Do gender and level of relational

functioning influence the long-term

treatment response in dynamic
psychotherapy?
RANDI ULBERG, ALICE MARBLE, PER HØGLEND

Ulberg R, Marble A, Høglend P. Do gender and level of relational functioning influence

the long-term treatment response in dynamic psychotherapy? Nord J Psychiatry 2009;63:
412
419.

Background: Gender as a moderator of long-term treatment effects has to a very little extent
been explored in individual psychotherapy. We have previously reported a short-term difference
in treatment-response to transference interpretations between women with poor relational
functioning (low Quality of Object Relations Scale; low QOR) and men with good relational
functioning (high QOR). The present study focuses on whether there also is a sustained
difference in treatment-response between those two subgroups. Material and method: In the First
Experimental Study of Transference-interpretations (FEST), patients (n100) were randomized
to receive dynamic psychotherapy over 1 year with either a moderate level of transference
interpretations or no transference interpretations. Assessments were made at pre-treatment, mid-
treatment, post-treatment, and at 1- and 3-year follow-ups. The outcome measures used were the
Psychodynamic Functioning Scales (PFS), Inventory of Interpersonal Problems (IIP-C), Global
Assessment of Functioning (GAF) and Symptom Checklist-90 (GSI). Change was assessed
using linear
mixed models. Results: In the moderator analyses, women with low QOR showed a
significant positive long-term treatment effect of transference interpretation (P0.005), while
men with high QOR responded equally well to both therapies. Conclusion: Women with poor
relational functioning and men with good relational functioning showed sustained different
treatment-response to transference interpretations.
’ Gender, Psychodynamic, Psychotherapy, Sex, Transference.
Randi Ulberg, Department of Psychiatry, PO Box 85, Vinderen Diakonhjemmet Hospital,
N-0319 Oslo, Norway, E-mail: Randi.Ulberg@medisin.uio.no; Accepted 27 April 2009.

eviews on the impact of patient gender in psy-
R chotherapy indicate that men and women respond
similarly across different types of psychotherapy (1
3).
As far as we know, only two studies have reported a
moderator effect of gender in individual psychotherapy
(4, 5).
From the First Experimental Study of Transference-
interpretations (FEST), we have reported a moderator
effect of gender during psychotherapy on symptom
change (5). FEST was designed to explore the long-term
treatment effects of transference interpretation using
an experimental dismantling design. Analysis of trans-
ference is one of the cornerstones of techniques in
dynamic psychotherapy and distinguishes dynamic psy-
chotherapy from other forms of psychotherapy. It has
been debated for whom this technique is suitable. In
FEST, pre-treatment Quality of Object Relations (QOR)
(6, 7) was found to be an important moderator of the
treatment effects of transference interpretation (8, 9). The
patients with a history of less mature object relations
benefited more from dynamic psychotherapy with trans-
ference interpretations than from therapy without such
interpretations during treatment, and the treatment effect
was sustained during the follow-up period (follow-ups at
1 and 3 years after treatment termination). Gender had a
moderator effect over and above the moderator effect of
QOR on short-term symptom change. Women with poor
relational functioning benefited significantly more from
dynamic psychotherapy with transference interpreta-
tions. Conversely, men with good relational functioning
benefited more from dynamic psychotherapy without
transference interpretations. Therefore, the focus in the
present study is whether there was a sustained difference
in treatment-response between those two subgroups.

# 2009 Informa UK Ltd. (Informa Healthcare, Taylor & Francis As) DOI: 10.1080/08039480903009126

Aims
We wanted to explore whether the subgroup of women
with poor relational functioning (low QOR women) and
the subgroup of men with good relational functioning
(high QOR men) showed different long-term treatment
responses to transference interpretations.
Subjects and Methods
Participants
The patients referred to the study sought psychotherapy
for depressive disorders, anxiety disorders, personality
disorders and interpersonal problems not caused by a
mental disorder. Patients with psychosis, bipolar illness,
organic mental disorder or substance abuse were ex-
cluded. The study therapists assessed 122 patients for
eligibility and 100 were included in the study.
Method
Detailed descriptions of the FEST study’s research
design and methods have previously been published
(8, 9) and are briefly described here. Treatment effects in
this study mean the long-term effects of transference
interpretation.
TREATMENT
The Regional Committee, Health-Region 1, Norway,
approved the study protocol. Written informed consent
was obtained from each patient. After completion of the
pre-treatment ratings, the patients where consecutively
randomized to one of two treatment groups. The FEST
study is a dismantling study which means that one-half
of the patients (n 52) received dynamic psychotherapy
with a low to moderate use of transference interpreta-
tion. The other half of the patients (n48) received
dynamic psychotherapy of the same kind, with the same
therapists, but without transference interpretation. The
patients were offered 45-min weekly sessions for 1 year.
All sessions were audio-recorded. A treatment manual
was used (10). In the transference group, specific
techniques were prescribed, e.g. the therapist was to
address transactions in the patient
therapist relation-
ship, encourage exploration of thoughts and feelings
about the therapy and the therapist, and interpret direct
manifestations of transference. The therapist was also to
link conflicts and repetitive interpersonal patterns to
transactions between the patient and the therapist. In
the comparison group, these techniques were proscribed.
In this group, the therapist was to use material about
interpersonal relationships outside of therapy as the
basis for similar interventions (extra-transference inter-
pretations). Both treatments were mainly exploratory in
nature. Patients in both treatment groups were encour-
aged to explore sensitive topics, which often involved
uncomfortable emotions, but the therapist abstained
NORD J PSYCHIATRY ×VOL 63 ×NO 5×2009
GENDER DIFFERENCES IN RESPONSE TO PSYCHOTHERAPY
from providing guidance or giving advice, praise or
reassurance.
THERAPISTS
Patients were assigned to one of seven therapists,
depending on availability. The therapists, who also
served as clinical evaluators of other patients, included
six psychiatrists and one clinical psychologist. There
were five men and two women. All had 10
25 years of
experience in practicing dynamic psychotherapy. They
were all trained for up to 4 years, to give either treatment
with a moderate frequency of transference interpreta-
tions (one to three per session), or treatment with-
out such interpretations, with equal ease and mastery.
All therapists treated patients in both groups. Only the
patients’ therapists learned the result of the random
assignment procedure. Since no therapist ratings of their
own patients were included in any of the statistical
analyses, all raters were blinded.
ASSESSMENTS
Before the randomization, each patient had a 2-hour
psychodynamic interview, modified after Malan (11) and
Sifneos (12), which was audio-recorded. At least three
raters rated the interview using the QOR (6, 7) and the
Psychodynamic Functioning Scales (PFS) (13). In addi-
tion, the patients completed a number of self-reports,
including the Inventory of Interpersonal Problems*
Circumplex Version (IIP-C) (14). The Axis-I diagnoses
were based on clinical history and assessment of back-
ground variables by the patient’s therapist. Diagnoses
according to DSM-III-R criteria were discussed before
randomization, by the therapist and at least one of the
other clinicians until consensus was reached. Axis-II
diagnoses were determined using the Structured Clinical
Interview for the DSM-III-R (SCID-II) (15).
OUTCOME MEASURES
The primary outcome measure in the FEST study was the
mean of PFS. To capture statistically significant clinician-
rated psychodynamic changes and interpersonal func-
tioning during 1 year of psychotherapy, PFS were
developed in the pilot phase of the FEST study. These
six scales have the same format as the Global Assessment
of Functioning (GAF), and measure psychological
capacities over the previous 3 months. The scales are
Quality of Family Relationships, Quality of Friendships,
Quality of Romantic/Sexual Relationships, Tolerance for
Affects, Insight, and Problem Solving Capacity. Aspects
of content validity, internal domain construct validity,
inter-rater reliability, discriminant validity from symp-
tom measures, and sensitivity for change in brief dynamic
psychotherapy have been established (13, 16
18). Three
clinical raters, blind to treatment group, made evalua-
tions at pre-treatment, and again at post-treatment,
413
R ULBERG ET AL.
1-year follow-up and 3-year follow-up. The inter-rater
reliability estimates (intraclass correlation coefficient) for
average scores of three raters were about 0.90 for the PFS.
IIP-C total mean score (14) was used to assess patients’
self-reported interpersonal problems at pre-treatment,
mid-treatment (session 16), post-treatment, 1-year fol-
low-up, and 3-year follow-up. PFS and IIP-C measure
psychodynamic and interpersonal functioning. GAF (19)
and Symptom Checklist-90 (GSI) (20) measure severity
of psychiatric psychopathology and symptom distress. At
all evaluations, patients rated 24 life events. Additional
treatment, such as contact with mental health profes-
sionals, psychotherapy, psychopharmacological treat-
ment and sick leave, was carefully recorded.
MODERATORS
In this study, patient gender was a putative moderator.
The QOR (6, 7) was a pre-selected moderator in the
study protocol (P.Høglend, unpublished 1994 manual).
The QOR measures the patient’s lifelong tendency to
establish certain kinds of relationships with others, from
mature to primitive. The predetermined cut-off value
for high versus low QOR scores was 5.00. The 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. The inter-rater
reliability for the average scores of three raters was 0.84.
Statistical analysis
Standard power calculation (end-point analyses) indi-
cated that moderate effects between groups (ES 0.55)
could be detected with an alpha level of 0.05 and a power
of 0.80. An alpha level of 0.10 was decided a priori for
the moderator analyses and the subgroup analyses in
this study in order to offset the risk of Type-II errors
(21). One outlier in the transference group was deleted
from analyses of the longitudinal data as it became clear
during treatment, that this patient abused sedatives and
painkillers. Longitudinal intention-to-treat analyses
were performed on a sample of 99 patients.
We used linear
mixed models to analyze longitudinal
data (SPSS version 16.0) (22). A randomly distributed
intercept and slope (over time) was fitted for each
patient.
The highest rate of improvement was during therapy,
with diminishing returns over time. Time was coded
1, 2, 3, 5 and 9, with one step for each half-year, and
transformed to a natural logarithm. 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, while treatment group (coded 1, 0) was
treated as a fixed effect. The fixed effects in the moderator
414
analyses were intercept, time, time treatment, modera-
tors, and time treatment moderators. By using this
model, we assume that treatment group means are equal
at baseline, by design (23, 24). The three-way interaction
terms (time treatment moderator) test changes in the
treatment effects as moderator changes (tests of mod-
erator effects). It has previously been reported that QOR
was an important moderator of the long-term treatment
effect of transference interpretation (9). We therefore,
through several analyses, explored whether gender had
moderator effects over and above the effects of QOR. The
moderators were centered at different levels so that
the treatment effect (time treatment) can be interpreted
directly. In these analyses, timetreatment may be
interpreted as treatment effects for low QOR women
when QOR is centered at 4.41 (the average level in the low
QOR subgroup) and gender is coded as 0 women and
1 men, and for high QOR men when QOR is centered
at 5.6 (the average level in the high QOR subgroup) and
1 women and 0 men. Same or different gender
between patient and therapist (match
mismatch) was
coded as 0 same and 1 different and also coded as
1 same and 0 different, in order to explore whether
gender match
mismatch influenced the results.
Results
One patient assigned to the comparison group withdrew
from the study after the randomization. Four other
patients, also in the comparison group, dropped out of
therapy before session 15. All patients, also the drop-
outs, completed the 3-year follow-up evaluation.
Patient characteristics at baseline
We could not detect any significant differences between
treatment groups in the whole patient sample or between
women and men on the pre-treatment variables includ-
ing demographic, diagnostic, initial severity, personality,
interpersonal functioning and expectancy (5, 8). The
average of the QOR scores was equal between women
and men (women (n56) 5.1, men (n44) 5.0). Some
pre-treatment characteristics for low QOR women and
for high QOR men are presented in Table 1.
Therapist effects
We did not detect any differences in effectiveness
between therapists. Two random factors for therapists,
intercepts and slopes gave almost zero variance and were
therefore deleted from the analyses. It should be noted
that this study did not have the power to detect small-to-
moderate therapist effects.
NORD J PSYCHIATRY ×VOL 63×NO 5×2009
 GENDER DIFFERENCES IN RESPONSE TO PSYCHOTHERAPY

Table 1. Pre-treatment characteristics for low QOR (Quality of Analysis of interpersonal relationships outside of
Object Relations Scale) women and high QOR men receiving therapy (extra-transference) was given somewhat more
dynamic psychotherapy for 1 year with and without transfer-
ence interpretations space and emphasis in the comparison group.

Transference Comparison Moderator analyses

n11 n11 In Table 2, we included both QOR and gender as
moderators in the linear-mixed model long-term ana-
Low QOR women Mean s Mean s
lyses.
Age 32.6 7.3 34.9 9.0 Using PFS as outcome variable, controlling the low
Education 14.8 1.0 14.6 0.9 and high QOR Scale scores, we found that time 
Expectancy* 8.5 1.8 7.0 2.7
treatment was significant (P 0.005) when women were
Motivation$ 5.5 0.7 5.2 0.7
QOR% 4.5 0.4 4.3 0.5 coded 0, men were coded 1 and QORwas centered at 4.41.
That is, the subsample of low QOR women showed a
n % n % positive long-term treatment effect of transference inter-
pretations. On the other hand, when women were coded
Single 4 36 8 73
Personality disorder 7 63 6 60 1, men were coded 0 and QORwas centered at 5.6, time 
treatment was not significant, which indicates that the
n11 n10 high QOR men did not differ significantly in treatment
High QOR men Mean s Mean s
response between the two treatment groups. On both PFS
and GSI women had a significantly more positive effect
Age 40.8 9.6 40.4 9.8 of transference interpretations than men (time 
Education 15.1 1.0 15.5 0.8 treatment gender: PFS P 0.059; GSI P 0.040). The
Expectancy* 8.4 2.4 10.1 1.2
pattern of results was sustained when controlling for same
Motivation$ 5.2 0.5 5.6 0.5
QOR% 5.5 0.3 5.7 0.6 or different patient
therapist gender.
With IIP
C and GAF as outcome variables, no
n % n % significant difference in treatment response in the low
QOR women or the high QOR men could be detected.
Single 3 27 3 30
Personality disorder 3 27 1 10
Subsample analyses
s, standard deviation. In Table 3 model predicted ‘‘true’’ values over time for
*Expectancy: Expectation for improvement of target problems. the PFS in low QOR women and high QOR men are
$Motivation: Motivation for improvement and self-understanding. presented.
%Quality of Object Relations Scale*lifelong pattern.
Fig. 1 shows the ‘‘true’’, that is the model predicted
Treatment fidelity trajectories, of the PFS for the subsample of low QOR
Treatment length, without the drop-outs, was equal in the women in the transference group (n 11) and the
transference and comparison groups, 34 (standard devia- comparison group (n 11), and the subsample of high
tion, s6.1) and 33 (s 6.6) sessions on average, respec- QOR men in the transference group (n11) and in the
tively. A manual for process ratings of audio-recorded comparison group (n 10).
During the whole 4-year study period, low QOR
sessions included the therapist performance scales: Gen-
women in the transference group had a significantly
eral Interpersonal Skill, Interpretive and Supportive
better outcome (slope) than low QOR women in the
Technique Scale (25) and Specific Transference Technique
comparison group. The mean difference on the model
Scales (P. Høglend, unpublished 1995 manual). All items
predicted PFS was 2.8 at the end of treatment (P 
from these scales have similar rating instructions and 0.045), 3.6 at 1-year follow-up (P 0.027) and 4.5 at
format, i.e. 5-step Likert scales (0, not at all; 1, very little; 2, 3-year follow-up (P 0.018). The between-group effect
moderately; 3, considerably; 4, very much). The use of sizes (Cohen’s d were respectively 0.91, 1.02 and 1.09. No
specific transference techniques differed significantly significant difference in slopes between the two treat-
between the two treatment groups. The average score in ment groups was found for high QOR men, but the
the transference group was 1.7 (s 0.7, moderately used) trajectory for men in the comparison group is on a
and in the comparison group 0.1 (s0.2, nearly not at all higher level than the trajectory in the transference group.
used), a significant difference (t14.8, df58,2, PB It should be noted that despite the fact that low QOR
0.0005) between the treatment groups. There were no women in the transference group started on a signifi-
differences in the use of specific transference techniques cantly lower level on PFS than high QOR men in the
between women and men in the transference group. transference group, the trajectory for low QOR women

NORD J PSYCHIATRY ×VOL 63 ×NO 5×2009 415

R ULBERG ET AL.

Table 2. Gender as moderator of long-term treatment effects, controlled for Quality of Object Relations Scale (QOR)*

Dependent variable and parameter Estimate 90% CI t df P

Psychodynamic Functioning Scales:

Women coded 0 and men coded 1
Intercept 60.47 59.48 to 61.45 101.9 120,520 0.000
Time (log) 3.75 3.18 to 4.32 11.0 117,001 0.000
Timetreatment 1.87 0.79 to 2.94 2.9 99,712 0.005
QOR (score minus 4.41)$ 4.37 3.60 to 5.13 9.5 113,545 0.000
Gender 1.36 0.18 to 2.53 1.9 112,507 0.059
Timetreatment(QOR score minus 4.41) 1.20 1.94 to 0.47 2.7 109,465 0.008
Timetreatmentgender 1.28 2.39 to 0.17 1.9 110,446 0.059
Men coded 0 and women coded 1
Intercept 67.02 66.00 to 68.00 109.4 118,407 0.000
Time (log) 3.75 3.18 to 4.32 11.0 117,001 0.000
Timetreatment 0.84 1.90 to 0.21 1.3 102,677 0.187
QOR (score minus 5.6)% 4.37 3.60 to 5.13 9.5 113,545 0.000
Gender 1.36 2.53 to 0.17 1.9 112,507 0.059
Timetreatment(QOR score minus 5.6) 1.20 1.94 to 0.47 2.7 109,465 0.008
Timetreatmentgender 1.28 0.17 to 2.39 1.9 110,446 0.059
Global Severity Index:
Women coded 0 and men coded 1
Intercept 1.16 1.03 to 1.29 14.7 109,337 0.000
Time (log) 2.60 0.31 to 0.21 8.6 107,821 0.000
Timetreatment 0.03 0.13 to 0.06 0.6 98,354 0.544
QOR (score minus 4.41)$ 0.10 0.20 to 0.00 1.7 111,813 0.092
Gender 0.27 0.43 to 0.12 2.9 110,585 0.004
Timetreatment(QOR score minus 4.41) 0.02 0.09 to 0.04 0.5 111,265 0.601
Timetreatmentgender 0.12 0.02 to 0.22 2.1 107,834 0.040

*Patients received dynamic psychotherapy for 1 year with transference interpretations (n51, one outlier deleted) and without transference
interpretations (n 48). Linear-mixed models. Type III decomposition of variance components.
$QOR centered at the average value in the low QOR subgroup.
%QOR centered at the average value in the high QOR subgroup.

reached the same level as the trajectory for high QOR
men at the 3-year follow-up.
Discussion
We have previously reported that there was an interac-
tion effect between gender, level of relational functioning
and treatment during psychotherapy with transference
interpretations (5). In the present study, we explored
long-term effects in the two contrasting groups of
women with poor relational functioning and men with
good relational functioning. Measured with the PFS,
low QOR women showed a more favorable dynamic
change after psychotherapy with transference interpre-
tations than without transference interpretations. On the
other hand, high QOR men in the two treatment groups

Table 3. Psychodynamic Functioning Scale measured (PFS)* over time in patients receiving 1 year of dynamic psychotherapy
with or without transference interpretation

Transference Comparison
Time point Mean s Mean s 90% CI t df P

Low QOR women (n22)

Pre-treatment 61.5 2.0 60.3 2.9 0.63 to 3.0 1.13 20 0.268
Post-treatment 67.6 2.9 64.7 3.3 0.5 to 5.1 2.14 20 0.045
1-year follow-up 70.4 3.4 66.8 3.6 1.0 to 6.2 2.39 20 0.027
3-year follow-up 73.6 4.1 69.1 4.1 1.5 to 7.5 2.6 20 0.018
High QOR men (n21)
Pre-treatment 66.1 2.3 68.4 2.6 4.1 to 0.5 2.17 19 0.043
Post-treatment 69.4 2.8 72.8 2.9 5.6 to 1.3 2.75 19 0.013
1-year follow-up 70.7 3.2 74.9 3.1 6.4 to 1.6 2.90 19 0.009
3-year follow-up 72.6 3.6 77.2 3.4 7.2 to 1.9 3.0 19 0.007

*Modell predicted PFS.

416 NORD J PSYCHIATRY ×VOL 63×NO 5×2009

 GENDER DIFFERENCES IN RESPONSE TO PSYCHOTHERAPY

Women low QOR Transference (N=11) Women low QOR Comparison (N=11)
Men high QOR Transference (N=11) Men high QOR Comparison (N=10)
78

76

74

72

70
PFS

68

66

64

62

60
pre-treatment post-treatment 1-year follow-up 3-year follow-up
Time
Between groups effect sizes
(Cohen’s d)
Women: 0.91 1.02 1.09
P=0.268 P=0.045 P=0.027 P=0.018

Fig. 1. Trajectories of the Psychodynamic Functioning Scales (PFS)* for the transference group and comparison group within the
subsamples of female patients with low scores and male patients with high scores on the Quality of Object Relations Scale (QOR).
*Model predicted PFS, controlled for the effect of low and high QOR.

did not differ significantly in dynamic treatment re-
sponse (slopes) throughout the whole 4-year study
period. A significant difference in levels also at pre-
treatment might explain this non-significant result over
time. Measured with the PFS and GSI, we found that the
difference in treatment response between women and
men were sustained (moderator effects of gender).
To our knowledge, only one other experimental study
in addition to the FEST study has investigated possible
long-term interaction effects between patient gender and
different forms of individual psychotherapy. However,
the findings from FEST and the study from Ogrodniczuk
et al. (4) are not directly comparable. Ogrodniczuk and
colleagues explored the effect of patient gender in
interpretive and supportive psychotherapy at 1-year
follow-up (4), while the present study explored two forms
of interpretive psychotherapy at follow-ups 1 and 3 years
after treatment termination. The patient samples in the
two studies are also different. The present study and the
Ogrodniczuk study included patients with a broad range
of symptoms; 87% of the patients in FEST and 67% of
the patients in the Ogrodniczuk study received an Axis-I
diagnosis. In FEST, 46% of the patients received an Axis-
II diagnosis, whereas 60% of the patients in the Ogrod-
niczuk study were diagnosed with an Axis-II diagnosis.
The present study showed a long-term moderator effect
of gender. However, Ogrodniczuk found no significant
difference between the two genders at 1-year follow-up.
The main goal for transference interpretations is to
improve interpersonal functioning and internal psycho-
logical capacities. PFS is the primary mode specific
outcome measure. It is to be expected that a long-term
treatment effect of transference interpretations is found
for this measure and not for the other measures for low
QOR women.
We could detect no significant differences in the
amounts of positive or negative life events, use of
medication, consulting mental health professionals,
general practitioners or additional psychotherapy during
the follow-up period between men and women or
between low QOR women in the two treatment groups
or between high QOR men in the two treatment groups.
To have a low Quality of Relations Scale score
indicates having had no stable, mutual and fulfilling
relationships. It might be that poor relational female
patients benefited from transference interpretations by
learning a particular way of thinking about personal
relationships, and therefore continued their improvement
and maturity process after treatment termination (26).

NORD J PSYCHIATRY ×VOL 63 ×NO 5×2009 417

R ULBERG ET AL.
A good outcome case vignette
This case vignette illustrates a good outcome process in
a poor relational functioning women during dynamic
psychotherapy with transference interpretations:
A 25-year-old single female hairdresser was referred to
psychotherapy because of recurrent depression with
suicidal attempts. She had failed to respond to various
antidepressant medications. She had never had a stable
sexual relationship. She was diagnosed with major
recurrent depression, dysthymia and mixed personality
disorder. Her QOR score was low. She was randomized
to psychodynamic psychotherapy with transference in-
terpretations.
As a child, she was physically punished by her father.
She felt little closeness or understanding from her parents.
The family had limited ability to verbalize emotions. At
school, she feared the teachers and experienced criticism
from them as punishment and hostility, which led to
frequent absences from school.
Her anxiety problems started at the age of 13. She felt
that she could not meet the expectations from her parents
and teachers and experienced repetitive nightmares
about failing at school.
In the initial phase of the treatment, she expected that
the therapy would be a repetition of previous experi-
ences at home and at school and she repeatedly
expressed her wish to stop therapy. She seemed to evade
talking about the relationship between herself and the
therapist:
Patient: I need something else.
Therapist: Do you mean that you need another thera-
pist than me?
Patient: No, that was not what I meant!
As therapy progressed, the patient more often could
express her thoughts about the therapist and the therapy.
The patient gradually seemed to trust the therapist who
through transference interpretations linked the patient

therapist interaction to old family patterns and school
problems:
Therapist: You really feel it is a challenge to come here,
talking with me. You interpret what happens
also here with me in a very pessimistic way
like you have always done.
Patient: Yes, I don’t think you really care about me.
But I do feel it’s better today*when I have
been talking about stuff with you, it feels
like a burden is gone.
After the end of therapy and also at follow-ups,
she emphasized that the focus on the relationship
between the therapist and herself had been helpful.
These experiences encouraged her to explore unpleasant
feelings, especially that ‘‘he showed me that many of my
418
worries were irrational.’’ Symptom and dynamic im-
provement was good.
Limitations
We used a relatively small group of specifically trained
psychotherapists, which is good for internal validity but
less than optimal for generalizing to standard practice.
The patient sample had mixed diagnoses; the effects
on depression, for example, cannot be evaluated. On the
other hand, we believe that the patients are relatively
typical for patients seeking psychotherapy. The study
was not large enough for precise estimates of effect sizes.
Conclusion
Poor relational functioning women benefit much more
from transference interpretations than good relational
functioning men. The difference in responses between
women and men were sustained over the whole study
period. Our findings, if replicated, may help clinicians
tailor treatment techniques to different types of patients.
Future research should explore in detail session tran-
scripts in order to study the immediate effects of
transference interpretation for women and men.
Acknowledgements*The authors thank Svein Amlo, M.D.; Kjell-
Petter Bøgwald, M.D.; Øystein Sørbye, M.D.; Oscar Heyerdahl M.D.
and Mary Cosgrove Sjaastad M.D., for their contribution in peer
supervision and providing treatment data to the study and Professor of
statistics Inge Helland, Department of Mathematics, University of
Oslo, for supervising the statistical analyses.
Disclosure
The authors have no potential conflicts of interest to
disclose.
Funding/support
Supported by grants from Norwegian Research Council,
The Norwegian Council of Mental Health, Health and
Rehabilitation, Diakonhjemmet Hospital.
Trial registry name
First Experimental Study of Transference-interpreta-
tions (FEST).Registration number: NCT00423462
URL: https://register.clinicaltrials.gov
References
1. Clarkin JF, Levy KN. The influence of client variables on
psychotherapy. In: Lambert MJ, editor. Bergin and Garfield’s
handbook of psychotherapy and behaviour change, 5th edition.
New York: John Wiley; 2004. p. 194
226.
2. Lam AG, Sue S. Client diversity. Psychotherapy 2001;38:479
86.
/ /
3. Ogrodniczuk JS, Piper WE, Joyce AS. Differences in men’s and
women’s responses to short-term group psychotherapy. Psychother
Res 2004;14:231
43.
/ /
4. Ogrodniczuk JS, Piper WE, Joyce AS, McCallum M. Effect of
patient gender on outcome in two forms of short-term individual
psychotherapy. J Psychother Pract Res 2001;10:69
78.
/ /
5. Ulberg R, Johansson P, Marble A, Høglend P. Patient sex as
moderator of effects of transference interpretation in a randomized
NORD J PSYCHIATRY ×VOL 63×NO 5×2009

controlled study of dynamic psychotherapy. Can J Psychiatry 2009;
54:78
86.
/
6. Azim HFA, Piper WE, Segal PM, Nixon GWH, Duncan SC. The
Quality of Object Relations Scale. B Menninger Clin 1991;55:/
323
43.
7. Høglend P. Long-term effects of brief dynamic psychotherapy.
Psychother Res 2003;13:271
92.
/ /
8. Høglend P, Amlo S, Marble A, Bøgwald KP, Sørbye Ø, Sjaastad
MC, et al. Analysis of the patient
therapist relationship in dynamic
psychotherapy: An experimental study of transference interpreta-
tions. Am J Psychiatry 2006;163:1739
46.
/ /
9. Høglend P, Bøgwald KP, Amlo S, Marble A, Ulberg R, Sjaastad
MC, et al. Transference interpretations in dynamic psychotherapy:
Do they really yield sustained effects? Am J Psychiatry 2008;165:
/ /
763
71.
10. Høglend P. Dynamisk korttidsterapi (Brief Dynamic Psychother-
apy). In: Alnes R, Ekern P, Jarval P, editors. Poliklinikken
Psykiatrisk Klinikk 25 år. Oslo: Psykiatrisk Klinikk Vinderen,
Universitetet i Oslo; 1990. p. 27
38.
11. Malan DH. The frontier of brief psychotherapy: An example of the
convergence of research and clinical practice. New York: Plenum;
1976.
12. Sifneos PE. Short-term anxiety-provoking psychotherapy: A
treatment manual. New York: Basic Books; 1992.
13. Høglend P, Bøgwald KP, Amlo S, Heyerdahl O, Sørbye O, Marble
A, et al. Assessment of change in dynamic psychotherapy. J
Psychother Pract Res 2000;9:190
9.
/ /
14. Alden LE, Wiggins JS, Pincus AL. Construction of circumplex
scales for the Inventory of Interpersonal Problems. J Pers Assess
1990;55:521
36.
/ /
15. Spitzer RL, Williams JBW, Gibbon M, First MB. Structured
Clinical Interview for DSM-III-R Personality Disorders (SCID-
II). Washington DC: American Psychiatric Press; 1990.
NORD J PSYCHIATRY ×VOL 63 ×NO 5×2009
GENDER DIFFERENCES IN RESPONSE TO PSYCHOTHERAPY
/ 16. Bøgwald KP, Dahlbender RW. Procedures for testing some aspects
of the content validity of the Psychodynamic Functioning Scales
and the Global Assessment of Functioning Scale. Psychother Res
/
2004;14:453
68.
/ /
17. Hersoug AG. Assessment of therapists and patients’ personality:
Relationship to therapeutic technique and outcome in brief
dynamic psychotherapy. J Pers Assess 2004;83:191
200.
/ /
18. Hagtvet KA, Høglend P. Assessing precision of change scores
in psychodynamic psychotherapy: A generalizability theory
approach. Meas Eval Couns Dev 2008;41:162
79.
/ /
19. Diagnostic and statistical manual of mental disorders, 3rd edition.
Washington DC: American Psychiatric Association; 1987.
20. Derogatis LR. SCL-90-R: Administration, scoring and procedures
manual II. Towson, MD: Clinical Psychometric Research; 1983.
21. Kraemer HC, Wilson GT, Fairburn CG, Agras WS. Mediators and
moderators of treatment effects in randomized clinical trials. Arch
Gen Psychiatry 2002;59:877
83.
/ /
22. SPSS Inc. Advanced Models 16.00. Chicago; SPSS Inc.; 2007.
23. Fitzmaurice G, Laird NM, Ware JH. Applied longitudinal
analysis. Hoboken, NJ: Wiley-Interscience; 2004.
24. Singer JD, Willet JB. Applied longitudinal data analysis. modeling
change and event occurrence. New York: Oxford University Press,
Inc.; 2003.
25. Ogrodniczuk JS, Piper WE. Measuring therapist technique in
psychodynamic psychotherapies. Development and use of a new
scale. J Psychother Pract Res 1999;8:142
54.
/ /
26. Gabbard GO. Long-term psychodynamic psychotherapy: A basic
text. Washington DC: American Psychiatric Publishing; 2004.
Randi Ulberg, M.D., Diakonhjemmet Hospital, Oslo, Norway.
Alice Marble, Psy.D., Department of Psychiatry, Vinderen,
University of Oslo, Norway.
Per Høglend, Ph.D., Department of Psychiatry, Vinderen, University
of Oslo, Norway.
419