnosed with BPD suffer from devastating behav-
ioral problems. Self-injurious behaviors are particularly prevalent
among these patients, occurring in an estimated 69%–75% of cases
(Kjellander, Bongar, & King, 1998). Other common self-
destructive behaviors include alcohol and drug abuse, risky sexual
behavior, and serious over- or undereating. Patients with BPD are
at high risk of suicide, with a completed suicide rate between 3%
and 9.5% (McGlashan, 1986; Paris, 1999; Stone, 1983). Addition-
ally, BPD is substantially comorbid with other personality disor-
ders as well as Axis I disorders (Zanarini et al., 1999). Not
surprisingly, patients with BPD are notoriously difficult to treat.
Patients with BPD use higher levels of services in emergency
rooms, day hospital and partial hospitalization programs, and
outpatient clinics and inpatient units, and these are often used in
chaotic ways with repeated patterns of dropout, erratic psychother-
apy attendance, refusal to take medications as prescribed, and
pervasive noncompliance (Bongar, Peterson, Golann, & Hardi-
man, 1990; Zanarini & Frankenburg, 2001). Given these facts,
BPD is clearly a major public health problem that is prevalent,
painful, debilitating, and deadly.
From its inception, John Bowlby conceptualized attachment
theory in terms of both normal and psychopathological develop-
ment. Bowlby (1977) believed that attachment difficulties increase
vulnerability to psychopathology and can help identify the specific
types of psychological difficulties that arise. Bowlby (1977) con-
tended that internal working models of attachment help explain
“the many forms of emotional distress and personality distur-
bances, including anxiety, anger, depression, and emotional de-
tachment, to which unwilling separations and loss give rise” (p.
201). He held that childhood attachment underlies the “later ca-
pacity to make affectional bonds as well as a whole range of adult
dysfunctions,” including “marital problems and trouble with chil-
dren, as well as . . . neurotic symptoms and personality disorders”
(p. 206). Thus, Bowlby postulated that early attachment experi-
ences have long-lasting effects that tend to persist across the life
span and are among the major determinates of personality organi-
On the basis of Bowlby’s attachment theory, Ainsworth, Blehar,
Waters, and Wall’s (1978) seminal study identified three major
styles of attachment in infancy—secure, avoidant, and anxious–
ambivalent—and traced these styles to caregivers’ parenting be-
havior. Subsequent longitudinal studies investigating the influence
of infant attachment styles on later functioning and adaptation
have found a remarkable stability of attachment classification
(Hamilton, 2000; Waters, Merrick, Treboux, Crowell, & Alber-
sheim, 2000; Weinfield, Sroufe, & Egeland, 2000), although this
stability is partially mediated by later life experiences (Lewis,
Feiring, & Rosenthal, 2000; Waters et al., 2000).
Measurement of Attachment
From the seminal work of Bowlby, attachment theory and
research has evolved into two traditions, each with its own meth-
odology for assessing attachment patterns (e.g., self-report and
interview). Mary Main and her colleagues (Main, Kaplan, &
Cassidy, 1985; Main & Goodwyn, in press) developed the Adult
Attachment Interview (AAI), a 1-hr attachment-history interview.
The interview (Main et al., 1985) inquires about early attachment
relationships as well as the interviewee’s sense of how these
experiences affected adult personality by probing for specific
memories that may corroborate or contradict the quality of attach-
ment history presented by the interviewee. Noting the discourse
features in the interviews, Main and colleagues identified three
major patterns of adult attachment: secure/autonomous (F), dis-
missing (D), and enmeshed/preoccupied (E); and more recently,
two additional categories have been identified: unresolved/disor-
ganized (U/d) and cannot classify (CC). The first three categories
parallel the attachment classifications originally identified in child-
hood (Ainsworth et al., 1978), the disorganized classification par-
allels a pattern Main later described in infants (Main & Weston,
1981), and cannot classify parallels the classification from Hesse
(1996). These attachment patterns in adults reliably predicted the
Strange Situation behavior of their children.
Security on the AAI is characterized by a well-organized, un-
defended discourse style in which emotions are freely expressed
and by a high degree of coherence, are exhibited in the discussion
of attachment relationships, regardless of how positively or nega-
tively these experiences are portrayed. These individuals maintain
a balanced and realistic-seeming view of early relationships, value
attachment relationships, and view attachment-related experiences
as influential to their development.
LEVY ET AL.
In contrast, dismissing individuals devalue the importance of
attachment relationships or portray them in an idealized fashion
with few corroborating concrete examples. They have difficulty
recalling specific events from their past and usually describe an
early history of rejection. These individuals are judged to have low
coherence of mind because of the vagueness and sparseness of
their descriptions as well as the inconsistency between vaguely
positive generalizations and “leaked” evidence to the contrary.
Preoccupied individuals have little difficulty talking about at-
tachment and expressing attachment-related feelings. However,
these individuals tend to display confusion about past experiences
and are unable to gain insight into early events. They often
describe early relationships with parents as overinvolved or as
guilt inducing. Descriptions of their current relationship with par-
ents are often characterized by pervasive anger, passivity, and
attempts to please parents, even when they describe the relation-
ship as positive. Perhaps of most noted importance, preoccupied
individuals have a tendency toward incoherence in their descrip-
tions. Specifically, their interviews are often excessively long and
are characterized by the use of lengthy, grammatically entangled
sentences, jargon and nonsense words, reversions to childlike
speech, and confusion regarding past and present relationships.
Preoccupied responses often fail to address the interviewer’s orig-
The unresolved/disorganized classification is assigned when an
individual displays lapses in the monitoring of reasoning or dis-
course when discussing experiences of loss and abuse. These
lapses include highly implausible statements regarding the causes
and consequences of traumatic attachment-related events, loss of
memory for attachment-related traumas, and confusion and silence
around discussion of trauma or loss.
Main’s fifth classification, cannot classify, is assigned when an
individual displays a combination of contradictory or incompatible
attachment patterns or when no single state of mind with respect to
attachment is predominant. This occurs when the patient shifts
attachment patterns in midinterview, when the patient demon-
strates different attachment patterns with different attachment fig-
ures, or when the patient shows a mixture of different attachment
patterns within the same transcript or passage.
The Emergence of Mentalization and Reflective Function
Over the last decade, the social-cognitive and affect concept of
mentalization has become increasingly important to theory in
psychoanalysis and for the conceptualization of the development
of BPD. Fonagy and colleagues coined the term mentalization
(Fonagy et al., 2002; Fonagy & Target, 1996) to describe the
developmental achievement whereby children acquire the capacity
to interpret or make sense of behavior in oneself and others in
terms of intentional mental states such as thoughts, feelings, and
beliefs. Thus, mentalization is the capacity to evoke and reflect on
one’s own experience to make inferences about behavior in oneself
and others. Drawing from developmental theory and research,
Fonagy and colleagues contend that the capacity for mentalization
is dependent on the quality of interpersonal interactions and the
emotional relationship between the infant and caregivers “who are
sufficiently benign and reflective” (Fonagy & Target, 1996, p.
RF has been operationalized by Fonagy and colleagues to eval-
uate the quality of mentalization in the context of attachment
relationships (Fonagy et al., 1995), and initial research using the
RF scale has been promising. In a study examining the role of the
parents’ mentalizing skills and its relation to their infant’s attach-
ment pattern, Fonagy et al. (1995) found that RF mediated the
relationship between parental attachment security and infant at-
tachment security in the Strange Situation (Ainsworth et al., 1978)
at 1 year and at 18 months. That is, insecurely attached parents
with high RF were more likely to have securely attached babies
than insecurely attached parents with low RF.
Consistent with this finding, Slade and colleagues (Grienen-
berger, Kelly, & Slade, 2005) recently have shown that a mother’s
RF mediates the relationship between atypical maternal behaviors
(e.g., affective communication errors, role/boundary confusion,
intrusiveness) and attachment security in their infants. Fonagy et
al. (1996) examined the interaction of abuse and RF in a large
sample of psychiatric inpatients. They found that among patients
reporting abuse, those who scored low on RF were more likely to
be diagnosed with BPD compared with those who were abused but
scored high on RF. Thus, high RF seems to be a possible buffer
against the development of BPD in individuals who have experi-
Psychotherapy With BPD
Several psychotherapy studies have reported evidence for the
efficacy (Bateman & Fonagy, 1999; Clarkin, Levy, Lenzenweger,
& Kernberg, 2006; Giesen-Bloo et al., 2006; Koons et al., 2001;
Linehan, Armstrong, Suarez, Allmon, & Heard, 1991; Linehan,
Kanter, & Comtois, 1999; Linehan et al., 2002; Linehan, Schmidt,
et al., 1999; Turner, 2000; Verheul et al., 2003) and effectiveness
(Blum, Pfohl, & St. John, 2002; Brown, Newman, & Charles-
worth, 2004; Clarkin et al., 2001; Ryle & Golynkina, 2000;
Stevenson & Meares, 1992) of specific treatments for patients with
BPD. Furthermore, studies testing the effectiveness and efficacy of
new treatments have recently been completed (Gratz & Gunder-
son, 2006) or are currently underway (Markowitz, Skodol,
Bleiberg, & Strasser-Vorus, 2004).
One such treatment that has garnered effectiveness and efficacy
data in clinical trials is transference-focused psychotherapy (TFP;
Clarkin, Yeomans, & Kernberg, 1999, Clarkin et al., 2001, Clar-
kin, Yeomans, & Kernberg, 2006), a highly structured, twice-
weekly modified psychodynamic treatment based on Kernberg’s
(1984) object relations model of BPD. Recent studies have dem-
onstrated TFP’s effectiveness using patients as their own control
participants (Clarkin et al., 2001) and in comparison with a
treatment-as-usual group diagnosed with BPD (Levy, Clarkin,
Foelsch, & Kernberg, 2006). In addition, a randomized control
trial (Clarkin, Levy, et al., 2006) comparing TFP, dialectical be-
havior therapy (DBT), and modified psychodynamic supportive
psychotherapy (SPT) found reduced suicidality and anger in pa-
tients treated with TFP and DBT but not in those treated with SPT.
Whereas all three treatments were effective in reducing depression
and anxiety and in improving global functioning and social adjust-
ment, only TFP was consistently related to reductions in aggres-
sion (Clarkin, Levy, et. al., 2006).
What is becoming clear is that although BPD is a chronic
problem functionally, it is also a treatable disorder (see Leichsen-
SPECIAL SECTION: ATTACHMENT, REFLECTIVENESS, AND BPD
ring & Leibing, 2003; Oldham et al., 2001; Perry, Banon, & Ianni,
1999, for reviews). What remains uncertain, however, are the
mechanisms in the development and maintenance of BPD, the
processes of change within patients during treatment, and the
specific therapeutic techniques that bring about such changes.
Therefore, despite the support for the effectiveness and even
efficacy of existing treatments for BPD, researchers are still con-
fronted with a high degree of uncertainty about the underlying
processes of change.
Like other theories of BPD (e.g., Bateman & Fonagy, 2003;
Linehan, 1993), TFP includes the conceptualization of the basic
etiological elements of BPD as an interaction between constitu-
tional and environmental factors that results in a personality struc-
ture or organization characterized by identity disturbance, use of
immature or low level defense mechanisms such as projective
identification,1splitting, omnipotent control (i.e., trying to control
the behavior of others, often subtly, although usually feeling as if
others are trying to control them), and deficits in social reality
testing (i.e., difficulty differentiating one’s own thoughts from
another’s or difficulty perceiving subtle social cues correctly,
which often results in transient paranoia and fears of abandonment;
perceptual reality testing is generally maintained).
Regarding the interaction between biological constitution and
environment, Kernberg (1984) posited that patients with BPD have
difficulty integrating disparate representations of themselves and
others, in part because negative emotions, particularly aggression,
disrupt one’s capacity to integrate these representations. These
unintegrated disparate representations result in what Kernberg
refers to as part or partial representations; that is, failing to evoke
the complete representation but instead only part of it. Strong
unmetabolized or unprocessed emotions have the capacity to over-
whelm positive representations. Kernberg hypothesized that the
individual, therefore, may be unconsciously motivated to keep
these representations separate or split in an effort to protect the
positive representations of themselves and others (or some com-
bination of self and other representations). These high levels of
negative emotionality and aggression can be constitutional or
engendered through experience, or some combination of the two.
Regardless of origin, high levels of aggression interfere with the
normative developmental process of integrating disparate repre-
sentations, and instead the high levels of aggression result in a
division between positive and negative representations. Likewise,
Siever and his colleagues (Gurvits, Koenigsberg, & Siever, 2000)
pointed out that affective instability may interfere with the ability
to develop stable perceptions of self and others. They note that
both the specific role of aggression and the more general role of
affective lability may make the developmental task of integrating
stable representations of self and others more difficult to accom-
plish. However, Kernberg and colleagues (Clarkin, Yeomans, &
Kernberg, 2006) also noted that emotional instability in BPD can
be secondary to a lack of differentiation and integration of internal
images of self and others, which leads to instability in one’s sense
of self and ultimately, in affective instability. Thus, the relation-
ship between lack of integration of representations and affective
instability may operate in a vicious circle with the intensity of
early affects resulting in a split experience of self and others.
According to Kernberg, positive and negative representations are
split to protect positive representations, however, such splitting
may lead to further affective instability by failing to provide a
sufficiently complete and accurate foundation from which to un-
derstand oneself and others.
As the patient progresses during the course of TFP moving from
split-off contradictory self-states to increased reflectiveness and
integration, from impulsive action to active reflection, the patient
develops better behavioral control. Over time, increased differ-
entiation and integration is theorized to allow patients with
BPD to think more flexibly and benevolently about the mental
states (or motives and intentions) of their therapist, significant
others (e.g., attachment figures), and themselves. With increased
differentiation–integration, impaired and distorted representations
of self and others are gradually transformed through new experi-
ence with significant others, beginning with the therapist. Over
time, this shift in attachment organization and RF is theorized to
assist patients in developing intimate relations that are infused with
less aggression, greater capacity for intimacy, increased coherence
of identity, and decreased self-defeating and destructive behaviors,
as well as general improvements in symptoms and functioning.
The Present Study
In the present study, we examined changes in attachment orga-
nization as measured by the AAI and changes in RF as measured
by the RF coding scale as a function of one of 3 year-long
intensive psychotherapies designed specifically for patients with
BPD. We hypothesized that the transference-focused psychother-
apy, as compared with DBT and SPT, will significantly increase
RF and narrative coherence and significantly reduce lack of reso-
lution of loss and trauma.
Patients with BPD were recruited between November 1999 and July
2002 from within a 50-mile radius of New York City. Patients were
referred by private practitioners, clinics, family members, and self-referral,
although 97% were referred by mental health professionals. Participants
were 90 adults (6 men and 84 women) between age 18 years and 50 years.
Patients with comorbid schizophrenia, schizoaffective disorder, bipolar I
disorder, delusional disorder, and/or delirium, dementia, and amnesia and
other cognitive disorders were excluded because of the influence of brain
pathology and thought disorder on the ability to provide meaningful
self-report data and complicated response to treatment. At the time that
participants were invited to participate in the study, written informed
consent was obtained after all study procedures had been explained. The
study was approved by the human participant institutional review board.
Of the 207 individuals clinically referred and interviewed for at least one
evaluation session, 109 were eligible for randomization. Most exclusions
were due to the absence of five criteria for BPD (n ? 34). Many patients
were excluded because of age (n ? 30) or because they met criteria for
1Filled with an intolerable idea or feeling about the self, a person acts
in a way to provoke behavior in another. This allows them to feel justified
that the other person is in fact this way and not them. For example, a person
filled with rage acts so to enrage another person and then feels relieved of
their own anger when they see the other person become angry.
LEVY ET AL.
current substance dependence2(n ? 9), schizophrenia or a schizophrenic
disorder (n ? 8), or bipolar I disorder (n ? 6). Patients were also excluded
following dropout from the evaluation process (n ? 8), IQ lower than 80
(n ? 2), and scheduling conflict (n ? 1). Of the 109 eligible for random-
ization, 90 (83%) were randomized to treatment. There were no differences
in terms of demographics, diagnostic data, and severity of psychopathology
between those randomized to treatment and those not (Levy, Critchfield, &
Clarkin, 2005). See Clarkin and colleagues (Clarkin, Levy, Lenzenweger,
& Kernberg, 2004, Clarkin, Levy, et al., 2006) for a full description of the
Treatments, Therapists, and Hypothesized Mechanisms of
TFP is a modified manualized psychodynamic treatment for patients
with BPD. The primary goal of TFP is to reduce symptomatology and
self-destructive behavior through the modification of representations of
self and others as they are enacted in the treatment (Clarkin et al., 1999,
Clarkin, Yeomans, & Kernberg, 2006; Kernberg, Selzer, Koenigsberg,
Carr, & Appelbaum, 1989). TFP is a highly structured, twice-weekly
treatment for 45 min per session that begins with explicit contract setting,
which clarifies the conditions of therapy, the method of treatment, hierar-
chy of target behaviors to be addressed during therapy sessions, the
respective roles of patient and therapist during the treatment, and how
suicidal urges and behaviors will be managed. The primary focus of TFP
is on the predominant affect-laden themes that emerge in the relationship
between patients who are borderline and their therapists in the here and
now of the transference. During the first year of treatment, TFP focuses on
the containment of acting out (parasuicidal) behaviors and the identifica-
tion and recapitulation of dominant relational patterns as they are experi-
enced and expressed in the here and now of the transference relationship.
The therapist uses techniques of clarification, confrontation, and transfer-
ence interpretation (that is, interpretation of the here and now patient–
therapist interactions that demonstrate the patient’s disparate perceptions of
self and others, including the therapist). In TFP, interpretation is viewed as
the route to integration of these disparate perceptions and representations.
DBT is a manualized cognitive–behavioral treatment with two compo-
nents: (a) individual therapy and (b) group skills training. The individual
treatment focuses on a hierarchy of target behaviors, which the patient
tracks on a daily basis with diary cards. Suicidal and self-mutilating
behaviors are at the top of the hierarchy and are examined in each session.
Behavioral analyses of the pattern and chain of thoughts, emotions, and
events resulting in suicidal and self-mutilating acts take place routinely to
help the patient identify triggers and alternative strategies for coping.
Change strategies such as problem solving and reinforcement techniques
are used in combination with acceptance and validation of the patient’s
experience. Group skills training is used to help patients develop less
self-destructive and more adaptive means of coping with intolerable af-
fects. Skills training sessions consist of teaching new skills to patients and
practicing these skills through specific assignments between sessions.
These skills include awareness of emotions and reactions, interpersonal
effectiveness, emotion regulation, and distress tolerance. The skills are then
integrated into the individual treatment when problem situations, such as
suicidal urges, present themselves. Therapists help patients identify appro-
priate skills to use instead of maladaptive coping strategies. Eventually,
patients can apply the skills on their own and develop more adaptive means
of functioning. Therapists are available to patients via pagers between
sessions for brief coaching to help patients apply appropriate skillful means
of coping with their emotions and stresses and to fight self-injurious
impulses. Individual therapy is provided once weekly for 60 min, and skills
training is provided weekly for 2.5 hr. Emergency telephone contact and
individual sessions are scheduled as needed.
SPT is a manualized psychoanalytically oriented treatment for border-
line patients (Appelbaum, 2005) adapted from one of the most common
SPT treatments (Rockland, 1992). SPT is a once or twice weekly treat-
ment3for 45 min per session with the primary goal of bringing about
changes through developing a healthy collaborative relationship with the
therapist and to replace self-destructive enactments with verbal expression
of conflicts. This transformation is thought to occur through the patient’s
identification with the reflective capacities of the therapist rather than
through interpretation as in TFP. Like TFP and DBT, SPT begins with a
contract setting phase, and the initial stages of treatment address behaviors
that threaten the patient’s safety, interfere with therapy, and disrupt the
patient’s psychosocial functioning. In addition, the initial phases of therapy
focus on fostering an atmosphere of safety and security for the patient and
a sense of collaboration between patient and therapist. The following
techniques are used in SPT: (a) remaining attuned to the dominant affect
and type of the transference without interpreting it; (b) accepting and using
the positive transference; (c) translating the patient’s tendency to act out
feelings into a more adaptive verbal expressive mode, which enhances
self-acceptance and self-awareness; (d) identifying and describing the
significant aspects of self to strengthen identity and sense of self-cohesion
and continuity; (e) limiting extreme dependency (inaction) by providing
cognitive support and by fostering the patient’s sense of self and agency;
(f) providing emotional support (encouragement, praise, reassurance, in-
spiring hope, expressing concern); (g) providing direct environmental
intervention when necessary; (h) offering advice and suggestions (indirect
environmental support); (i) supporting mastery of impulse and affect; and
(j) encouraging sublimations and socially acceptable modes of impulse
expression (e.g., exercise, physical sports). The SPT group was conceptu-
alized as a component control condition, with the intended active ingredi-
ent (transference interpretation) proscribed.
Therapists in each of the three treatment conditions were selected on the
basis of prior demonstration of competence in their respective treatment.
To ensure ongoing therapist adherence and competence, experts supervised
all treatments on a weekly basis. Barbara Stanley, an acknowledged expert
in DBT and a National Institute of Mental Health funded researcher in this
area, supervised DBT therapists. Frank E. Yeomans, an expert therapist in
TFP and contributor to the clinical literature on BPD and TFP, supervised
TFP therapists. Ann Appelbaum, expert therapist and contributor to the
clinical literature on BPD and SPT techniques, supervised the supportive
treatment (Appelbaum, 1994, 1996, 2005). Prior to being assigned patients,
all therapists selected for the study were judged by treatment cell leaders to
be both adherent to their respective manual and competent in using the
specific techniques of their respective modality. Throughout the study, all
therapists regularly videotaped their sessions and were supervised in a
group on a weekly basis.
The TFP therapists were 8 experienced individuals
with postdoctoral training. Experience level ranged from faculty/staff psy-
2Patients with current substance dependence were referred to substance
dependence treatment but were eligible for randomization after the sub-
stance dependence was treated or resolved.
3Therapists typically see patients once weekly; however, many see
patients twice weekly if clinically indicated, if there is an emergency, or to
make up a missed session from the prior week.
SPECIAL SECTION: ATTACHMENT, REFLECTIVENESS, AND BPD
chiatrists with at least 10 years of experience to faculty/staff psychologists
with at least 2 years of experience treating patients with BPD as well as
specific training in TFP.
The DBT therapists were 5 experienced individuals
with postdoctoral training. Experience level ranged from faculty/staff psy-
chologist with 10 years of experience to faculty/staff psychologists with at
least 2 years of experience treating patients with BPD as well as specific
training in DBT (all therapists had attended multiple intensive trainings
with Linehan or other certified trainers).
The SPT therapists were 7 experienced individuals
with postdoctoral training. Experience level ranged from faculty/staff psy-
chiatrists with at least 15 years of experience to faculty/staff psychologists
with at least 2 years of experience treating patients with BPD as well as
specific training in SPT.
Monitoring Treatment Conditions
Treatment integrity was monitored in a number of ways. First, we chose
experienced and expert treatment cell leaders who were responsible for
recruiting therapists for their respective treatment cell. Therapists were
known to the treatment cell leaders and chosen because they were expe-
rienced and adherent therapists with proven track records. Second, thera-
pists in each treatment cell attended weekly group supervisions in which
treatment cell leaders were able to observe videotaped sessions. Additional
feedback to therapists was provided by treatment cell leaders if a therapist
fell below an acceptable level of either adherence to the manual or
competence. When a therapist’s ratings were consistently low for adher-
ence, then ratings were made more frequently (approximately every four
sessions) for the succeeding 3-month interval, and supervision focused on
the difficulties identified by raters. Additional individual supervision was
provided when either adherence and/or competence fell below acceptable
levels. When a therapist fell below acceptable levels, no new cases were
assigned to them. Third, we asked treatment cell leaders to rate and rank
therapists on each case.
Ninety participants were randomized into TFP, SPT, or DBT. Simple
randomization was used to guard against unseen threats to validity rather
than stratified or minimization random assignment procedures that match
participants on prognostic variables. A study-independent person generated
the allocation sequence and assigned participants to their groups. Initial
assessments were made after inclusion and before randomization.
Randomization resulted in 31 patients randomized to TFP, 29 patients
randomized to DBT, and 30 patients randomized to SPT. Of the patients,
1 in TFP was removed early on from the study because of misdiagnosis
when it became apparent that she had a psychotic disorder. In addition, 1
patient was removed from DBT group because of the individual withdraw-
ing from the study prior to attending the first therapy session.
Diagnostic and Borderline Symptom Instruments
Structured Clinical Interview for DSM–IV–Research Version (First,
Gibbon, Spitzer, & Williams, 1997).
for DSM–IV–Research Version is a structured clinical interview used for
making DSM–IV Axis I diagnoses in patients older than 18 years.
International Personality Disorder Examination (Loranger, Sartorius,
Andreoli, & Berger, 1994).
The International Personality Disorder Ex-
amination is a semistructured diagnostic interview for diagnosing person-
ality disorders. It consists of 99 items arranged in six categories (e.g., self
or work), along with a detailed scoring manual (Loranger et al., 1994).
Each item assesses part or all of a DSM–IV personality disorder criterion
and is rated on a three-point scale ranging between 0 (absent or normal),
The Structured Clinical Interview
1 (exaggerated or accentuated), and 2 (meets criteria or pathological).
Items consist of one or several primary questions and follow-up questions.
All positive responses are followed by requests for examples. After the
provided questions are exhausted, the clinical interviewer is free to ask
additional questions until he or she is able to score the item. The Interna-
tional Personality Disorder Examination generates probable (subthreshold
number of DSM–IV criteria met) and definite diagnoses for each of the
DSM–IV diagnoses. It also generates dimensional scores for each diagnosis
by adding the ratings on all the criteria composing a diagnosis.
Reliability of assessment interviews indicated good to excellent levels of
interrater reliability for all Axis I and II disorders with kappas ranging from
.59 for anxiety disorders to 1.00 for alcohol/substance dependence. The
kappa for BPD was .64 and the intraclass correlation (ICC) for dimensional
criteria ratings was .86. All kappa and ICC coefficients were in the good to
excellent range (Fleiss, 1971). See Critchfield, Levy, and Clarkin (in press)
for more detail regarding diagnostic interviewers, interviewer credentials
and training, and reliability procedures.
Assessment of Attachment
AAI (George, Kaplan, & Main, 1985).
clinical interview designed to elicit thoughts, feelings, and memories about
early attachment experiences and to assess the individual’s state of mind or
internal working model with regard to early attachment relationships. The
interview consists of 20 questions asked in a set order with standard probes.
Individuals are asked to describe their childhood relationship with their
parents, choosing five adjectives to describe each relationship and support-
ing these descriptors with specific memories. To elicit attachment-related
information, they are asked how their parents responded to them when they
were in physical or emotional distress (e.g., during times when they were
upset, injured, and sick as children). They are also asked about memories
of separations, loss, experiences of rejection, and times when they might
have felt threatened including, but not limited to, those involving physical
and sexual abuse. The interview requires that they reflect on their parents’
styles of parenting and that they consider how their childhood experiences
with their parents have influenced their lives. The technique has been
described as having the effect of “surprising the unconscious” (George et
al., 1985, p. 3) and allowing numerous opportunities for the interviewee to
elaborate on, contradict, or fail to support previous statements.
The AAI is transcribed verbatim, and trained coders first score the
transcripts with subscales ratings, which are then used to assign individuals
to one of five primary attachment classifications (secure/autonomous,
dismissive, preoccupied, unresolved, and cannot classify). The unresolved
classification can be a primary or secondary designation, and a patient
classified as unresolved is always given an additional organized style. In
addition to attachment classification, we focused on the narrative coher-
ence subscale of the AAI, which has been found to be the best predictor of
attachment security, r ? .96, p ? .001 (Waters, Treboux, Fyffe, & Crowell,
2001). The AAI is administered and scored by raters who have completed
a 2-week training workshop conducted by Mary Main and Eric Hesse and
who have achieved reliability on an extensive set of training transcripts.
Raters are blind to all identifying characteristics of the participants, includ-
ing attachment status and the nature and purpose of the study. After
training was completed and reliability was established, the coders coded a
subset of each other’s transcripts (n ? 22). Raters agreed on 86% of the
categorical classifications, ? ? .80, t(20) ? 6.11, p ? .001. The ICC for
dimensional ratings of narrative coherence was .88
The AAI was also scored with the RF scale (Fonagy, Steele,
Steele, & Target, 1998), an 11-point scale that evaluates the quality of
mentalization in the context of attachment relationships. The RF scale
ranges from –1 (negative RF, in which interviews are overly concrete,
totally barren of mentalization, or grossly distorting of the mental states of
others) to 9 (exceptional RF, in which interviews show unusually complex,
elaborate, or original reasoning about mental states). Coders were trained
by Kenneth N. Levy, who had received training from the developers of the
The AAI is a semistructured
LEVY ET AL.
coding manual. Reliability was obtained between the coders and one of the
developers of the coding manual on practice sets. After training was
completed and reliability was established, the two coders coded a subset of
each other’s transcripts (n ? 28, ICC ? .86). As with the AAI coding,
coders were blind to both time and treatment condition.
In the original outcome report (Clarkin, Levy, et al., 2006)
examining the efficacy of TFP in this sample, we reported intent-
to-treat (ITT) analyses for all primary and secondary outcome
variables. However, in the current article we only report completer
analyses because the goal of the study was to identify mechanism
of change rather than to test the efficacy of the treatment. ITT
analyses are less relevant for the study of mechanisms of change,
in which it is more important ensure that participants receive a
sufficient dose of the treatment and to guard against threats to
validity from insufficient or diluted doses of treatments.4
Distribution of Attachment Patterns at Time 1
As shown in Table 1, with the five-category system (i.e., secure,
preoccupied, dismissing, unresolved, and cannot classify), we
found 5% of our patients were classified as secure with respect to
attachment. This rate is similar to other studies examining attach-
ment patterns in samples of patients with BPD (Barone, 2003;
Diamond et al., 1999; Fonagy et al. 1996; Stovall-McClough &
Cloitre, 2003).5With regard to the insecure attachment patterns,
we found that 31.7% of patients were classified as unresolved with
respect to attachment, and 18.3% were classified in the cannot
classify category. Thirty percent of patients were classified as
dismissive with respect to attachment, and 15% of patients were
classified as preoccupied with respect to attachment. Table 1 also
shows the three-way classifications of attachment patterns on the
basis of secondary classifications. By using the secondary classi-
fications, we found that 50% of patients can be classified into the
preoccupied category and 45% into the dismissing category. These
findings represent a more even distribution between the different
attachment categories than was found in two earlier studies (Fon-
agy et al., 1996; Patrick, Hobson, Castle, Howard, & Maughan,
1994), both of which found that patients with BPD predominately
were classified as unresolved and preoccupied with respect to
attachment. Our patterns of findings are more consistent with those
of Barone (2003), who also found a more even distribution of
attachment patterns in a sample of outpatients diagnosed with
BPD. It is interesting to note that we found that secondary classi-
fications for patients in the cannot classify group were mostly
dismissive and for unresolved patients were mostly preoccupied,
with all but 3 of the 11 patients in the cannot classify group having
a secondary best-fitting classification of dismissive and all but 2 of
the 19 unresolved patients having a secondary best-fitting classi-
fication of preoccupied. Chi-square analyses indicated that these
associations are significant.
Change in the Distribution of Attachment Patterns
Between Time 1 and Time 2
As shown in Table 2, there were significant changes in attach-
ment patterns during the course of the year of psychotherapy.
Overall, there were increases in the percentage of patients now
classified as securely attached. Whereas at Time 1 there were only
3 patients (5%) classifiable as secure with respect to attachment, at
Time 2 there was a threefold increase in attachment security, with
9 patients (15%) now classifiable as secure with respect to attach-
ment. This difference was significant, McNemar’s ?2(1, N ? 60)
? 34.03, p ? .001. Although there was a decrease in the number
of patients classified with unresolved attachment (n ? 19, 31.6%,
vs. n ? 13, 21.6%), this decrease was nonsignificant, McNemar’s
?2(1, N ? 60) ? 0.90, p ? .05.
4Noncompliance and dropout in the treatment of BPD is common, and
therefore ITT analysis may underestimate the real benefits of treatment.
Completer analyses are also especially important if there is differential
dropout. A number of recent studies that had differential dropout have
neglected to report completer analyses (Giesen-Bloo et al., 2006; Linehan
et al., 2006), which limits conclusions about relative efficacy (Levy, 2006).
5Although this finding is consistent across a number of studies, it is
obviously particular that one can be diagnosed with BPD, a disorder
characterized by extreme insecurity, and rated as securely attached on the
AAI. There are a number of possibilities. First, the finding may simply
represent measurement error. On the AAI, if one can describe experiences
of insecurity or behaving insecurely in a coherent manner, all other things
equal, that individual would most likely be classified as securely attached.
Those borderline patients classified as securely attached often were coher-
ent, albeit moderately, in describing what could be classified as insecure
behaviors in relationships with others.
Distribution of Attachment Patterns With the Five-Way Classification System at Time 1
Five categoryFour categoryThree category
SPECIAL SECTION: ATTACHMENT, REFLECTIVENESS, AND BPD
Change in the Distribution of Attachment Patterns
Between Time 1 and Time 2 as a Function of Treatment
As shown in Table 3, we also examined change in attachment
classification as a function of treatment group. In the TFP group,
1 (4.5%) of the 22 patients was classified as secure with respect to
attachment at Time 1; however, 7 (31.8%) of the 22 patients were
classified as secure with respect to attachment at Time 2. This
difference was significant, McNemar’s ?2(1, N ? 22) ? 4.17, p ?
.04. There was no change, however, for both DBT and SPT in the
number of patients who went from insecure to secure attachment.
One patient in each treatment (6.3% and 4.5%, respectively) was
classified as securely attached at both Times 1 and 2. The differ-
ence in the number of securely attached patients between the TFP
and DBT and SPT cells was obviously not significant at Time 1,
?2(2, N ? 60) ? 1.46, p ? .05, but was significant at the end of
treatment, ?2(2, N ? 60) ? 8.25, p ? .02.
Relationship Between Attachment Coherence, RF, and
Lack of Resolution of Loss and Trauma
Table 4 shows the relationship between the dimensional
ratings of attachment coherence, RF, and lack of resolution of
loss and trauma. As can be seen, Time 1 coherence and RF were
significantly positively correlated at a moderate level, suggest-
ing that the two constructs are related but are not necessarily
measuring the same construct. The correlation between coher-
ence and RF at Time 2 was also significant and in the same
direction and of about the same magnitude. It is important to
point out that although the two constructs are significantly
associated, the magnitude of the correlation is much less than in
previous studies, in which correlations as high as .73 were
found (Fonagy, Steele, Moran, Steele, & Higgitt, 1991). Coher-
ence was significantly related to lack of resolution of loss and
trauma; however, RF was not. This finding is in contrast to a
previous finding that found high RF was related to higher
resolution of trauma (Fonagy et al., 1996). This previous study
involved severely disturbed inpatients with high rates of severe
trauma and loss; it may be that in cases of severe trauma the
buffering effects of high RF become more apparent.
Change in RF, Coherence, and Lack of Resolution of
Loss and Trauma From Time 1 to Time 2
With respect to attachment coherence, the covariate, Time 1
coherence, was significantly related to participants’ Time 2
coherence, F(1, 54) ? 32.94, p ? .05, r ? .97. There was also
a significant effect of treatment group on Time 2 coherence
after the effects of Time 1 coherence were controlled for, F(3,
54) ? 6.28, p ? .05, r ? .64. Planned contrasts (TFP ? 2,
DBT ? –1, SPT ? –1) revealed that those in the TFP condition
significantly increased coherence compared with both the DBT,
t(54) ? 2.06, p ? .05, r ? .26, and SPT groups, t(54) ? 2.72,
p ? .05, r ? .34. Means and standard deviations are shown in
Association Between Attachment Patterns at Time 1 (T1) and Time 2 (T2)
SecurePreoccupiedDismissive UnresolvedCannot classifyTotal
?2(16, N ? 60)
100.0 181911 60
****p ? .001.
Association Between Attachment Security Time 1 and Time 2 as a Function of Treatment Group
Time 2 attachment
Time 1 attachment
SecureInsecure SecureInsecure SecureInsecure
psychotherapy. Percentages are for the columns.
**p ? .02.
TFP ? transference-focused psychotherapy; DBT ? dialectical behavior therapy; SPT ? supportive
LEVY ET AL.
With respect to RF, the covariate, Time 1 RF, was signifi-
cantly related to the participants’ Time 2 RF, F(1, 54) ? 7.23,
p ? .01, r ? .69. There was also a significant effect of treatment
group on Time 2 RF after the effects of Time 1 RF were
controlled for, F(3, 54) ? 15.05, p ? .05, r ? .89. Planned
contrasts (TFP ? 2, DBT ? –1, SPT ? –1) revealed that those
in the TFP condition significantly increased RF compared with
both the DBT, t(54) ? 2.10, p ? .05, r ? .27, and SPT groups,
t(54) ? 3.24, p ? .05, r ? .39. Means and standard deviations
are shown in Table 5.
With respect to lack of resolution of loss, Time 1 lack of
resolution of loss was significantly related to the participants’
Time 2 scores, F(1, 54) ? 63.50, p ? .01, r ? .99. However, the
effect of treatment group on Time 2 lack of resolution of loss
after the effects of Time 1 scores were controlled for, F(3,
54) ? 1.14, p ? .05, r ? .15, was not significant. With regard
to lack of resolution of trauma, Time 1 lack of resolution of
trauma, was significantly related to the participants’ Time 2
lack of resolution of trauma, F(1, 54) ? 43.05, p ? .01, r ? .98.
However, the effect of treatment group on Time 2 lack of
resolution of trauma after the effects of Time 1 lack of resolu-
tion of trauma were controlled for, F(3, 54) ? 1.34, p ? .05,
r ? .89, was not significant. Means and standard deviations are
shown in Table 5.
In addition to relevant domain specific outcomes (reported
elsewhere in Clarkin, Levy, et al., 2006), we examined changes in
attachment organization and RF during one of three (TFP, DBT, or
SPT) year-long intensive psychotherapies for patients diagnosed
with BPD. On the basis of prior theorizing, we conceptualized
these constructs as putative mechanisms of change within patients
with BPD (Fonagy et al., 2002; Levy, 2005; Levy, Clarkin, Yeo-
mans, et al., 2006).
RF, attachment coherence, and security of attachment were
found to increase over the year of treatment as a function of
treatment group. Specifically, we found that those patients
treated with TFP evidenced significant increases in RF, attach-
ment coherence, and rates of being classified as secure with
respect to attachment as compared with the other treatment
conditions. However, there were no significant changes in terms
of resolution of loss or trauma across treatment groups. There
were no adverse events or side effects in any of the intervention
groups. It is important to also note that our sample size was
relatively small, and this may have limited ability to defect
The findings from this study, coupled with our other work
(Clarkin, Levy, et al., 2006), show that TFP is not only an effica-
cious treatment for BPD but works in a theoretically predicted way
and that TFP does better on those variables than DBT and SPT.
Our findings are especially important given the literature suggest-
ing that many treatments do not show specific effects on theory-
driven mechanisms (Ablon & Jones, 1998; Ablon, Levy, & Ka-
zenstein, 2006; Castonguay, Goldfried, Wiser, & Raue, 1996;
DeRubeis & Feeley, 1990; DeRubeis et al., 1990; Ilardi & Craig-
head, 1994; Jones & Pulos, 1993; Shaw et al., 1999; Trepka, Rees,
Shapiro, Hardy, & Barkham, 2004).
Correlations Between Dimensional Measures of Attachment
1. RF T1
2. RF T2
3. Coherence T1
4. Coherence T2
5. Resolution of Loss T1
6. Resolution of Loss T2
7. Resolution of Trauma T1
8. Resolution of Trauma T2
.02.03 .01 .06
*p ? .05.
RF ? reflective function; T1 ? Time 1; T2 ? Time 2.
***p ? .01.
Change in RF, Coherence, and Lack of Resolution of Loss and Trauma From Time 1 to Time 2
TFP (N ? 22) DBT (N ? 15) SPT (N ? 23)
Time 1 Time 2Time 1 Time 2Time 1 Time 2
M SDM SDM SDM SDM SDM SD
Resolution of Loss
Resolution of Trauma
TFP ? transference-focused psychotherapy; DBT ? dialectical behavior therapy; SPT ? supportive psychotherapy; RF ? reflective function.
SPECIAL SECTION: ATTACHMENT, REFLECTIVENESS, AND BPD
These findings also have implications for available treatment
choices for patients with BPD. Although Linehan’s (Linehan et
al., 1991) seminal randomized clinical trail of DBT was a
breakthrough for the research on BPD, and DBT has marshaled
a large amount of evidence for its efficacy (more evidence than
any other treatment for BPD), clinicians now have two addi-
tional efficacious treatments available in Mentalization Based
Therapy (Bateman & Fonagy, 1999) and in TFP. Given the
heterogeneity seen in patients with BPD, future research might
focus on the prescriptive implications of client factors as well as
which patients might be most responsive to a particular treat-
ment (Clarkin & Levy, 2003).
The Meaning of Attachment Security in the Context of
One of the more provocative findings in our study was the
increased percentage of those patients classified as secure with
respect to attachment; we found a threefold increase in those
patients classified as securely attached. This finding is striking
because insecurity in interpersonal relationships is theorized to
be central to borderline pathology (Gunderson, 1996). This
finding raises the broader question of the meaning of security
with respect to attachment in this context; specifically, have this
subset of patients achieved secure attachment in the course of 1
year of psychotherapy? According to Bowlby, the hallmarks of
attachment security are secure base behavior (the use of an
attachment figure as home base from which the individual can
explore the world) and safe haven behavior (to seek support,
comfort, and protection from an attachment figure in the face of
danger or distress). Thus, if these patients with BPD were
secure with respect to attachment, there would be an accompa-
nying shift in their interactions with their attachment figures.
The individual would be able to turn to others as a safe haven
in times of distress though when not in distress, would use
attachment figures as a secure base from which to explore both
the physical and psychological world. Security with respect to
attachment in these patients would also accompany a shift in
their interactions with their own children, as the most consistent
finding with regard to security of attachment on the AAI is that
it predicts the interviewees’ infants’ Strange Situation behavior.
Clearly, some patients with BPD in our study were more
coherent with respect to their attachment representations during
the course of 1 year of psychotherapy. However, we do not have
data to speak to either of these interpersonal behavioral markers
of attachment security. Thus, future research will need to es-
tablish these individuals are evidencing both safe haven and
secure base behaviors that would indicate secure attachment. It
is possible that these individuals, despite increased narrative
coherence, continue to engage in self-destructive behaviors.
Regardless, over time it is predicted that this increased coher-
ence will result in better integration of experience, increased
flexibility of thought processes, and better self-regulation, and
this may put these patients on the road to eventually developing
secure attachment behaviors. Additionally, regardless of their
own secure-like behaviors, those patients showing increased
coherence and indicators of secure attachment organization may
be more likely to have children that evidence secure attachment
behavior in the Strange Situation. That is, the increased coher-
ence is predicted to result in more sensitive, responsive parent-
ing and less oscillations between intrusive and neglectful par-
enting. The AAI and the Strange Situation were developed and,
until relatively recently, used almost exclusively within non-
clinical samples. A finding in which patients treated for BPD,
and showing increased coherence as well as indicators of secure
attachment organization, have children that evidence secure
attachment behavior in the Strange Situation would provide
needed validity for both the AAI and Strange Situation proce-
Changes in RF
The changes observed in RF in this study represent a signif-
icant shift in patients’ capacity to mentalize the thoughts, feel-
ings, intentions, and desires of self and others. Patients in the
TFP group entered the study with a mean RF score of 2.86,
which is similar to findings from an earlier study examining RF
in a sample of participants with BPD (Fonagy et al., 1996).
According to Fonagy and colleagues (Fonagy et al., 1998), a
score of 3 on the RF scale is considered questionable or low and
indicates naive or simplistic reflections on the mental states of
self and others. A score of 3 on the scale may also indicate an
overanalytic or hyperactive style. In such cases, attributions
may appear at face value quite reflective and may show greater
depth than expected but, on closer inspection, may be diffuse
and unintegrated in a manner that does not create links between
disparate aspects of the individual’s experience or result in
increased understanding. Instead, these individuals tend to dis-
cuss internal thoughts and feelings with a certainty about oth-
ers’ mental states in a manner that belies the open, flexible, and
integrative processes found in high RF individuals (Fonagy et
al., 1991, 1996; Grienenberger et al., 2005). In our study, only
4 of the 22 TFP patients (15%) entered the study with a score
higher than 3 on the RF scale. As compared with the other
treatment conditions, patients in TFP showed a significant
increase over the course of treatment in RF, with a mean score
of 4.11 posttreatment (approaching ordinary or adequate RF)
and almost two thirds of the patients scoring 4 or better (72.7%,
with 31.8% scoring 5 or above). A score of 5 on the RF scale
is the most common rating in nonclinical samples and is char-
acterized by instances of reflective thinking in the context of
otherwise ordinary or adequate attributions of the thoughts,
feelings, intentions, and desires of self and others. Individuals
in this range may reflectively discuss internal thoughts and
feelings, but such thinking may not be exhibited spontaneously
and may be limited when tackling complex social situations in
which there is conflict and ambivalence. Those receiving scores
of 4 on the RF scale may have a model of the mind similar to
those who score a 5 but with less integration.
Resolution of Loss and Trauma
There was little change in patient unresolved/not unresolved
status during treatment (63% maintained their unresolved classi-
fication). This is less change than was found in an earlier study
examining stability of attachment representations in adults transi-
tioning to marriage (Crowell, Treboux, & Waters, 2002), with only
46% of participants initially classified as unresolved retaining the
LEVY ET AL.
classification. In our study, we also found that 7.3% of patients
with BPD who were not unresolved became unresolved over the
course of 1 year of treatment; a similar finding to Crowell et al.
(2002) who found that 9% of individuals who were not unresolved
become unresolved. Thus, although patients with BPD appear to
be at greater risk for unresolved status over the course of 1 year of
treatment, it does not appear that they are at any greater risk above
and beyond their initial risk of becoming unresolved over a 1-year
Nevertheless, our findings, if replicated, suggest that patients
with BPD may be particularly resistant to the resolution of loss and
trauma, even with intensive interventions. Additionally, both TFP
and DBT explicitly hold off on addressing trauma until later in
treatment because of the tendency of borderline patients to become
disorganized, dysregulated, and at times dissociated in the context
of discussing these issues. Further, TFP and DBT both deempha-
size past experiences in the service of understanding what is
occurring for the patient in the here and now. In may be the case
that because of this emphasis on the present there is less time spent
discussing past traumatic experiences as compared with SPT;
future research exploring treatment-specific mechanisms of
change should address this question.
The Relationship of RF to Lack of Resolution of Trauma
Unexpectedly, we found that RF was not significantly related to
lack of resolution of loss and trauma. This finding is surprising
given the centrality of both constructs for thinking about BPD
(Fonagy et al., 2002; Holmes, 2003, 2004; Liotti & Pasquini,
2000). Our finding suggests that low RF and lack of resolution of
trauma may operate relatively independently.
Studying Mechanisms of Change in the Psychotherapy for
Future research should seek to establish the hypothesized link
between increased narrative coherence and RF over the course of
TFP treatment and improvements in level of symptomatology and
global functioning. In addition, it will be important to identify the
psychotherapy processes related to these observed changes in RF
and narrative coherence. On the basis of Kernberg’s (1984) devel-
opmentally based theory of BPD, the hypothesized mechanism of
change in TFP stems from the integration of polarized affect states
and representations of self and other into a more coherent whole.
Through the exploration and integration of these “split-off”
cognitive–affective units of self- and other representations, Kern-
berg postulated that the patient develops the capacity to think more
coherently and reflectively, with more realistic, complex, and
differentiated appraisals of the thoughts, feelings, intentions, and
desires of self and others. This integration in psychological struc-
ture is hypothesized to allow for increased modulation of affect
and coherence of identity, a greater capacity for intimacy in
relationships, a reduction in self-destructive behaviors, and general
improvement in functioning.
Summary and Conclusion
In a sample of patients with BPD, we studied the differential
effect of TFP on changes in attachment organization, RF, and lack
of resolution of loss and trauma. Consistent with our hypotheses,
we found that TFP resulted in unique changes in narrative coher-
ence and RF not observed in other treatment conditions. However,
there was little support for an effect on lack of resolution of loss or
trauma. These findings have implications for conceptualizing the
mechanism by which patients with borderline personality may
change. Howard et al. (Howard, Kopta, Krause, & Orlinsky, 1986)
distinguished between three levels of change in psychotherapy:
remoralization, remediation, and rehabilitation. Remoralization is
characterized as the initial improvement in mood that patients
show when they first enter treatment because of expectancies and
the instillation of hope. Remediation is characterized by symptom
improvement, which depending on the treatment and the disorder
can occur relatively quickly, between 4 and 16 weeks of beginning
treatment. Rehabilitation is characterized by personality change.
We hypothesize that changes in RF and narrative coherence are
akin to rehabilitative changes in the internal structure of represen-
tations of self and other that will provide patients with buffers
against internal and external stressors. These patients may be less
likely to create stress (e.g., acting in ways that gets one fired) and
better able to withstand natural stressors (e.g., bad economy or
natural disaster). Thus, patients who have made rehabilitative
changes may be better suited to interact with the world at large. Of
course, confirmation of these speculations in follow-up studies is
Ablon, J. S., & Jones, E. E. (1998). How expert ’clinicians’ prototypes of
an ideal treatment correlate with outcome in psychodynamic and
cognitive-behavioral therapy. Psychotherapy Research, 8, 71–83.
Ablon, J. S., Levy, R. A., & Kazenstein, T. (2006). Beyond brand names
of psychotherapy: Identifying empirically supported change processes.
Psychotherapy: Theory, Research, Practice, Training, 42, 216–231.
Ainsworth, M. S., Blehar, M. C., Waters, E., & Wall, S. (1978). Patterns
of attachment: A psychological study of the Strange Situation. Hillsdale,
American Psychiatric Association. (2000). Diagnostic and statistical man-
ual of mental disorders (4th ed., text revision). Washington, DC: Author.
Appelbaum, A. H. (1994). Psychotherapeutic routes to structural change.
Bulletin of the Menninger Clinic, 58, 37–54.
Appelbaum, A. H. (1996). Why traumatized borderline patients relapse.
Bulletin of the Menninger Clinic, 60, 449–463.
Appelbaum, A. H. (2005). Supportive psychotherapy. In J. Oldham, A.
Skodol, & D. Bender (Eds.), The American Psychiatric Publishing
textbook of personality disorders (pp. 311–326). Washington, DC:
American Psychiatric Publishing.
Barone, L. (2003). Developmental protective and risk factors in borderline
personality disorder: A study using the adult attachment interview.
Attachment & Human Development, 5, 64–77.
Bateman, A., & Fonagy, P. (1999). Effectiveness of partial hospitalization
in the treatment of borderline personality disorder: A randomized con-
trolled trial. American Journal of Psychiatry, 156, 1563–1569.
Bateman, A., & Fonagy, P. (2003). The development of an attachment-
based treatment program for borderline personality disorder. Bulletin of
the Menninger Clinic, 67, 187–211.
Blatt, S. J., & Levy, K. N. (2003). Attachment theory, psychoanalysis,
personality development, and psychopathology. Psychoanalytic Inquiry,
Blum, N., Pfohl, B., & St. John, D. (2002). STEPPS: A cognitive-
behavioral systems-based group treatment for outpatients with border-
SPECIAL SECTION: ATTACHMENT, REFLECTIVENESS, AND BPD
line personality disorder—A preliminary report. Comprehensive Psychi-
atry, 43, 301–310.
Bongar, B., Peterson, L. G., Golann, S., & Hardiman, J. J. (1990). Self-
mutilation and the chronically “suicidal” patient: An examination of the
frequent visitor to the psychiatric emergency room. Annals of Clinical
Psychiatry, 2, 217–222.
Bowlby, J. (1977). The making and breaking of affectional bonds: I.
Aetiology and psychopathology in the light of attachment theory. British
Journal of Psychiatry, 130, 201–210.
Brown, G. K., Newman, C. F., & Charlesworth, S. E. (2004). An open
clinical trial of cognitive therapy for borderline personality disorder.
Journal of Personality Disorders, 18, 257–271.
Castonguay, L. G., Goldfried, M. R., Wiser, S., & Raue, P. J. (1996).
Predicting the effect of cognitive therapy for depression: A study of
unique and common factors. Journal of Consulting & Clinical Psychol-
ogy, 64, 497–504.
Clarkin, J. F., Foelsch, P. A., Levy, K. N., Hull, J. W., Delaney, J. C., &
Kernberg, O. F. (2001). The development of a psychodynamic treatment
for patients with borderline personality disorders: A preliminary study of
behavioral change. Journal of Personality Disorders, 15, 487–495.
Clarkin, J. F., & Levy, K. N. (2003). Influence of client variables on
psychotherapy. In M. Lambert (Ed.), Bergin and Garfield’s handbook of
psychotherapy and behavior change (5th ed.). New York: Wiley.
Clarkin, J. F., Levy, K. N., Lenzenweger, M. F., & Kernberg, O. F. (2004).
The Personality Disorders Institute/Borderline Personality Disorder Re-
search Foundation randomized control trial for borderline personality
disorder: Rationale, methods, and patient characteristics. Journal of
Personality Disorders, 18, 52–72.
Clarkin, J. F., Levy, K. N., Lenzenweger, M. F., & Kernberg, O. F. (2006).
The Personality Disorders Institute/Borderline Personality Disorder
Research Foundation randomized control trial for borderline personal-
ity disorder: Treatment outcome. Manuscript submitted for publication.
Clarkin, J. F., Yeomans, F., & Kernberg, O. F. (1999). Psychotherapy of
borderline personality: Focusing on object relations. New York: Wiley.
Clarkin, J. F., Yeomans, F., & Kernberg, O. F. (2006). Psychotherapy of
borderline personality: Focusing on object relations. Arlington, VA:
American Psychiatric Publishing.
Critchfield, K. L., Levy, K. N., & Clarkin, J. F. (2005). The relationship
between impulsivity, aggression, and impulsive-aggression in borderline
personality disorder: An empirical analysis of self-report measures.
Journal of Personality Disorders, 18, 555–570.
Critchfield, K. L., Levy, K. N., & Clarkin, J. F. (in press). The Personality
Disorders Institute/Borderline Disorder Research Foundation random-
ized control trial for borderline personality disorder: Axis I and II
diagnoses. Psychiatric Quarterly.
Crowell, J. A., Treboux, D., & Waters, E. (2002). Stability of attachment
representations: The transition to marriage. Developmental Psychology,
Cyranowski, J. M., Bookwala, J., Feske, U., Houck, P., Pilkonis, P.,
Kostelnik, B., & Frank, E. (2002). Adult attachment profiles, interper-
sonal difficulties, and response to interpersonal psychotherapy in women
with recurrent major depression. Journal of Social and Clinical Psychol-
ogy, 21, 191–217.
DeRubeis, R. J., Evans, M. D., Hollon, S. D., Garvey, M. J., Grove, W. M.,
& Tuason, V. B. (1990). How does cognitive therapy work? Cognitive
change and symptom change in cognitive therapy and pharmacotherapy
for depression. Journal of Consulting and Clinical Psychology, 58,
DeRubeis, R. J., & Feeley, M. (1990). Determinants of change in cognitive
therapy for depression. Cognitive Therapy and Research, 14, 469–482.
Diamond, D., Clarkin, J. F., Levine, H., Levy, K., Foelsch, P., & Yeomans,
F. (1999). Borderline conditions and attachment: A preliminary report.
Psychoanalytic Inquiry, 19, 831–884.
Dozier, M. (1990). Attachment organization and treatment use for adults
with serious psychopathological disorders. Development and Psychopa-
thology, 2, 47–60.
Dozier, M., Cue, K. L., & Barnett, L. (1994). Clinicians as caregivers: The
role of attachment organization in treatment. Journal of Consulting and
Clinical Psychology, 62, 793–800.
Eagle, M. (2003). Clinical implications of attachment theory. Psychoana-
lytic Inquiry, 23, 12–27.
Eagle, M., & Wolitzky, D. L. (in press). The perspectives of attachment
theory & psychoanalysis: Adult psychotherapy. In E. Berant & J. Obegi
(Eds.), Clinical applications of adult attachment. New York: Guilford
Farber, B. A., Lippert, R. A., & Nevas, D. B. (1995). The therapist as
attachment figure. Psychotherapy, 32, 204–212.
First, M. B., Gibbon, M., Spitzer, R. L., & Williams, J. B. W. (1997).
Structured clinical interview for Axis I DSM–IV disorders (SCID-I),
clinical version. New York: Biometrics Research Department, New
York State Psychiatric Institute.
Fleiss, J. L. (1971). Measuring nominal scale agreement among many
raters. Psychological Bulletin, 76, 378–381.
Fonagy, P. (1991). Thinking about thinking: Some clinical and theoretical
considerations in the treatment of a borderline patient. International
Journal of Psycho-Analysis, 72, 639–656.
Fonagy, P., Gergely, G., Jurist, E., & Target, M. (2002). Affect regulation,
mentalization, and the development of the self. New York: Other Press.
Fonagy, P., Leigh, T., Steele, M., Steele, H., Kennedy, R., Mattoon, G., et
al. (1996). The relation of attachment status, psychiatric classification,
and response to psychotherapy. Journal of Consulting & Clinical Psy-
chology, 64, 22–31.
Fonagy, P., Steele, H., Moran, G., Steele, M., & Higgitt, A. (1991). The
capacity for understanding mental states: The reflective self in parent
and child and its significance for security of attachment. Infant Mental
Health Journal, 13, 200–217.
Fonagy, P., Steele, M., Steele, H., Leigh, T., Kennedy, R., Mattoon, G., &
Target, M. (1995). Attachment, the reflective self, and borderline states.
In S. Goldberg & J. Kerr (Eds.), Attachment research: The state of the
art (pp. 233–278). New York: Analytic Press.
Fonagy, P., Steele, M., Steele, H., & Target, M. (1998). Reflexive-function
manual: Version 5.0 for application to the adult attachment interview.
Unpublished manual, University College, London.
Fonagy, P., & Target, M. (1996). Playing with reality: I. Theory of mind
and the normal development of psychic reality. International Journal of
Psychoanalysis, 77, 217–234.
George, C., Kaplan, N., & Main, M. (1985). The Berkeley Adult Attach-
ment Interview. Unpublished manuscript, University of California,
Giesen-Bloo, J., van Dyck, R., Spinhoven, P., van Tilburg, W., Dirksen, C.,
van Asselt, T., et al. (2006). Outpatient psychotherapy for borderline
personality disorder: Randomized trial of schema-focused therapy vs.
transference-focused psychotherapy. Archives of General Psychiatry,
Gratz, K. L., & Gunderson, J. G. (2006). Preliminary data on an
acceptance-based emotion regulation group intervention for deliberate
self-harm among women with borderline personality disorder. Behavior
Therapy, 37, 25–35.
Grienenberger, J. F., Kelly, K., & Slade, A. (2005). Maternal reflective
functioning, mother-infant affective communication, and infant attach-
ment: Exploring the link between mental states and observed caregiving
behavior in the intergenerational transmission of attachment. Attachment
and Human Development, 7, 299–311.
Gross, R., Olfson, M., Gameroff, M., Shea, S., Feder, A., Fuentes, M., et
al. (2002). Borderline personality disorder in primary care. Archives of
Internal Medicine, 162, 53–60.
Gunderson, J. G. (1996). The borderline patient’s intolerance of aloneness:
LEVY ET AL.
Insecure attachments and therapist availability. American Journal of
Psychiatry, 153, 752–758.
Gurvits, I. G., Koenigsberg, H. W., & Siever, L. J. (2000). Neurotransmit-
ter dysfunction in patients with borderline personality disorder. Psychi-
atric Clinics of North America, 23, 27–40.
Hamilton, C. E. (2000). Continuity and discontinuity of attachment from
infancy through adolescence. Child Development, 71, 690–694.
Hesse, E. (1996). Discourse, memory, and the Adult Attachment Interview:
A note with emphasis on the emerging cannot classify category. Infant
Mental Health Journal, 17, 4–11.
Holmes, J. (1995). Supportive psychotherapy. The search for positive
meanings. British Journal of Psychiatry, 167, 439–447.
Holmes, J. (1996). Attachment, intimacy, autonomy: Using attachment
theory in adult psychotherapy. Northvale, NJ: Aronson.
Holmes, J. (2003). Borderline personality disorder and the search for
meaning: An attachment perspective. Australian and New Zealand Jour-
nal of Psychiatry, 37, 524–531.
Holmes, J. (2004). Disorganized attachment and borderline personality
disorder: A clinical perspective. Attachment and Human Development,
Howard, K. I., Kopta, S. M., Krause, M. S., & Orlinsky, D. E. (1986). The
dose-response relationship in psychotherapy. American Psychologist,
Ilardi, S. S., & Craighead, W. E. (1994). The role of nonspecific factors in
cognitive behavior therapy for depression. Clinical Psychology: Science
and Practice, 1, 138–155.
Jones, E. E., & Pulos, S. M. (1993). Comparing the process in psychody-
namic and cognitive-behavioral therapies. Journal of Consulting and
Clinical Psychology, 61, 306–316.
Kernberg, O. F. (1984). Severe personality disorders: Psychotherapeutic
strategies. New Haven, CT: Yale University Press.
Kernberg, O. F., Selzer, M., Koenigsberg, H. W., Carr, A., & Appelbaum,
A. (1989). Psychodynamic psychotherapy of borderline patients. New
York: Basic Books.
Kjellander, C., Bongar, B., & King, A. (1998). Suicidality in borderline
personality disorder. Crisis, 19, 125–135.
Koons, C. R., Robins, C. J., Tweed, J. L., Lynch, T. R., Gonzalez, A. M.,
Morse, J. Q., et al. (2001). Efficacy of dialectical behavior therapy in
women veterans with borderline personality disorder. Behavior Therapy,
Leichsenring, F., & Leibing, E. (2003). The effectiveness of psychody-
namic therapy and cognitive behavior therapy in the treatment of per-
sonality disorders: A meta-analysis. American Journal of Psychiatry,
Lenzenweger, M. F., Loranger, A. W., Korfine, L., & Neff, C. (1997).
Detecting personality disorders in a nonclinical population: Application
of a 2-stage procedure for case identification. Archives of General
Psychiatry, 54, 345–351.
Levy, K. N. (2005). The implications of attachment theory and research for
understanding borderline personality disorder. Development and Psy-
chopathology, 17, 959–986.
Levy, K. N. (2006). Letter to the editor. Unpublished manuscript.
Levy, K. N., & Blatt, S. J. (1999). Attachment theory and psychoanalysis:
Further differentiation within insecure attachment patterns. Psychoana-
lytic Inquiry, 19, 541–575.
Levy, K. N., Clarkin, J. F., Foelsch, P. A., & Kernberg, O. F. (2006).
Transference focused psychotherapy for patients diagnosed with bor-
derline personality disorder: A comparison with a treatment-as-usual
cohort. Manuscript submitted for publication.
Levy, K. N., Clarkin, J. F., Yeomans, F. E., Scott, L. N., Wasserman, R. H.,
& Kernberg, O. F. (2006). The mechanisms of change in the treatment
of transference focused psychotherapy. Journal of Clinical Psychology,
Levy, K. N., Critchfield, K. L., & Clarkin, J. F. (2006). Generalizability of
the Personality Disorder Institute/Borderline Personality Disorder Re-
search Foundation randomized clinical trial: A comparison of random-
ized participants to those not randomized. Unpublished manuscript.
Lewis, M., Feiring, C., & Rosenthal, S. (2000). Attachment over time.
Child Development, 71, 707–720.
Linehan, M. M. (1993). Cognitive-behavioral treatment of borderline
personality disorder. New York: Guilford Press.
Linehan, M. M., Armstrong, H. E., Suarez, A., Allmon, D., & Heard, H. L.
(1991). Cognitive-behavioral treatment of chronically parasuicidal bor-
derline patients. Archives of General Psychiatry, 48, 1060–1064.
Linehan, M. M., Comtois, K. A., Murray, A. M., Brown, M. Z., Gallup,
R. J., Heard, H. L., et al. (2006). Two-year randomized controlled trial
and follow-up of dialectical behavior therapy vs. therapy by experts for
suicidal behaviors and borderline personality disorder. Archives of Gen-
eral Psychiatry, 63, 757–766.
Linehan, M. M., Dimeff, L. A., Reynolds, S. K., Comtois, K. A., Welch,
S. S., Heagerty, P., & Kivlahan, D. R. (2002). Dialetical behavior
therapy versus comprehensive validation therapy plus 12-step for the
treatment of opioid dependent women meeting criteria for borderline
personality disorder. Drug and Alcohol Dependence, 67, 13–26.
Linehan, M. M., Kanter, J. W., & Comtois, K. A. (1999). Dialectical
behavior therapy for borderline personality disorder: Efficacy, specific-
ity, and cost effectiveness. In D. S. Janowsky (Ed.), Psychotherapy
indications and outcomes (pp. 93–118). Washington, DC: American
Linehan, M. M., Schmidt, H., Dimeff, L. A., Craft, J. C., Kanter, J., &
Comtois, K. A. (1999). Dialectical behavior therapy for patients with
borderline personality disorder and drug-dependence. The American
Journal on Addictions, 8, 279–292.
Liotti, G., & Pasquini, P. (2000). Predictive factors for borderline person-
ality disorder: Patients’ early traumatic experiences and losses suffered
by the attachment figure: The Italian Group for the Study of Dissocia-
tion. Acta Psychiatrica Scandinavica, 102, 282–289.
Loranger, A. W., Sartorius, N., Andreoli, A., & Berger, P. (1994). The
international personality disorder examination: The World Health Orga-
nization/Alcohol, Drug Abuse, and Mental Health Administration inter-
national pilot study of personality disorders. Archives of General Psy-
chiatry, 51, 215.
Main, M., & Goldwyn, R. (in press). Adult attachment rating and classi-
fication systems. In M. Main (Ed.), [A typology of human attachment
organization assessed in discourse, drawings, and interviews]. New
York: Cambridge University Press.
Main, M., Kaplan, N., & Cassidy, J. (1985). Security in infancy, childhood,
and adulthood: A move to the level of representation. Monographs of the
Society for Research in Child Development, 50, 66–104.
Main, M., & Weston, D. R. (1981). The quality of the toddler’s relationship
to mother and to father: Related to conflict behavior and the readiness to
establish new relationships. Child Development, 52, 932–940.
Mallinckrodt, B., Gantt, D. L., & Coble, H. M. (1995). Attachment patterns
in the psychotherapy relationship: Development of the client attachment
to therapist scale. Journal of Counseling Psychology, 42, 307–317.
Markowitz, J., Skodol, A. E., Bleiberg, K., & Strasser-Vorus, T. (2004,
July). IPT for borderline personality disorder. Paper presented to the
National Institute of Mental Health International Think Tank for the
More Effective Treatment of Borderline Personality Disorder, Lincthi-
McGlashan, T. H. (1986). The Chestnut Lodge follow-up study: III. Long-
term outcome of borderline personalities. Archives of General Psychi-
atry, 43, 20–30.
Meyer, B., Pilkonis, P. A., Proietti, J. M., Heape, C. L., & Egan, M. (2001).
Attachment styles and personality disorders as predictors of symptom
course. Journal of Personality Disorders, 15, 371–389.
Oldham, J. M., Gabbard, G. O., Goin, M. K., Gunderson, J., Soloff, P.,
Spiegel, D., et al. (2001). Practice guideline for the treatment of patients
SPECIAL SECTION: ATTACHMENT, REFLECTIVENESS, AND BPD
with borderline personality disorder. American Journal of Psychiatry, Download full-text
Paris, J. (1999). Borderline personality disorder. In T. Millon, P. H. Blaney,
& R. D. Davis (Eds.), Oxford textbook of psychopathology (pp. 625–
652). New York: Oxford University Press.
Patrick, M., Hobson, P., Castle, D., Howard, R., & Maughan, B. (1994).
Personality disorder and the mental representation of early social expe-
rience. Development and Psychopathology, 6, 375–388.
Perry, J. C., Banon, E., & Ianni, F. (1999). Effectiveness of psychotherapy
for personality disorders. American Journal of Psychiatry, 156, 1312–
Rockland, L. H. (1992). Supportive therapy for borderline patients: A
psychodynamic approach. New York: Guilford Press.
Ryle, A., & Golynkina, K. (2000). Effectiveness of time-limited cognitive
analytic therapy of borderline personality disorder: Factors associated
with outcome. British Journal of Medical Psychology, 73, 197–210.
Shaw, B. F., Elkin, I., Yamaguchi, J., Olmsted, M., Vallis, M., Dobson,
K. S., et al. (1999). Therapist competence ratings in relation to clinical
outcome in cognitive therapy of depression. Journal of Consulting and
Clinical Psychology, 67, 837–846.
Skodol, A. E., Gunderson, J. G., Pfohl, B., Widiger, T. A., Livesley, W. J.,
& Siever, L. J. (2002). The borderline diagnosis: I. Psychopathology,
comorbidity, and personality structure. Biological Psychiatry, 51, 936–
Slade, A. (1999). Representation, symbolization, and affect regulation in
the concomitant treatment of a mother and child: Attachment theory and
child psychotherapy. Psychoanalytic Inquiry, 19, 797–830.
Stevenson, J., & Meares, R. (1992). An outcome study of psychotherapy
for patients with borderline personality disorder. American Journal of
Psychiatry, 149, 358–362.
Stone, M. H. (1983). Long-term outcome in personality disorders. British
Journal of Psychiatry, 162, 299–313.
Stovall-McClough, K. C., & Cloitre, M. (2003). Reorganization of unre-
solved childhood traumatic memories following exposure therapy. An-
nuals of the New York Academy of Sciences, 1008, 297–299.
Telch, C. F., Agras, W. S., & Linehan, M. M. (2001). Dialectical behavior
therapy for binge eating disorder. Journal of Consulting & Clinical
Psychology, 69, 1061–1065.
Torgersen, S., Kringlen, E., & Cramer, V. (2001). The prevalence of
personality disorders in a community sample. Archives of General
Psychiatry, 58, 590–596.
Trepka, C., Rees, A., Shapiro, D. A., Hardy, G. E., & Barkham, M. (2004).
Therapist competence and outcome of cognitive therapy for depression.
Cognitive Therapy and Research, 25, 143–157.
Turner, R. M. (2000). Naturalistic evaluation of dialectical behavior
therapy-oriented treatment for borderline personality disorder. Cognitive
and Behavioral Practice, 7, 413–419.
Tyrrell, C. L., Dozier, M., Teague, G. B., & Fallot, R. D. (1999). Effective
treatment relationships for persons with serious psychiatric disorders:
The importance of attachment states of mind. Journal of Consulting &
Clinical Psychology, 67, 725–733.
Verheul, R., van den Bosch, L. M. C., Koeter, M. W. J., de Ridder,
M. A. J., Stijnen, T., & van den Brink, W. V. (2003). Dialectical
behavior therapy for women with borderline personality disorder. British
Journal of Psychiatry, 182, 135–140.
Waters, E., Merrick, S., Treboux, D., Crowell, J., & Albersheim, L. (2000).
Attachment security in infancy and early adulthood: A 20-year longitu-
dinal study. Child Development, 71, 684–689.
Waters, E., Treboux, D., Fyffe, C., & Crowell, J. (2001). Secure versus
insecure and dismissing versus preoccupied attachment representation
scored as continuous variables from AAI state of mind scales. New
York: Stony Brook University, State University of New York.
Weinfield, N. S., Sroufe, L. A., & Egeland, B. (2000). Attachment from
infancy to early adulthood in a high-risk sample: Continuity, disconti-
nuity, and their correlates. Child Development, 71, 695–702.
Westen, D., & Morrison, K. (2001). A multidimensional meta-analysis of
treatment for depression, panic, and generalized anxiety disorder: An
empirical examination of the status of empirically supported therapies.
Journal of Consulting & Clinical Psychology, 69, 875–899.
Westen, D., Novotny, C. M., & Thompson-Brenner, H. (2004). The em-
pirical status of empirically supported psychotherapies: Assumptions,
findings, and reporting in controlled clinical trials. Psychological Bul-
letin, 130, 631–663.
Zanarini, M. C., & Frankenburg, F. R. (2001). Olanzapine treatment of
female borderline personality disorder patients: A double-blind, placebo-
controlled pilot study. Journal of Clinical Psychiatry, 62, 849–854.
Zanarini, M. C., Frankenburg, F. R., Reich, D., Marino, M. F., Haynes,
M. C., & Gunderson, J. G. (1999). Violence in the lives of adult
borderline patients. Journal of Nervous and Mental Disease, 187, 65–71.
Received May 2, 2006
Revision received September 18, 2006
Accepted September 18, 2006 ?
LEVY ET AL.