[acb-hsp] Effects of transference work ...

[J.Rayl]

Effects of transference work in the context of therapeutic alliance and quality

of object relations.

____________________________________________________________

Document 1 of 1

Effects of transference work in the context of therapeutic alliance and quality of

object relations.

Author: Høglend, Per 1 ; Hersoug, Anne Grete 1 ; Bøgwald, Kjell-Petter 2 ; Amlo,

Svein 3 ; Marble, Alice 1 ; Sørbye,Øystein 4 ; Røssberg, Jan Ivar 5 ; Ulberg, Randi

6 ; Gabbard, Glen O. 7 ; Crits-Christoph, Paul 8 1 Department of Psychiatry, University

of Oslo, Olso, Norway

p.a.hoglend at medisin.uio.no

2 Department of Research and Education, Diakonhjemmet Hospital, Olso, Norway 3 Vestre

Viken Hospital, Akershus, Norway 4 Child and Adolescent Mental Health Clinic, Ulleval

University Hospital, University of Oslo, Norway 5 Department of Psychiatry, Ulleval

University Hospital, University of Oslo, Norway 6 Vestfold Mental Health Care Trust,

Tonsberg, Norway 7 Baylor Psychiatric Clinic, TX, US 8 Department of Psychiatry,

University of Pennsylvania, PA, US

Publication info: Journal of Consulting and Clinical Psychology 79. 5 (Oct 2011):

697-706.

https://login.libproxy.edmc.edu/login?url=http://search.proquest.com/docview/885055405?accountid=34899

Abstract: Objective: Transference interpretation is considered as a core active ingredient

in dynamic psychotherapy. In common clinical theory, it is maintained that more mature

relationships, as well as a strong therapeutic alliance, may be prerequisites for

successful transference work. In this study, the interaction between quality of object

relations, transference interpretation, and alliance is estimated. Method: One hundred

outpatients seeking psychotherapy for depression, anxiety, and personality disorders

were randomly assigned to 1 year of weekly sessions of dynamic psychotherapy with

transference interpretation or to the same type and duration of treatment, but without

the use of transference interpretation. Quality of Object Relations (QOR)-lifelong

pattern was evaluated before treatment (P. Høglend, 1994). The Working Alliance Inventory

(A. O. Horvath & L. S. Greenberg, 1989; T. J. Tracey & A. M. Kokotovic, 1989) was

rated in Session 7. The primary outcome variable was the Psychodynamic Functioning

Scales (P. Høglend et al., 2000), measured at pretreatment, posttreatment, and 1

year after treatment termination. Results: A significant Treatment Group × Quality

of Object Relations × Alliance interaction was present, indicating that alliance

had a significantly different impact on effects of transference interpretation, depending

on the level of QOR. The impact of transference interpretation on psychodynamic functioning

was more positive within the context of a weak therapeutic alliance for patients

with low quality of object relations. For patients with more mature object relations

and high alliance, the authors observed a negative effect of transference work. Conclusion:

The specific effects of transference work was influenced by the interaction of object

relations and alliance, but in the direct opposite direction of what is generally

maintained in mainstream clinical theory. (PsycINFO Database Record (c) 2012 APA,

all rights reserved)(journal abstract)

Links: null

Full Text: Contents - Abstract

Method Patients

Treatment Conditions and Therapists

Treatment Fidelity

Assessments

Outcome Measure

Moderator

Statistical Analysis

Results

Discussion

Show less

Figures and Tables - Figure 1

- Figure 2

- Table 1

Show less Abstract Objective: Transference interpretation is considered as a core

active ingredient in dynamic psychotherapy. In common clinical theory, it is maintained

that more mature relationships, as well as a strong therapeutic alliance, may be

prerequisites for successful transference work. In this study, the interaction between

quality of object relations, transference interpretation, and alliance is estimated.

Method: One hundred outpatients seeking psychotherapy for depression, anxiety, and

personality disorders were randomly assigned to 1 year of weekly sessions of dynamic

psychotherapy with transference interpretation or to the same type and duration of

treatment, but without the use of transference interpretation. Quality of Object

Relations (QOR)-lifelong pattern was evaluated before treatment ( P. Høglend, 1994

). The Working Alliance Inventory ( A. O. Horvath & L. S. Greenberg, 1989 ; T. J.

Tracey & A. M. Kokotovic, 1989 ) was rated in Session 7. The primary outcome variable

was the Psychodynamic Functioning Scales ( P. Høglend et al., 2000 ), measured at

pretreatment, posttreatment, and 1 year after treatment termination. Results: A significant

Treatment Group × Quality of Object Relations × Alliance interaction was present,

indicating that alliance had a significantly different impact on effects of transference

interpretation, depending on the level of QOR. The impact of transference interpretation

on psychodynamic functioning was more positive within the context of a weak therapeutic

alliance for patients with low quality of object relations. For patients with more

mature object relations and high alliance, the authors observed a negative effect

of transference work. Conclusion: The specific effects of transference work was influenced

by the interaction of object relations and alliance, but in the direct opposite direction

of what is generally maintained in mainstream clinical theory.

In the psychotherapy clinical and research literature in general, the role of unspecific

factors (e.g., the alliance) versus technique factors (e.g., transference interpretation)

in producing change continues to be a source of controversy ( Barber, 2009 ; Beutler

& Kendall, 1995 ; Crits-Christoph & Connolly, 1999 ; Weinberger, 1995 ). In the research

literature, all too often these components have been pitted against one another,

fostering the idea that it is either technique or alliance that is most responsible

for change ( Barber, 2009 ; Goldfried & Davila, 2005 ). There are only a few studies

exploring alliance and technique within the same study, and even fewer actually studying

the interaction of technique and alliance.

Transference has been a core concept in dynamic psychotherapy for over a century

( Freud, 1905/1953 ). Freud originally regarded transference as a living reconstruction

of the patient's repressed historical past "transferred" onto the relationship with

the therapist. Later theorists have questioned the notion of transference as a pure

enactment of early relationships and emphasized how transference is partly a new

experience ( Cooper, 1987 ). Dynamic psychotherapy is interpersonal in nature, and

the transference is also influenced by the therapist. Additional concepts such as

the therapeutic alliance and the real relationship with the therapist may be needed

to account for the patient's reactions to the therapist ( Ehrenreich, 1989 ; Gabbard

& Westen, 2003 ). More recently, clinical theorists and researchers have relied on

broader definitions of transference and transference interpretation that are more

experience near. In this study, we define transference work as all therapist interventions

with explicit linking to the patient-therapist interaction ( Høglend, 1993b ; Piper,

Azim, Joyce, & McCallum, 1991 ). Influential theorists maintain that the ongoing

interaction between patient and psychotherapist is influenced by the patient's past

or current relationships and affective experiences. Therefore, focusing on the themes

and conflicts that arise in the therapeutic relationship will have immediate affective

resonance and illuminate the true nature of problems in the patient's relationships

outside of therapy ( Kernberg, Diamond, Yeomans, Clarkin, & Levy, 2008 ; Strachey,

1934 ). Focus on transference can enable the patient (and therapist) to distinguish

what is real in the therapeutic relationship from what are enactments influenced

by earlier experiences. Analysis of transference may increase insight regarding intrapsychic

conflicts and problematic relations, which may in turn lead to better adaptive and

interpersonal functioning ( Gabbard & Westen, 2003 ; McGlashan & Miller, 1982 ; Messer

& McWilliams, 2007 ; Strachey, 1934 ).

As psychoanalytic thinking has broadened to include more relational perspectives,

the importance of the alliance has increased substantially ( Messer & Warren, 1995

). Most contemporary models of exploratory dynamic psychotherapy suggest that the

effects of specific techniques are dependent on a positive therapeutic alliance (

Crits-Christoph & Barber, 1991 ; Luborsky, 1984 ; Malan, 1976 ; Strupp & Binder,

1984 ). The term working alliance was introduced by Greenson (1965) , who emphasized

the patient's capacity to work purposefully in the treatment situation. Bordin (1976)

proposed a pantheoretical conceptualization of working alliance that incorporated

a mutual understanding of the purpose of therapy (goal); agreement on how to work

together toward the goal (task); and the patient's personal liking, trusting, and

valuing of the therapist (bond) ( Bordin, 1976 ; Gabbard et al., 1994 ; Horvath &

Greenberg, 1986 , 1989 ).

Several studies have included both technique and alliance variables, but without

actually studying the interaction of alliance and technique ( Barber, Crits-Christoph,

& Luborsky, 1996 ; Marmar, Gaston, Gallagher, & Thompson, 1989 ; Ogrodniczuk, Piper,

Joyce, & McCallum, 2000 ; Piper, McCallum, Azim, & Joyce, 1993 ; Spinhoven, Giesen-Bloo,

van Dyck, Kooiman, & Arntz, 2007 ).

Other studies that did examine interactions between the alliance and techniques have

yielded mixed results. Crits-Christoph, Cooper, and Luborsky (1988) failed to find

a significant interaction between accuracy of interpretation and the alliance in

the prediction of outcome of moderate length dynamic therapy. Svartberg and Stiles

(1994) reported no interaction of alliance and competently delivered interventions

in short-term dynamic psychotherapy. Similarly, Gaston, Piper, Debbane, Bienvenu,

and Garant (1994) reported no significant interaction between working and therapeutic

alliance and exploratory technique in short-term therapy. In long-term therapy, however,

the interaction of working alliance and exploratory interventions significantly predicted

both symptoms and interpersonal problems at termination. These interactions were

in the direction of better outcome associated with the use of exploratory interventions

in the context of a good alliance. However, the sample size in the study was small

( N = 15 in long-term therapy). In another study, Gaston, Thompson, Gallahager, Cournoyer,

and Gagnon (1998) found that in Session 10, exploratory interventions provided in

the context of better alliances tended to be associated with lower depression at

termination. However, Gaston et al. (1998) also reported that in behavior therapy

and cognitive therapy, more exploratory interventions worked best in the context

of poorer alliance. Barber et al. (2006) reported that high adherence to the techniques

of drug counseling worked best when the alliance was weak.

Patient characteristics are important to consider as moderating factors in the relationship

between transference interpretations and outcome. Luborsky (1984) recommended that

exploratory-interpretive work in psychodynamic therapy (including transference interpretation)

should be emphasized with patients who have good ego strength, anxiety tolerance,

and a capacity for reflection about their interpersonal relationships, whereas supportive

(alliance-building) techniques should be emphasized with patients who lack these

capacities.

Earlier naturalistic studies have indicated that the frequency of transference interpretations

has a nonsignificant, or a negative correlation with treatment outcome ( Høglend,

2004 ). Several of these studies have examined when the impact of the use of transference

interpretations on outcome is moderated by patient "quality of object relations"

(QOR) or interpersonal functioning. In short-term interpretative psychotherapy, Piper

et al. (1991) found an inverse relationship between high levels of transference interpretations

and outcome for patients who were characterized by high QOR. Høglend (1993b) also

found that frequency of transference interpretations was inversely related to better

outcome for patients with high QOR. However, two other studies ( Connolly et al.,

1999 ; Ogrodniczuk, Piper, Joyce, & McCallum, 1999 ) reported that patients with

relatively poorer interpersonal functioning had a less favorable outcome when transference

interpretations were used more frequently.

Although some of the above studies are suggestive of an interaction between alliance

and techniques, and quality of object relations and techniques, none of the studies

explored the three-way interaction of transference work, alliance, and object relations.

Furthermore, none of the studies used an experimental dismantling design. As Stiles

and Shapiro (1994) have argued, correlational studies attempting to link levels of

process components to outcome can be problematic. To the extent that therapists actually

respond to patient needs and requirements, process-outcome correlations can be misleading.

The therapeutic alliance may change directly as a consequence of technique ( Foreman

& Marmar, 1985 ), and/or the use of interpretations may be influenced by the strengths

of the alliance ( Marmar, Weiss, & Gaston, 1989 ). Crits-Christoph and Connolly (1999)

and Webb, DeRubeis, and Barber (2010) suggest that more studies of the interaction

of alliance and techniques using experimental designs are needed. Ideally, one process

component should be experimentally manipulated (e.g., high rates of transference

interpretation vs. low rates), whereas all others are held constant, and patients

are randomly assigned. Such rigorously designed studies might help in establishing

causal connections. Crits-Christoph and Connolly (1999) add that several factors

such as initial health or quality of object relations, the alliance, techniques,

and improvement over time should be examined within the same study. Sample sizes

larger than those used to date, in this area, would be needed to provide strong tests

of hypotheses and to explore possible three-way interactions.

Høglend et al. (2006) reported the results of a study using a randomized experimental

design. This study was the first dismantling, randomized clinical trial to test the

long-term effects of transference interpretation. One hundred patients were randomized

to 1-year dynamic psychotherapy with a moderate level of transference work or to

the same type of therapy without use of transference work. There was no overall effect

of transference work. However, patients with low QOR benefited significantly more

from therapy with transference interpretation than without ( Høglend et al., 2006

). This effect was sustained during a follow-up period ( Høglend et al., 2008 ).

Patients with mature relationships and greater psychological resources benefited

equally well from both treatments.

The goal of the present article was to further clarify these findings by examining

the interaction between quality of object relations, transference work, and therapeutic

alliance in relation to treatment outcome over time. Specifically, our aim was to

examine whether or not the association between alliance and the specific effects

of transference work changed as a function of different levels of QOR. Our focus

in particular was how these interactions relate to change in psychodynamic functioning,

over the 2-year study period. On the basis of the clinical literature and the limited

research, our hypothesis was that the effects of transference interpretations on

outcome will increase with stronger alliance. Moreover, for the typical patient with

low QOR, we hypothesized that the effect of transference work would be stronger in

the context of a high alliance, but for patients with more mature relationships,

alliance might be less important.

Method

Patients From 1994 to 2001, 122 patients were referred to the study therapists by

primary care physicians, private specialist practitioners, and public outpatient

departments. These patients sought psychotherapy due to depressive disorders, anxiety

disorders, personality disorders, and interpersonal problems. The study therapists

assessed the patients for eligibility. Patients with psychosis, bipolar illness,

organic mental disorder, or substance abuse were excluded. Patients with mental health

problems that caused long-term inability to work (>2 years) were also excluded. Written

informed consent was obtained from each of the 100 participants included in the study.

Treatment Conditions and Therapists The Regional Ethics Committee, Health-region

1, Norway, approved the study protocol. Patients were allocated by simple randomization,

without stratification, to two treatment groups after completion of the pretreatment

ratings. Only the patients' therapist learned the result of the random assignment

procedure. The random assignment code was kept on a separate computer, which belonged

to our research assistant. The other clinicians remained blind to the patient's treatment

group. Fifty-two patients were assigned to dynamic psychotherapy with low to moderate

use of transference work (transference group). Forty-eight patients were assigned

to dynamic psychotherapy of the same kind but without transference work (comparison

group).

Patients were assigned to one of seven therapists based on availability. The clinical

research team consisted of six psychiatrists and one clinical psychologist, all of

whom had 10-25 years of experience in practicing psychodynamic psychotherapy. Four

were fully trained psychoanalysts. Each therapist treated from 10 to 17 patients

in the study. All therapists treated patients from both groups. The patients were

offered 45-min sessions weekly for up to 1 year. All sessions were audio recorded.

Treatment manuals were used for both treatment conditions ( Høglend, 1990 ). Manuals

in dynamic psychotherapy are manuals of principles, not step-by-step procedures.

Our treatment model was based on general psychodynamic treatment techniques, such

as focus on affects, exploration of warded off material, focus on current relationships,

past relationships and the therapeutic relationship, interpretations of wishes, needs

and motives, and the principles outlined by Malan and Ferruccio (1992) and Sifneos

(1992) . In the pilot phase of the study, the therapists were trained for up to 4

years to enable them to provide treatment with a low to moderate level of transference

work and treatment without such interventions with equal ease and mastery.

Transference work included the following specific techniques: (a) The therapist was

to address transactions in the patient-therapist relationship; (b) the therapist

was to encourage exploration of thoughts and feelings about the therapy and therapist

including repercussions on the transference by high therapist activity; (c) the therapist

was to encourage the patient to discuss how he or she believed the therapist might

feel or think about him or her; (d) the therapist was to include himself or herself

explicitly in interpretive linking of dynamic elements (conflicts), direct manifestations

of transference, and also allusions to the transference; and (e) the therapist was

to interpret repetitive interpersonal patterns (including genetic interpretations)

and link these patterns to transactions between the patient and the therapist. The

first three techniques are not interpretations per se, but preparatory interventions.

In contrast, in the comparison group, the therapist consistently focused on interpersonal

relationships outside of therapy as the basis for similar interventions (extratransference

work) and did not link these patterns to the interaction between the patient and

the therapist. For both treatment groups, psychotherapy was exploratory in nature:

Patients were encouraged to explore sensitive topics that often involved uncomfortable

emotions, and the therapist abstained from giving advice, praise, or reassurance.

The small to moderate level of transference interventions recommended in the treatment

manual was based on 10 previous studies. The level of transference interpretation

in those studies varied from 1 to 6, on average per session ( Høglend, 2004 ). The

patients were not informed about which technique was used or the study hypotheses.

They were told that the aim of this study was to explore the long-term efficacy of

psychodynamic psychotherapy. Treatment completers were patients who terminated treatment

in agreement with the therapist.

Treatment Fidelity Treatment fidelity was assessed by three blind, independent raters,

using a manual for process ratings ( Høglend, 1994 ). The raters, two psychiatrists

and one psychologist, had 15-30 years of clinical experience as dynamic psychotherapists.

Two of them were fully trained psychoanalysts. The training period for the raters

included 15 full sessions from each treatment group. A global rating method was used

rather than rating the exact frequency of different interventions. The frequency

of a certain intervention does not necessarily give a valid measure of how important

this type of intervention was in a given session. Both how clearly an intervention

is offered and how much it is emphasized should be given weight in the rating process.

All items in the manual therefore use a 5-point Likert scale ranging from 0 ( not

at all used ), 2 ( moderately used ), to 4 ( very much used ). Four or five full

sessions of each therapy (452 sessions total) were rated by two of the raters. Using

average scores of the two raters, the reliability estimates (intraclass correlation

[ICC]) was above .70 for all items. Treatment integrity was excellent ( Bøgwald,

Høglend, & Sørbye, 1999 ; Høglend et al., 2006 , 2008 ). The only difference between

the two treatments was use of the specific transference interventions. The average

score across the five specific interventions was 1.7 ( SD = 0.7) in the transference

group, indicating moderate use of transference work, and 0.1 ( SD = 0.2) in the comparison

group, indicating nearly no use at all, t (58.2) = 14.8, p < .005. The average use

of supportive interventions was low and equal in the two treatment groups. The therapists'

skill in delivering the interventions was high and equal in the two treatment groups.

Assessments Before randomization, each patient had a 2-hr semistructured psychodynamic

interview, modified from Sifneos (1992) and Malan and Ferruccio (1992) , with an

independent evaluator. The interview was audio recorded, and also two other clinicians

rated the interview using the Quality of Object Relations Scale (QOR; Høglend, 1993a

; Piper et al., 1993 ), motivation for active change and self-understanding ( Høglend,

1994 ), and the Psychodynamic Functioning Scales (PFS; Høglend et al., 2000 ). The

PFS was also used at treatment termination and at follow-up 1 year after treatment

termination. The raters were independent (i.e., not the patient's therapist) and

blind with regard to treatment group. No structured interview was used in this study

to determine Axis I diagnoses. These diagnoses were based on the clinical history

and assessment of background variables by the patient's therapist. Diagnoses according

to the Diagnostic and Statistical Manual of Mental Disorders , third edition revised

(DSM-III-R; American Psychiatric Association, 1987 ) criteria were discussed before

randomization until consensus was reached ( Spitzer, 1983 ). Axis II diagnoses were

determined before the start of therapy by the patient's therapist, using the Structured

Clinical interview for DSM-III-R (SCID-II; Spitzer, Williams, Gibbon, & First, 1990

). All therapists were trained to use SCID-II. The patient filled in an expectancy

measure ( Battle, Imber, Hoehn-Saric, Nash, & Frank, 1966 ).

Outcome Measure Change on the PFS from baseline to 1 year after treatment termination

was the primary outcome measure in this study, chosen a priori. Six scales are used

in the PFS, with the same format as the Global Assessment of Functioning ( American

Psychiatric Association, 1987 ), to measure psychological functioning over the 3

previous months. Three of the scales measure interpersonal aspects: Quality of Family

Relationships, Quality of Friendships, and Quality of Romantic/Sexual Relationships.

The other three measure intrapersonal functioning: Tolerance for Affects, Insight,

and Problem Solving Capacity. Interrater reliability (ICC) for the average scores

of three raters on PFS was 0.91. Aspects of content validity, internal domain construct

validity, discriminant validity from symptom measures, and sensitivity for change

in dynamic therapy have been established in different samples of patients and evaluators

( Bøgwald & Dahlbender, 2004 ; Hagtvet & Høglend, 2008 ; Hersoug, 2004 ; Høglend,

2004 ; Høglend et al., 2000 ).

Moderator The primary putative moderator, QOR, was chosen a priori. The most common

definition of object relations in the literature appears to be an individual's representations

of self and other, and affect associated with these representations. This leads to

relatively stable patterns of interpersonal functioning ( Huprich & Greenberg, 2003

). QOR is the best studied predictor or moderator of treatment response and seems

to offer better utility than conventional diagnostic categories in predicting outcome

( Connolly et al., 1999 ; Høglend, 1993b ; Høglend et al., 2006 , 2008 ; Piper et

al., 1991 ). QOR was measured on three 8-point scales: Evidence of at Least one Stable

and Mutual Interpersonal Relationship in the Patient's Life, History of Adult Sexual

Relationships, and History of Non-sexual Adult Relationships ( Høglend, 1994 ; Høglend,

1993a ; Piper et al., 1993 ). QOR measure the patient's lifelong tendency to establish

relationships with others, ranging from mature to primitive, using the average of

the three scales. Higher QOR Scale scores indicate evidence of at least one stable

and mutual interpersonal relationship in the patient's history. Lower scores indicate

a lifelong history of less gratifying relationships characterized by less stability,

less emotional investment, and need for dependency or overcontrol. Interrater reliability

(ICC) for the average QOR scores of three raters was 0.84. The mean score in the

sample was 5.1 ( SD = 0.8; range = 2.6-7.3). Sixty percent of the patients with QOR

Scale scores below mean had one or more personality disorders in this study.

The Working Alliance Inventory, short form, with 12 items (WAI-S; Horvath & Greenberg,

1989 ; Tracey & Kokotovic, 1989 ), is the most widely used measure of alliance in

psychotherapy research and captures Bordin's three aspects of alliance: task, goal,

and bond ( Bordin, 1976 ). The WAI was reported by patients in Session 7, which is

assumed to more clearly represent the actual collaborative process between the patient

and the therapist than alliance in the very first sessions, which may reflect mainly

the bond aspect of alliance. We assume that alliance is a relatively stable aspect

of the therapy process, as is usually reported in the research literature ( Horvath

& Symonds, 1991 ). Total scores of the WAI were used. Cronbach's alpha was 0.76.

The mean score in the study sample was high: 5.2 at Session 7 ( SD = 0.7; range =

3.4-6.6).

Motivation for active change and self-understanding ( Høglend, 1994 ) was rated by

at least three clinicians, using four 8-point scales. Interrater reliability (ICC)

of average scores was 0.77. The patients' own target objectives for therapy were

formulated and scored using the method developed by Battle et al. (1966) . Prior

to treatment, the patients formulated their main target objectives for therapy and

the expected improvement for this. Target Expectancy was rated on a 12- point Likert

scale ranging from 1 ( deteriorate ) to 12 ( disappear totally ).

Statistical Analysis One outlier in the transference group was deleted from analyses

of longitudinal data as it became clear during treatment that this patient had been

abusing sedatives and painkillers over many years. Including this case also significantly

worsened goodness-of-fit measures (change in -2 log likelihood).

Longitudinal analyses were performed on 99 patients. Linear mixed models were used

to analyze longitudinal data (SPSS version 16.0). "Subject" was treated as a random

effect. That is, randomly distributed intercepts and slopes were fitted for each

patient. The highest rate of improvement was during therapy, with diminishing returns

over time following the end of therapy. Time was coded 1, 3, and 5, with one step

for each (1/2) year, and transformed to a natural logarithm. Time at baseline thereby

became 0. The log transformation of time fit the data discernibly better than a linear

time slope (change in -2 log likelihood). Intercept and time were treated as both

random and fixed effects, whereas treatment group (coded 1, 0) was treated as a fixed

effect. A variance component covariance matrix yielded the best goodness-of-fit measures.

In order to test the possible combined, interactive influence of QOR and WAI on the

effect of transference work, the following composite model equation was used:

Y ij is change of PFS over the 2-year study period. B 0 - B 9 are the fixed effects,

and ? oi + ? 1i TIME ij + e ij ] are random intercept, random time, and error term,

respectively. By design, treatment group means were equal at baseline. The statistical

model forces both treatments to have a common intercept. This model is more powerful

and is routinely recommended for analysis of randomized clinical trials ( Fitzmaurice,

Laird, & Ware, 2004 ; Kenny et al., 2004 ).

QOR and WAI were centered at their overall mean values. The relevant parameters are

B 2 , the treatment effect (difference in slopes between the two treatment groups)

for the typical patient in the sample. B 5 (QOR × Time × Treatment) indicates how

that treatment effect changes for the patients with typical (average) WAI as a function

of QOR. B 8 (WAI × Time × Treatment) indicates how the treatment effect changes for

the patients with typical QOR as a function of WAI. B 9 tests whether the interaction

of QOR and WAI influence the treatment effect-or put differently, whether the association

between WAI and the treatment effect changes as a function of different levels of

QOR. No statistical analyses were done on subgroups of patients. The full sample

of patients ( N = 99) was used in all analyses.

Effect sizes (converted to Cohen's d ), derived from the F test for mixed effects

model, were calculated as d = 2 , where F is the F test statistic for the effect

of interest in the repeated model as well as other multilevel designs ( Rosenthal

& Rosnow, 1991 ; Verbeke & Molenberg, 2000 ).

Results Figure 1 shows the flow of patients in the study. Five patients, all of them

in the comparison group, dropped out of therapy before Session 16. Only two of them

dropped out before Session 9. The mean (and SD ) number of sessions was 34 (6.1)

for the transference group, and 33 (6.6) for the comparison group. We could detect

no significant differences in patient characteristics between the two treatment groups

at baseline (see Table 1 ).

600?'600px':'auto');"> Enlarge this Image.

Patient flow in the randomized clinical trial.

600?'600px':'auto');"> Enlarge this Image.

Pretreatment Characteristics of 100 Patients Receiving 1 Year of Dynamic Psychotherapy

With or Without Transference Interpretation

In the whole study sample, there was no overall difference in long-term outcomes

between the two treatments ( Høglend et al., 2008 ). The mean WAI in the transference

group was 5.2 ( SD = 0.7) and 5.1 ( SD = 0.7) in the comparison group. QOR and treatment

group were uncorrelated ( r = -.03). WAI and treatment group were uncorrelated (

r = .05), and QOR and WAI were also uncorrelated ( r = -.12). Within the transference

group ( n = 51), we could not detect any associations between the amount of transference

work in early, middle, or late sessions, and WAI in Session 7. The correlations were

.02, .08, and .07, respectively. This indicates that the level of WAI was not influenced

by the amount of transference work, or vice versa, that WAI did not influence the

amount of transference work over treatment.

B 2 (Time × Treatment), the treatment effect for the typical patient in the sample,

was not significant ( B 5 = .47), F (1, 81) = 0.7, p = .40. The effect size was .20

(small). The moderator term B 5 (Time × Treatment × QOR) indicated that the differences

in slopes for treatment and control group (the treatment effect) for patients with

typical WAI increased with lower levels of the QOR Scale ( B 5 = -1.9), F (1, 81)

= 5.8, p = .02. The effect size of this term was .54 (moderate). B 8 (Time × Treatment

× WAI) indicated that the treatment effect for patients with typical QOR changed

in a negative direction as a function of higher WAI scores, but not to a significant

degree ( B 8 = -1.3), F (1, 81) = 2.7, p = .10. The effect size was, however, not

negligible: .37 (small to moderate). The interaction term B 9 (Time × Treatment ×

QOR × WAI) indicates how nonadditive the effects B 5 and B 8 are. The association

between WAI and the effects of transference work varied significantly, depending

on level of QOR ( B 9 = -1.7), F (1, 81) = 4.8, p = .03. The effect size for this

interaction term was .49 (moderate).

The term B 7 (Time × WAI) was significant ( B 7 = 1.7), F (1, 97) = 8.3, p = .005.

This reflects that WAI predicts overall outcome (across treatment groups), as generally

reported in the literature. Overall outcome is not the specific effect of transference

work.

To assess the potential effect of dropouts in the comparison group, we included treatment

completion status by time in the statistical model. This did not change the results.

In order to illustrate our findings, especially what the three-way interaction (B

9 ) means, we computed the specific treatment effect: B2 + B5QOR + B8WAI + B9QOR

× WAI, for patients with different pairs of values on QOR and WAI. The treatment

effects computed are estimated unstandardized mean differences in slopes between

the two treatment groups for hypothetical patients with the same pair of values on

QOR and WAI, for various values of QOR and WAI. Figure 2 shows that the association

between WAI and the specific treatment effect changes in a negative direction as

a function of higher QOR Scale scores. For patients with a typical (median) WAI score

(WAI = 5.1), the treatment effect is positive for patients with QOR Scale scores

below median (QOR = 5.2). For patients with typical QOR Scale scores, the association

between WAI and the treatment effect is negative.

600?'600px':'auto');"> Enlarge this Image.

The association between WAI (Working Alliance Inventory) and the specific effects

of transference work for patients with different QOR (Quality of Object Relations)

Scale scores. The treatment effects are estimated unstandardized mean differences

in slopes between the two treatment groups for hypothetical patients with the same

pair of values on QOR and WAI, for different values of QOR and WAI. WAI on the x-axis

between 10th and 90th percentiles. Selected values of QOR for the lines in the figure

are the 20th, 40th, 50th, 60th, and 80th percentiles.

Discussion Contrary to our hypothesis and common clinical wisdom, transference work

had the strongest specific effect for patients with low QOR Scale scores within the

context of weaker alliance (low WAI). For patients with a high alliance score and

high QOR Scale scores, the specific effect of transference work was negative. Unexpectedly,

patients with more mature relationships and the ability to form a favorable alliance

with the therapist did relatively better in the treatment that had no transference

interpretations compared with the treatment that had transference interpretations.

The conventional clinical wisdom in predicting psychotherapy outcome has been that

patients with greater psychological resources and more mature relationships will

benefit from transference interpretation ( Gabbard, 2006 ; Sifneos, 1992 ). However,

this study indicates that transference work is crucial when treating patients with

more severe and chronic difficulties in establishing stable and fulfilling relationships

outside therapy and also difficulties in establishing a favorable alliance within

therapy. Transference work may be helpful to the more disturbed patients' understanding

of the distortions they bring to the transference, such as fear of rejection, avoidance,

dependency, need for overcontrol, and devaluation/idealization ( Gabbard, 2006 ).

If the therapist does not address this, the patient may easily feel less understood

and less contained by the therapy, leading to deterioration in an already weak alliance

( Safran & Muran, 2000 ).

Although contrary to clinical wisdom, at least within the psychodynamic tradition,

our findings are consistent with experimental clinical research in medicine and psychiatry.

>From as early as 1950, researchers conducting randomized clinical trials for a wide

range of medical and psychiatric conditions have observed that patients with more

serious disturbances showed greater specific effects from active treatments ( Elkin

et al., 1989 ; Fisher, Lipman, Uhlenhuth, Rickels, & Park, 1965 ; Fournier et al.,

2010 ).

We used an experimental dismantling design in our study. The association between

transference work and outcome can therefore be interpreted as a causal relation.

However, the level of alliance cannot be experimentally manipulated. Alliance might

be dependent on technique, or techniques might be adjusted in response to the level

of alliance. However, we found that the average level of transference work in early,

middle, and late sessions was not associated with alliance. The correlations were

very small or zero. The alliance might also be a function of early improvement in

therapy. More frequently, measurement of psychodynamic outcomes would be needed to

unravel the causal connections between clinical improvement, alliance, and outcomes

at posttherapy and follow-up.

The relative advantage of dynamic psychotherapy without transference interpretations,

compared with that with transference interpretations, for patients with greater capacity

for mature relationships and a high alliance with the therapist is surprising. However,

a long-standing admonition that stems from Freud's work is that one should not interpret

transference until it becomes a resistance ( Gabbard, 2010 ). When a high alliance

exists, there may be minimal resistance. Hence, a therapist who is overly zealous

in interpreting transference may appear to the patient to be narcissistically and

needlessly focusing the patient's attention on the therapist. Resourceful patients

with a positive alliance may feel that they have only minimal problems with the therapist

so that the interpretation of transference is experienced as jarring or strange.

One may also speculate that patients who habitually establish mutual relationships

with others present more subtle transference cues, thus leading therapists to base

transference interpretation more on inference than on sufficient evidence. Such interventions

in the absence of concrete evidence may result in increased resistance. On the other

hand, the "spontaneous" transferences of patients with low QORs may take on a "dependent

" or "pathological" form that is more suitable for transference work. It should be

noted, however, that patients in the transference interpretation treatment with high

QOR and high alliance improved considerably on psychodynamic functioning from baseline

to posttherapy and follow-up. Thus, it is not the case that these patients who receive

transference interpretation do poorly in psychodynamic therapy. The data here, however,

suggest that such patients have particularly good outcomes in the context of a treatment

without transference work.

Regardless of the explanation for our results, the findings presented here indicate

that the relationships between therapist technique (transference work), patient characteristics

(QOR), therapy process (alliance), and outcome are complex. Examination of any one

of these variables in isolation from the others may provide a misleading understanding

of their role in relation to outcome. The presence of a three-way interaction between

technique, alliance, and a patient characteristic may explain why some previous studies

that have attempted to examine only two of these factors failed to find two-way interactions

(e.g., Crits-Christoph et al., 1988 ; Svartberg & Stiles, 1994 ). Lack of statistical

power to detect interactions may also have hampered previous studies. Findings from

the present study indicate that larger sample sizes than typically used in many psychotherapy

process-outcome studies may be needed to adequately test the complex interacting

relationships among key clinical concepts.

Several limitations of this research are important to note. Because our aim was not

to study treatment of specific disorders, the extent to which the findings of the

study are relatively specific to certain disorders and not others is not known. However,

the treatments studied do not target specific psychiatric disorders. The wide variety

of diagnoses in this sample may in fact increase generalizability to patients seeking

outpatient psychotherapy. A second limitation is that the alliance was only measured

at a single session.

For high-QOR patients, therapists were clearly in favor of using transference work

with 65% of the patients. For low-QOR patients, the proportion was 50%. The therapists

were specifically trained over a long period of time to be able to perform both treatments

equally well. Therapist allegiance effects can hardly explain our findings.

The alpha level was liberal in this study, which may have increased the risk of Type

I errors. Three-way interactions may be unreliable in moderately large patient samples.

This study was not large enough to provide accurate estimates of effect sizes. The

population estimates may range from small to large. Furthermore, our analyses were

exploratory in nature. Until the findings are replicated in future studies with larger

samples, they must be considered preliminary. Finally, there are likely other technique

variables (e.g., timing of interventions), patient variables (e.g., chronicity of

problems), and process variables (e.g., achievement of key insights) beyond the ones

measured here that are important to treatment outcome and may moderate the impact

of transference interpretations, QOR, and alliance. Despite these limitations, the

present study adds to an evolving literature that suggests that patient characteristics,

technique variables, and so-called common factors process variables are all important,

and interact in complex ways, to determine psychotherapy outcome (cf. Castonguay

& Beutler, 2005 ).

In this study, those patients who need to improve the most benefit the most from

explicit analysis of the patient-therapist interaction. One may speculate that such

work for the more disturbed patients may have the potential to improve patient-therapist

collaboration and outcome in other types of treatment as well.

References

1 . (1987). Diagnostic and statistical manual of mental disorders. Washington, DC:

Author.

2 . Barber, J. P. (2009). Psychotherapy Research.

3 . Barber, J. P., Crits-Christoph, P., & Luborsky, L. (1996). Journal of Consulting

and Clinical Psychology.

4 . Barber, J. P., Gallop, R., Crits-Christoph, P., Frank, A., Thase, M. E., Weiss,

R. D., & Connolly Gibbons, M. B. (2006). Psychotherapy Research.

5 . Battle, C. C., Imber, S. D., Hoehn-Saric, R., Nash, E. R., & Frank, J. D. (1966).

American Journal of Psychotherapy.

6 . Beutler, L. E., & Kendall, P. C. (1995). Journal of Consulting and Clinical Psychology.

7 . Bøgwald, K-P., & Dahlbender, R. W. (2004). Psychotherapy Research.

8 . Bøgwald, K-P., Høglend, P., & Sørbye, Ø. (1999). Journal of Psychotherapy Practice

and Research.

9 . Bordin, E. S. (1976). Psychotherapy: Theory, Research and Practice.

10 . Castonguay, L., & Beutler, L. (2005). Empirically supported principles of change.

New York, NY: Oxford University Press.

11 . Connolly, M. B., Crits-Christoph, P., Shappel, S., Barber, J. P., Luborsky,

L., & Shaffer, C. (1999). Psychotherapy Research.

12 . Cooper, A. M. (1987). Journal of the American Psychoanalytic Association.

13 . Crits-Christoph, P., & Barber, J. (1991). Handbook of short-term dynamic therapy.

New York, NY: Basic Books.

14 . Crits-Christoph, P., & Connolly, M. B. (1999). Clinical Psychology Review.

15 . Crits-Christoph, P., Cooper, A., & Luborsky, L. (1988). Journal of Consulting

and Clinical Psychology.

16 . Ehrenreich, J. H. (1989). Psychoanalytic Review.

17 . Elkin, I., Shea, M. T., Watkins, J. T., Imber, S. D., Sotsky, S. M., Collins,

J. F., & Parloff, M. B. (1989). Archives of General Psychiatry.

18 . Fisher, S., Lipman, R. S., Uhlenhuth, E. H., Rickels, K., & Park, L. C. (1965).

Psychopharmacologia.

19 . Fitzmaurice, G., Laird, N. M., & Ware, J. H. (2004). Applied longitudinal analysis.

New York, NY: John Wiley & Sons.

20 . Foreman, S. A., & Marmar, C. R. (1985). American Journal of Psychiatry.

21 . Fournier, J. A., DeRubeis, R. J., Hollon, S. D., Dimidjian, S., Amsterdam, J.

D., Shelton, R. C., & Fawcett, J. (2010). Journal of the American Medical Association.

22 . Freud, S., & Strachey, J. (1953). The standard edition of the complete psychological

works of Sigmund Freud. London, England: Hogarth Press.

23 . Gabbard, G. O. (2006). American Journal of Psychiatry.

24 . Gabbard, G. O. (2010). Long-term psychodynamic psychotherapy: A basic text.

Arlington, VA: American Psychiatric Publishing.

25 . Gabbard, G. O., Horwitz, L., Allen, J. G., Frieswyk, S., Newsom, G., Colson,

D. B., & Coyne, L. (1994). Harvard Review Psychiatry.

26 . Gabbard, G. O., & Westen, D. (2003). International Journal of Psychoanalysis.

27 . Gaston, L., Piper, W. E., Debbane, E. G., Bienvenu, J. P., & Garant, J. (1994).

Psychotherapy Research.

28 . Gaston, L., Thompson, L., Gallagher, D., Cournoyer, L. G., & Gagnon, R. (1998).

Psychotherapy Research.

29 . Goldfried, M. R., & Davila, J. (2005). Psychotherapy: Theory, Research, Practice,

Training.

30 . Greenson, R. R. (1965). Psychoanalytic Quarterly.

31 . Hagtvet, K. A., & Høglend, P. (2008). Measurement and Evaluation in Counseling

and Development.

32 . Hersoug, A. G. (2004). Journal of Personality Assessment.

33 . Høglend, P., Alnes, R., Ekern, P., & Jarval, P. (1990). Poliklinikken Psykiatrisk

Klinikk 25 år. Oslo, Norway: Universitetet i Oslo.

34 . Høglend, P. (1993a). Acta Psychiatrica Scandinavica.

35 . Høglend, P. (1993b). American Journal of Psychotherapy.

36 . Høglend, P. (1994). Manual for process ratings of general skill, supportive

interventions, and specific techniques. Oslo, Norway: Department of Psychiatry, University

of Oslo.

37 . Høglend, P. (2004). Canadian Journal of Psychoanalysis.

38 . Høglend, P., Amlo, S., Marble, A., Bøgwald, K-P., Sørbye, Ø., Sjaastad, M. C.,

& Heyerdahl, O. (2006). American Journal of Psychiatry.

39 . Høglend, P., Bøgwald, K-P., Amlo, S., Heyerdahl, O., Sørbye, Ø., Marble, A.,

& Bentsen, H. (2000). Journal of Psychotherapy Practice and Research.

40 . Høglend, P., Bøgwald, K-P., Amlo, S., Marble, A., Ulberg, R., Sjaastad, M. C.,

& Johansson, P. (2008). American Journal of Psychiatry.

41 . Horvath, A. O., Greenberg, L. S., Greenberg, L. S., & Pinsoff, W. M. (1986).

The psychotherapeutic process: A research handbook. New York, NY: Guilford Press.

42 . Horvath, A. O., & Greenberg, L. S. (1989). Journal of Counseling Psychology.

43 . Horvath, A. O., & Symonds, B. D. (1991). Journal of Counseling Psychology.

44 . Huprich, S. K., & Greenberg, R. P. (2003). Clinical Psychology Review.

45 . Kenny, D. A., Calsyn, R. J., Morse, G. A., Klinkenberg, W. D., Winter, J. P.,

& Trusty, M. L. (2004). Evaluation Review.

46 . Kernberg, O. F., Diamond, D., Yeomans, F. E., Clarkin, J. F., Levy, K. N., Jurist,

E., Slade, A., & Bergner, S. (2008). Mind to mind: Infant research, neuroscience

and psychoanalysis. New York, NY: Other Press.

47 . Luborsky, L. (1984). Principles of psychoanalytic psychotherapy. A manual for

supportive expressive treatment. New York, NY: Basic Books.

48 . Malan, D. H. (1976). Toward the validation of dynamic psychotherapy. New York,

NY: Plenum Press.

49 . Malan, D. H., & Ferruccio, O. (1992). Psychodynamics, training, and outcome

in brief psychotherapy. Oxford, England: Butterworth-Heinemann.

50 . Marmar, C. R., Gaston, L., Gallagher, D., & Thompson, L. W. (1989). Journal

of Nervous and Mental Disease.

51 . Marmar, C. R., Weiss, D. S., & Gaston, L. (1989). Psychological Assessment:

A Journal of Consulting and Clinical Psychology.

52 . McGlashan, T. H., & Miller, G. H. (1982). Archives of General Psychiatry.

53 . Messer, S., McWilliams, N., Castonguay, L. G., & Hill, C. E. (2007). Insight

in psychotherapy. Washington, DC: American Psychological Association.

54 . Messer, S., & Warren, S. (1995). Models of brief psychodynamic therapy: A comparative

approach. New York, NY: Guilford Press.

55 . Ogrodniczuk, J. S., Piper, W. E., Joyce, A. S., & McCallum, M. (1999). Journal

of Nervous and Mental Disease.

56 . Ogrodniczuk, J. S., Piper, W. E., Joyce, A. S., & McCallum, M. (2000). Canadian

Journal of Psychiatry.

57 . Piper, W. E., Azim, H. F., Joyce, A. S., & McCallum, M. (1991). Archives of

General Psychiatry.

58 . Piper, W. E., McCallum, M., Azim, H. F., & Joyce, A. S. (1993). American Journal

of Psychotherapy.

59 . Rosenthal, R., & Rosnow, R. L. (1991). Essentials of behavioral research: Methods

and data analysis. New York, NY: McGraw-Hill.

60 . Safran, J. D., & Muran, J. C. (2000). Negotiating the therapeutic alliance:

A relational treatment guide. New York, NY: Guilford Press.

61 . Sifneos, P. E. (1992). Short-term anxiety-provoking therapy: A treatment manual.

New York, NY: Basic Books.

62 . Spinhoven, P., Giesen-Bloo, J., van Dyck, R., Kooiman, K., & Arntz, A. (2007).

Journal of Consulting and Clinical Psychology.

63 . Spitzer, R. L. (1983). Comprehensive Psychiatry.

64 . Spitzer, R. L., Williams, J. B. W., Gibbon, M., & First, M. B. (1990). Structured

Clinical Interview for DSM-III-R Personality Disorders (SCID-II). Washington, DC:

American Psychiatric Publishing.

65 . Stiles, W. B., & Shapiro, D. A. (1994). Journal of Consulting and Clinical Psychology.

66 . Strachey, J. (1934). International Journal of Psychoanalysis.

67 . Strupp, H. H., & Binder, J. L. (1984). Psychotherapy in a new key: A guide to

time-limited dynamic psychotherapy. New York, NY: Basic Books.

68 . Svartberg, M., & Stiles, T. (1994). Psychotherapy Research.

69 . Tracey, T. J., & Kokotovic, A. M. (1989). Psychological Assessment: A Journal

of Consulting and Clinical Psychology.

70 . Verbeke, G., & Molenberg, G. (2000). Linear mixed models for longitudinal data.

New York, NY: Springer.

71 . Webb, C. A., DeRubeis, R. J., & Barber, J. P. (2010). Journal of Consulting

and Clinical Psychology.

72 . Weinberger, J. (1995). Clinical Psychology: Science and Practice.

Show less

Address for Correspondence: Per Høglend, Department of Psychiatry, Vinderen, University

of Oslo, P.O. Box 85, Vinderen, N-0319 Oslo, Norway

Email: p.a.hoglend at medisin.uio.no

© 2011 American Psychological Association

Subject: Object Relations (major); Psychotherapeutic Transference (major); Therapeutic

Alliance (major); Psychotherapy

Classification: 3310: Psychotherapy&Psychotherapeutic Counseling

Age: Adulthood (18 yrs&older)

Population: Human, Male, Female

Identifier (keyword): alliance, outcome, quality of object relations, transference

interpretation

Test and measure: Quality of Object Relations Scale, Psychodynamic Functioning Scales-Quality

of Family Relationships scale, Global Assessment of Functioning, Structured Clinical

interview for DSM-III-R,; Psychodynamic Functioning Scales-Quality of Friendships

scale, Psychodynamic Functioning Scales-Quality of Romantic/Sexual Relationships,

Psychodynamic Functioning Scales-Tolerance for Affects Scale,Psychodynamic Functioning

Scales-Insight scale, Psychodynamic Functioning Scales-Problem Solving Capability

scale

Methodology: Empirical Study, Quantitative Study

Title: Effects of transference work in the context of therapeutic alliance and quality

of object relations.

Author e-mail address: p.a.hoglend at medisin.uio.no

Contact individual: Høglend, Per, Department; of Psychiatry, University of Oslo,

Vinderen, P.O. Box 85, Vinderen, Olso, N-0319, Norway,p.a.hoglend at medisin.uio.no

Publication title: Journal of Consulting and Clinical Psychology

Volume: 79

Issue: 5

Pages: 697-706

Publication date: Oct 2011

Format covered: Electronic

Publisher: American Psychological Association

Country of publication: United States

ISSN: 0022-006X

eISSN: 1939-2117

Peer reviewed: Yes

Document type: Journal, Journal Article, Peer Reviewed Journal

Number of references: 72

Publication history :

Accepted date: 13 Jun 2011

Revised date: 31 May 2011

First submitted date: 23 Aug 2010

DOI: <a href="http://dx.doi.org/10.1037/a0024863">http://dx.doi.org/10.1037/a0024863</a>

Release date: 22 Aug 2011 (PsycINFO); ; 22 Aug 2011 (PsycARTICLES);

Correction date: 26 Sep 2011 (PsycINFO)

Accession number: 2011-18188-001

PubMed ID: 21859184

ProQuest document ID: 885055405

Document URL: https://login.libproxy.edmc.edu/login?url=http://search.proquest.com/docview/885055405?accountid=34899

Copyright: ©American Psychological Association 2011

Database: PsycARTICLES

____________________________________________________________

Contact us at: http://www.proquest.com/go/contactsupport

Copyright © 2012 ProQuest LLC. All rights reserved.

Terms and Conditions: http://search.proquest.com/info/termsAndConditions

Jessie Rayl
thedogmom63 at frontier.com
www.facebook.com/Eaglewings10
www.pathtogrowth.org
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.acb.org/pipermail/acb-hsp/attachments/20121119/fa77efc7/attachment-0001.html>