Summary notes are offered.

Psychotherapy: Theory, Research, Practice, Training Copyright 2007 by the American Psychological Association
2007, Vol. 44, No. 1, 1­13 0033-3204/07/$12.00 DOI: 10.1037/0033-3204.44.1.1




PSYCHOANALYSIS WITH AVOIDANT PERSONALITY
DISORDER: A SYSTEMATIC CASE STUDY

JOHN H. PORCERELLI V. BARRY DAUPHIN
Wayne State University School of Medicine University of Detroit Mercy


J. STUART ABLON SUSAN LEITMAN
Massachusetts General Hospital, Harvard Private Practice
Medical School Bloomfield Hills, MI


MICHAEL BAMBERY
University of Detroit Mercy
This case study of process and outcome Process data was gathered to determine
is based upon data obtained during a the extent to which the treatment ad-
5-year psychoanalysis of an adult male hered to prototypes of psychodynamic,
with avoidant personality disorder cognitive­ behavioral, and interpersonal
(AVPD). To date, no known systematic therapy. Results indicated that the pa-
case studies, effectiveness studies, or tient achieved clinically significant re-
randomized control trials exist for psy- ductions in PD, symptom severity, and
choanalysis in the treatment of AVPD. relational pathology. Gains were main-
In this study, self-reported symptoms tained at 1-year follow-up. The treat-
and observer-rated personality disorder ment significantly adhered to psychody-
(PD), global functioning, object rela- namic principles throughout, with some
tions, and psychological health were use of cognitive­ behavioral and inter-
gathered at intake, after each year of personal principles in the third year of
treatment, and at 1-year follow-up. treatment. These findings warrant fur-
ther investigation of psychoanalysis for
AVPD and demonstrate the usefulness
John H. Porcerelli, Department of Family Medicine and of assessing multiple domains of patient
Public Health Sciences, Wayne State University School of functioning and treatment process.
Medicine, Detroit, Michigan; V. Barry Dauphin, Department
of Psychology, University of Detroit Mercy, Detroit, Michi-
gan; J. Stuart Ablon, Department of Psychiatry, Massachu- Keywords: psychoanalysis, avoidant
setts General Hospital, and Harvard Medical School, Boston, personality disorder, case study, treat-
Massachusetts; Susan Leitman, Private Practice, Bloomfield ment outcome
Hills, Michigan; Michael Bambery, Doctoral Student, Depart-
ment of Psychology, University of Detroit Mercy, Detroit,
Michigan. Psychoanalytic clinicians have long been inter-
This study was funded through a grant from the Fund for
ested in and have treated individuals with
Psychoanalytic Research, American Psychoanalytic Associa-
tion.
avoidant personality pathology (one of the many
Correspondence regarding this article should be addressed forms of object relations pathology). Surpris-
to John H. Porcerelli, Department of Family Medicine and ingly, however, there are no known systematic
Public Health Sciences, 15400 W. McNichols, Room 210, case studies or effectiveness studies of psycho-
Detroit, MI 48235. E-mail: jporcer@med.wayne.edu analysis for avoidant personality disorder



1
Porcerelli et al.


(AVPD). In fact, Fonagy (2002) has indicated than no-PD groups. They did not attempt to mea-
that there are "no definitive studies that show sure any therapeutic effects on PD symptoms per
psychoanalysis to be unequivocally effective rel- se. Effect sizes for symptoms and functioning
ative to an active placebo or an alternative measures were large for both treatments at termi-
method of treatment" (p. 287) for any particular nation. However, the effect sizes for CBT were
disorder. There are only five studies assessing the larger than those for PI.
effectiveness of time-limited psychodynamic Hoglend (1993) studied short-term psychody-
psychotherapy for Cluster C personality disorders namic psychotherapy therapy (9 to 53 sessions)
(PD), which includes avoidant, obsessive­ for patients without PDs (n 30, 20 of whom
compulsive, and dependent PDs, or for AVPD itself had Axis I disorders) and with PDs (n 15, 8 of
(Barber, Morse, Krakauer, Chittams, and Crits- whom had Cluster C PDs). Six of 15 PD patients
Christoph, 1997; Hardy et al., 1995; Hoglend, 1993; no longer met criteria for PD at 4-year follow-up,
Svartberg, Stiles, & Seltzer, 2004; Winston et al., and none of the 3 patients with AVPD showed
1994), and only one study assessing the effective- significant intrapsychic improvement, whereas 2
ness Cluster C PDs with long-term psychodynamic showed significant Axis I symptom improve-
therapy. This is somewhat surprising given that ment.
AVPD is one of the most prevalent of all PDs in Monsen and colleagues (1995) treated 25 pa-
the general population (Torgersen, Kringlen, & tients with a PD and Axis I diagnosis with psy-
Cramer, 2001; Samuels et al., 2002; Coid, Yang, chodynamic psychotherapy. Treatment length av-
Tirer, Roberts, & Ullrich, 2006), is associated eraged 24 months, and the mean follow-up
with disability (Grant et al., 2004), and impedes assessments occurred at 5.2 years. At the end of
recovery from Axis I disorders (Hardy et al., treatment, 75% and 72% of patients no longer
1995; Reich, 2003; Shea et al., 2004; Viinamaki met DSM­III­R criteria for Axis I and Axis II
et al., 2002; Viinamaki et al., 2003). disorders, respectively. These findings essentially
Barber and colleagues (1997) assessed the ef- held up at follow-up. However, only 4 patients
fectiveness of time-limited (52-session) met Cluster C PD criteria, and the only patient
supportive­ expressive psychodynamic psycho- who met AVPD criteria at intake still had the
therapy (Luborsky, 1984) in an open trial with diagnosis at termination and follow-up.
patients diagnosed with DSM­III­R AVPD (n Svartberg and colleagues (2004) conducted a
24) and obsessive­ compulsive PD (OCPD) (n randomized trial of 40-session psychodynamic
14). Patients also met criteria for at least one Axis (n 25) and cognitive (n 25) psychotherapy
I disorder (anxiety or depression). Although both of patients with DSM­III­R Cluster C PDs, 31 of
treatment groups improved across time on symp- whom met criteria for AVPD. Most of the pa-
tom and personality measures, patients with tients (94%) also met criteria for an Axis I dis-
OCPD lost their PD diagnosis more quickly, order. Patients in both treatment groups exhibited
fewer met criteria for a PD at the end of treatment significant changes in interpersonal problems and
(15.4% vs. 38.5%), and OCPD patients had a dimensional personality scales pre- and posttreat-
significantly lower dropout rate (7% vs. 54%) ment. Patients in psychodynamic therapy showed
than did patients with AVPD. Barber et al. (1997) significant decreases in symptom distress at the
concluded that supportive­ expressive psychody- end of treatment, whereas the patients in cogni-
namic psychotherapy was more effective for pa- tive therapy did not. However, scores on all mea-
tients with OCPD than for patients with AVPD. sures decreased at the 2-year follow-up. Symp-
Hardy and colleagues (1995) compared pa- tomatic recovery rates at follow-up were 54%
tients with Cluster C PDs (n 13) and depres- and 42% for the psychodynamic and cognitive
sion with patients absent a PD diagnosis (n 44) therapy groups, respectively. Recovery rates
for 8- or 16-week psychodynamic­interpersonal from interpersonal problems were approximately
(PI) or cognitive­ behavioral therapy (CBT). 40%. Svartberg and colleagues (2004) concluded
They found that PD patients had significantly that both forms of therapy were effective in the
poorer outcomes in PI therapy than in CBT, and treatment of Cluster C PDs. Unfortunately, sep-
that PD patients had significantly worse out- arate analyses were not conducted on patients
comes than no-PD for PI therapy but not for with AVPD.
CBT. Depression severity appeared to interact Winston and colleagues (1994) randomized 81
with PD, resulting in poorer outcomes for PD patients with a DSM­III­R PD to brief dynamic/



2
Avoidant Personality Disorder


affective-focused or adaptive/cognitive-focused live with his anxiety. At the age of 50, with the
psychodynamic psychotherapy and a waiting-list support of his wife, the patient sought psycho-
control group. Patients were seen for an average analysis at four times per week.
of 40 sessions. Thirty-six (44.4%) of the patients The purpose of this systematic case study is to
had an AVPD. More than one half of patients met begin to assess the effectiveness of psychoanaly-
criteria for an Axis I disorder. Both treatments sis for AVPD with comorbid Axis I pathology
groups showed significant differences pre- and using a longitudinal design. Our methodology
posttreatment on all of the outcome measures as included yearly and follow-up assessments of
compared to no differences in the waiting-list Axis II pathology along with Axis I diagnosis,
group. Thirty-eight (47%) of patients were avail- symptom severity, pathological and healthy as-
able for follow-up (average of 1.5 years) for pects of functioning, and internalized object re-
target complaint assessments. Target symptom lations. In addition to a self-report measure of
ratings did not differ between the treatment symptom severity, independent external raters
groups, but both groups showed additional im- completed all other outcome assessments.
provement from scores at the end of treatment. Based upon the phase model of psychotherapy
Thus, both treatments were effective for treating (Howard, Maling, & Martinovich, 1993; Howard,
Cluster C PDs. Moras, Brill, Martinovich, & Lutz, 1996; Kopta,
Each of these psychotherapy studies for AVPD Howard, Lowry, & Beutler, 1994; Lueger, Lutz,
included many patients with comorbid Axis I & Howard, 2000), we hypothesized that the pa-
disorders. Thus the findings are likely to be more tient would evidence clinically significant
generalizable to real-world clinical practice than changes in symptom severity/global functioning,
studies that attempt to isolate symptoms using followed by clinically significant changes in per-
stringent exclusion criteria. In sum, the findings sonality pathology, including internalized object
from these six studies suggest that short-term relations.
psychodynamic psychotherapy may be a viable
approach to treating AVPD. To the best of our
knowledge, there are no studies supporting the Method
effectiveness of long-term psychodynamic psy-
chotherapy or psychoanalysis for AVPD. The Patient
Although short-term psychodynamic psycho- Mr. A, a computer technician, was 50 years old
therapy may be helpful for some patients with when he began treatment. His immediate impetus
AVPD, clearly not all patients benefit from this for being evaluated for psychoanalysis was his
approach. More intensive psychotherapies, in- long-standing anxiety and fears about driving (es-
cluding psychoanalysis, may have a place in the pecially on expressways). However, his wife (of
treatment of patients who do not improve in 7 years) was also urging him to try psychoanal-
other treatments (Fonagy, Roth, & Higgitt, 2005; ysis because she found living with him to be
Gabbard, Gunderson, & Fonagy, 2002). More- increasingly difficult. Problems that unfolded as a
over, it is important to study patient change with function of intensive treatment included: hostility
more complexity because psychoanalysis empha- toward his wife, who was disabled from a slowly
sizes multiple layers of meaning to people's ex- but progressively deteriorating neuromuscular
periences (Blatt & Auerbach, 2003), information condition, inhibitions at work with regard to ad-
that is often lacking in many large outcome stud- vancement, lifelong feelings of inadequacy, so-
ies. In the present case study, the patient sought cial inhibitions, sensitivity to criticism, and sex-
psychoanalysis because he had tried multiple ual conflicts.
therapies: crisis therapy plus medication in his
late teens, psychodynamic and eclectic therapy in
his 20s, behavior therapy in his 30s, and support- The Therapist
ive psychotherapy in his early 40s. Symptom
relief from these treatments was temporary at The therapist (J. H. P.) was a middle-aged
best, and the patient continued to have disabling Caucasian male in independent practice in the
symptoms (e.g., a phobia of driving on express- midwestern United States. He holds a doctorate
ways) and major interpersonal problems. He re- in clinical psychology from an American Psycho-
signed himself to the fact that he would have to logical Association­accredited graduate program,



3
Porcerelli et al.


is board certified in clinical psychology by the designed to assess recent psychiatric symptoms
American Board of Professional Psychology, and and includes nine symptom subscales and a mea-
is a graduate of a psychoanalytic institute accred- sure of global symptom severity (Global Severity
ited by the American Psychoanalytic Association. Index; GSI). Higher scores represent greater psy-
chological distress. A t score of 65 or higher is
The Treatment considered positive for psychopathology. The
mean GSI raw scores for a normal population
The treatment was psychoanalysis four times (N 974) is .30 (SD .31) and test­retest
per week. After the evaluation sessions, the pa- reliability for a 2-week period utilizing a nonpa-
tient chose to use the analytic couch. During the tient sample was .90. The BSI is a reliable and
last year of treatment, the patient reduced his valid measure of psychological distress for both
sessions to three times per week. The patient research and clinical settings. For this study, only
negotiated a reduced fee (three fourths of the the GSI was used to track symptom severity.
therapist's standard fee), which was paid for
through a combination of private funds and Outcome Measures (External Raters)
health insurance.
Two experienced clinical psychologists
Procedures (V. B. D. and S. L.) in independent practice
independently rated all session transcripts using
Mr. A. agreed to complete the Brief Symptom the Global Assessment of Functioning Scale
Inventory (BSI) at intake, after every year of (GAF; American Psychiatric Association, 1994),
therapy and at 1-year and 2-year follow-up. Re- Social Cognition and Object Relations Scale--
sults from the BSI were always discussed as part Global (SCORS-G; Westen, 1995), and the
of the treatment. Mr. A gave his signed consent to Shedler­Westen Assessment Procedure--200
the therapist's use of audiotaping for both clinical (SWAP-200; Shedler & Westen, 1998). Both rat-
and research purposes. He declined the need to ers had experience in coding clinical data for
read any article that included any material about research purposes. Prior to using these instru-
him and his treatment. He was assured that any ments, each rater read the relevant literature for
presentation or article would omit identifying in- each scale, studied each of the scales, and applied
formation. The patient agreed to the taping of 25 them to current cases in their practices. Raters
sessions, which included 4 intake sessions, 4 ses- were given the intake sessions first. Sessions
sions at the end of each year, and 1 follow-up from each year and follow-up were given in ran-
session. Three of the 4 sessions at the end of each dom order. Sessions were not disguised, and
year with the best audio quality were transcribed therefore it is likely that the raters were aware of
and used for the study. Therefore, 4 intake ses- coding sessions close to the termination of the
sions, 15 therapy sessions, and 1 follow-up ses- treatment and the 1-year follow-up session. They
sion were used to assess personality pathology, were unaware of the details of the case at the time
global functioning, and object relations by two of coding. The means of the raters' scores were
independent raters (experienced clinical psychol- used for data presentation.
ogists). The 15 therapy sessions were also used to Global Assessment of Functioning Scale
assess psychotherapy process by two different (GAF; American Psychiatric Association, 1994).
raters (the therapist and an advanced doctoral The GAF (Axis V of DSM­IV) is a measure of
student). All process and outcome data were dou- overall functioning designed to track clinical
ble coded for interrater reliability. Raters used the progress. Clinicians are to take into account a
entire session for rating all observer-based pro- patient's psychological, social, and occupational
cess and outcome measures. Raters' responses functioning at the time of the assessment and rate
were averaged for all outcome measures for final them on a scale from 0 to 100, with higher scores
data presentation. indicative of better functioning. Scores from 1 to
50 indicate severe psychopathology and severe
Outcome Measure (Patient Self-Report) impairment in social, occupational, or school
functioning; 51 to 70 indicates moderate to mild
Brief Symptom Inventory (BSI; Derogatis, symptom severity and moderate to mild func-
1993). The BSI is a 53-item self-report scale tional impairment; 71 to 80 indicate transient



4
Avoidant Personality Disorder


symptoms and slight impairment; 81 to 90 indi- (SD .25), and interrater reliability was .85. For
cate absent or minimal symptoms and generally this study, only the AVPD and Health Functioning
good functioning; 91 to 100 indicates an absence scales are reported.
of symptoms and superior functioning. The GAF Social Cognition and Object Relations Scale--
has demonstrated moderate to high interrater Global (SCORS-G; Westen, 1995; Hilsenroth,
reliability and validity (Goldman, Skodol, & Stein, & Pinsker, 2004). The SCORS-G is a
Lave, 1992; Hilsenroth et al., 2000). measure of global dimensions of interpersonal
Shedler­Westen Assessment Procedure--200 representations for narrative data, including
(SWAP-200; Shedler & Westen, 1998). This is psychotherapy sessions. The scale integrates
a Q-sort instrument that includes 200 descriptive object relations, social­ cognitive, and develop-
statements describing both pathological and mental theories. The SCORS-G comprises
health aspects of personality. The statements are eight dimensions: Complexity of representa-
sorted into eight categories, ranging from 0 (ir-
tions (Complexity) refers to the degree of self/
relevant to the patient) to 7 (highly descriptive of
other differentiation and degree of complexity
the patient). SWAP-200 statements are written in
a manner close to the data (e.g., "Tends to be in which others are experienced; Affect­tone of
passive and unassertive" or "Living arrangements relationship paradigms (Affect­tone) refers to
are chaotic and unstable"), and items that require the overall affective quality of interpersonal
inference about internal processes are written in representations from malevolent to benevolent;
clear and unambiguous language (e.g., "Is unable Capacity for emotional investment in relation-
to describe important others in a way that con- ships (Relationships) refers to the degree of
veys a sense of who they are as people; descrip- mutuality (vs. ego-centricity) of relationships;
tions lack fullness and color" or "Tends to blame Capacity for emotional investment in values
others for own failures or shortcomings; tends to and morals (Morals) refers to the degree to
believe his or her problems are caused by exter- which moral issues are considered in relation-
nal factors"). Reliable descriptions with the ships; Understanding social causality refers to
SWAP-200 have been obtained from clinicians the degree to which thought, feeling, and be-
from a variety of theoretical orientations (Westen havior is logical, accurate, and psychologically
& Shedler, 1999a, 1999b). Clinician ratings are minded. Experience and management of ag-
converted to t scores (M 50; SD 10) for each gression (Aggression) refers to the degree to
of the DSM­IV PDs. The SWAP-200 also in- which aggression is appropriately controlled
cludes a Healthy Functioning scale, a dimen- and expressed. Self-esteem refers to degree to
sional measure of psychological strengths and which the self is experienced as generally pos-
adaptive functioning. T scores from 55 to 59 itive (vs. self-loathing); Identity and self-
indicate PD features, whereas a t score of 60 is coherence (Identity) refers to the degree to
the cutoff for PD (J. Shedler, personal communi- which identity is stable, enduring, and purpose-
cation, January 20, 2003). Thus the scale can be ful (vs. fragmented, unstable, and inconsistent).
used categorically and/or dimensionally. The Each dimension is rated on a 7-point scale
SWAP-200 scales have good internal consistency
where scores of 1 or 2 indicate immature/
(Westen & Shedler, 1999b), interrater reliability
pathological object relations and scores of 6
(Marin-Avellan, McGauley, Campbell, & Fonagy,
2005; Westen & Muderrisoglu, 2003, 2006), and and 7 indicate mature/healthy object relations.
convergent/discriminant (Marin-Avellan et al., Each of the SCORS-G dimensions has demon-
2005; Westen & Shedler, 1999a), incremental strated good interrater reliability and validity
(Westen & Harnden-Fischer, 2001), and known (Huprich & Greenberg, 2003; Stricker &
groups validity (Porcerelli, Cogan, & Hibbard, Gooen-Piels, 2004). Peters, Hilsenroth, Eudell,
2004). The SWAP-200 has demonstrated sensitivity Blagys, and Handler (2006) have reported the
to changes brought about in an intensive psychoan- reliability and convergent validity of the
alytic psychotherapy of a patient with Axis I (sub- SCORS, as rated through relational narrative
stance abuse) and Axis II (borderline personality and self-statements told during psychotherapy
disorder) pathology (Lingiardi, Shedler, & Gazzillo, sessions, with DSM­IV Axis V measures of
2006). The mean SWAP-200 AVPD raw score in global functioning in outpatients undergoing
the standardization sample (N 530) was .29 short-term psychodynamic psychotherapy.



5
Porcerelli et al.


DSM­IV Axis I Diagnosis strated over several studies and treatment sam-
ples (Ablon & Jones, 2002). Ablon and Jones
The patient was diagnosed with DSM­IV specific (1998, 1999, 2002) developed prototypes for psy-
phobia, situational type, by both the therapist chodynamic, cognitive­ behavioral, and interper-
(J. H. P.) following the four evaluation sessions and sonal therapy by having experts in each of these
by an independent rater (V. B. D). theoretical orientations rate an "ideal" therapy
process for their respective therapies. By corre-
Process Measure lating the 100 PQS items with PQS ratings from
actual therapy sessions, it can be determined the
Two raters, an advanced psychology graduate degree to which a session adhered to a particular
student (M. B.) and the therapist (J. P.), indepen- brand of therapy. Examples of the top four psy-
dently rated each of 15 transcribed session for years chodynamic therapy prototype items are "Pa-
1 through 5 using the Psychotherapy Q-Set (PQS; tient's dreams and fantasies are discussed" (Item
Jones, 2000). The raters read the relevant PQS 90); "Therapist is neutral" (Item 93); "Therapist
literature and coded five psychotherapy sessions points outpatient's use of defensive maneuvers"
from another psychotherapy case. Discrepancies in (Item 36); and "Therapist draws connections be-
scoring were discussed. The average interrater reli- tween therapeutic relationship and other relation-
ability for the five practice cases was greater than ships" (Item 100). Examples of the top four
.70. To eliminate therapist bias, only process ratings cognitive­ behavioral therapy prototypes items
from the graduate student were used to assess treat- are "There is discussion of specific activities or
ment fidelity. tasks for the patient to attempt outside of session"
Psychotherapy Q-Set (PQS; Jones, 2000). (Item 38); "Discussion centers on cognitive
The PSQ is a pantheoretical 100-item Q-sort rat- themes, that is, about ideas or belief systems"
ing scale designed to provide a comprehensive (Item 30); "Patient's treatment goals are dis-
description of therapist­patient interactions (ther- cussed" (Item 4); and "Therapist encourages pa-
apist interventions and attitudes, patient attitudes tient to try new ways of behaving with others"
and behaviors, and therapist/patient interaction). (Item 85). Examples of the top four interpersonal
The PQS can be used with videotaped or audio- therapy prototypes are "Patient's interpersonal
taped psychotherapy sessions. Entire sessions are relationships are a major theme" (Item 63);
coded, not just a portion of the session, in order to "Love or romantic relationships are a topic of
capture the complexity of the treatment process. discussion" (Item 64); "Patient's current or recent
Coders sort the 100 items along a 9-point contin- life situation is emphasized in discussion" (Item
uum. The most characteristic items are placed in 69); and "Therapist asks for more information or
Category 9, neutral (or irrelevant) items are elaboration" (Item 31).
placed in Category 5, and the least characteristic
items are placed in Category 1. Items placed into Determining Clinically Significant Change
the least characteristic category are important be-
cause they indicate what aspects of the treatment We employed criteria established by Jacobson
and interaction are not present in a given therapy and Truax (1991), which include Reliable
session. The placement of the 100 items along the Change Index (RCI 1.96) and return to a func-
9-point continuum is governed by a fixed distri- tional distribution or out of a dysfunctional dis-
bution approximating a normal curve. Only 5 tribution (SD 2.0). An RCI was calculated for
items can be placed at each end of the continuum each of the outcome variables with a slight mod-
(Categories 1 and 9), whereas 18 items can be ification to the formula as suggested by Wise
placed in the center (Category 5). A fixed distri- (2004). Because this is a case study, pretreatment
bution is a hallmark of the Q-technique and re- mean scores and SDs were borrowed from other
quires coders to make multiple evaluations studies for the GAF and SCORS. Because there
among items and therefore avoid response sets are no nonpatient norms for the GAF scale, the
and halo effects. Interrater reliability is computed SD and mean from 41 patients beginning psycho-
by correlating (intraclass correlation; ICC) PSQ analysis for personality pathology (Cogan &
ratings from two independent coders for the same Porcerelli, 2006) were used to calculate change in
psychotherapy session. The reliability and dis- SD units. In that sample, the pretreatment mean
criminant validity for the PQS has been demon- was 67.80, and SD was 11.70. Nonpatient norms



6
Avoidant Personality Disorder


do not exist for the SCORS-G. However, in a the actual process ratings and the cognitive­
sample of 90 outpatients (all with Axis I diag- behavioral and interpersonal therapy prototypes
noses and 48 with comorbid Axis II diagnoses), were statistically significant (p .05), indicating
Peters and colleagues (2006) reported means that the process at that point in the treatment was
(3.12 for Self-Esteem to 4.02 for Complexity) not solely psychodynamic. A composite profile of
and SDs (.81 for Self-Esteem to 1.17 for Aggres- the set of three yearly sessions across each of 5
sion) for SCORS's dimensions from psychother- years of treatment (15 total sessions) of the most
apy sessions. Thus, we set SDs for all SCORS characteristic and least characteristic processes is
dimensions at 1.00. reported in Table 2.

Results Outcome
Interrater Agreement for Outcome and Process Table 3 displays the results of the outcome
Measures measures as rated by the patient and external
observers at the end of each year of treatment and
Interrater agreements for the outcome and pro- at follow-up.
cess measures were calculated using ICC (two-
way random effects). The mean ICC for the Symptom Severity, Global Functioning, and
SWAP-200 was .68 (.81 with Spearman­Brown Strengths
correction for double coding). The ICC for the
GAF scale was .67 (.80 corrected), and the ICC With regard to symptom severity, BSI Global
for the SCORS dimensions ranged from .54 to Severity Index t scores were in the pathological
.97 (.70 to .98 corrected). The mean ICC for the range from intake through Year 3 of treatment. T
process ratings for the PQS was .82 (.90 cor- scores were in the nonclinical range at Year 4
rected; range .76 to .87). through 2 years posttreatment. Changes in raw
BSI scores between intake, Years 4 and 5, and
Treatment Process: Did the Treatment Conform follow-up reached clinical significance. GAF
to the Psychoanalytic Prototype? scores changed from the "moderate" symptom
severity range (51­ 60) to the "mild" range (61 ­
Correlations between the sessions of psycho- 70) at Year 2 and into the "transient" symptom
analysis and expert-developed psychodynamic, range (71 ­ 80) at Year 5. They were maintained
cognitive­ behavioral, and interpersonal ideal at the 1-year follow-up assessment. Changes in
prototypes are presented in Table 1. These results GAF scores from intake to Year 5 and 1-year
indicate that a strong psychodynamic process was follow-up were clinically significant. Psycholog-
demonstrated from three sessions at the end of ical strengths, as indicated by SWAP-200 High
each year of treatment. Significant correlations Functioning scale scores, evidenced clinically
(p .001) ranged from .49 to .62. At the end of significant changes by Year 4 through follow-up.
the third year of treatment, correlations between
Personality Disorder
TABLE 1. Correlations Between Psychoanalytic Therapy
Sessions and Psychodynamic, Cognitive-Behavioral, and Mr. A met SWAP-200 criteria for AVPD with
Interpersonal Therapy Prototypes a t score of 60.50 at intake. At the end of Year 1
of treatment, he no longer met criteria for the
Treatment prototypes disorder. However, in Years 3 and 4, he reached
Cognitive- the "features" range with t scores of 55 and 59,
Year Psychodynamic Behavioral Interpersonal respectively. At Year 5, he no longer exhibited
1 49*** .14 .07
AVPD features. Changes in Mr. A's SWAP-200
2 56*** .04 .01 AVPD score at follow-up were clinically signif-
3 62*** .23* .24* icant.
4 62*** .03 .01
5 .56*** .12 .06
Composite of
Specific Phobia
Years 1­5 .64*** .04 .05
Mr. A reported having avoided driving ex-
*
p .05. ***
p .001. pressways 18 times in the 2 weeks prior to be-



7
Porcerelli et al.

TABLE 2. Rank Ordering of PQS Process Means in Psychoanalysis for Avoidant Personality Disorder

Item no. PQS items M

Most characteristic

88 Patient brings up significant issues and material 8.13
73 Patient is committed to the work of therapy 8.10
90 Patient's dreams or fantasies are discussed 7.80
50 Therapist draws attention to feelings regarded by the patient as unacceptable (e.g.,
anger, envy, or excitement) 7.77
72 Patient understands the nature of therapy and what is expected 7.73
67 Therapist interprets warded-off or unconscious wishes, feelings, or ideas 7.40
81 Therapist emphasizes patient's feelings in order to help him or her experience them
more deeply 7.27
49 Patient experiences ambivalent or conflicted feelings about the therapist 7.13
33 Patient talks of feelings about being close to or needing someone 7.00
53 Patient is concerned about what the therapist thinks of him 6.90
18 Therapist coveys a sense of nonjudgmental acceptance 6.87
28 Therapist accurately perceives the therapeutic process 6.83
98 Therapy relationship is focus of discussion 6.83
Least characteristic

38 There is a discussion of specific activities or tasks for the patient to attempt outside
of sessions 1.23
77 Therapist is tactless 1.63
27 Therapist gives explicit advice and guidance (vs. defers even when pressed to do so) 1.77
37 Therapist behaves in a teacher-like (didactic) manner 1.90
51 Therapist condescends to or patronizes the patient 2.03
85 Therapist encourages patient to try new ways of behaving with others 2.07
57 Therapist explains rationale behind his or her technique or approach to treatment 2.13
21 Therapist self-discloses 2.13
17 Therapist actively exerts control over the interaction 2.30
89 Therapist acts to strengthen defenses 2.40
4 Patient's treatment goals are discussed 2.60
24 Therapist's own emotional conflicts intrude into the relationship 2.60
Note. PQS Psychotherapy Q-Set. The mean score is the average score for each item across each of the 5 yearly sets
of sessions (15 total sessions).



ginning his evaluation for treatment and reported clinically significant at follow-up (Complexity,
six avoidances of expressway driving at 6 months Affect­tone, Social causality, Aggression, Self-
into treatment. At Year 1, he was no longer esteem, and Identity). Interestingly, the Self-
avoided expressway driving and no longer met esteem ratings were clinically significant at Year
DSM­IV criteria for specific phobia, situational 2 through follow-up.
type, as reported by the therapist and an indepen-
dent rater. Discussion

Object Relations Short-term psychodynamic therapies (9 to 53
sessions) show promise as interventions for
At intake, the majority of Mr. A's object rela- AVPD. Treatment results reported by Barber et
tions ratings were around the midpoint of the al. (1997), Hardy et al. (1995), Svartberg et al.
7-point scales (3.50 ­ 4.50), except for Complex- (2004), and Winston et al. (1994) are likely gen-
ity (3.00) and Self-esteem (2.50), which were eralizable to real-world clinical practice given
clearly in the pathological range. Clinically sig- that in all four studies most of the patients had a
nificant changes were noted for four of eight comorbid Axis I pathology. However, not all
dimensions at Year 4 and seven of eight by Year patients improved with short-term therapy, war-
5, with six of these seven dimensions remaining ranting an evaluation of the effectiveness of a



8
Avoidant Personality Disorder

TABLE 3. Mean Scores of Symptom Severity, Global Functioning, Personality Disorder, and Object Relations by Year of
Treatment

Follow-up
Effect size
Intake Year 1 Year 2 Year 3 Year 4 Year 5 1 Year 2 Years Cohen's d

Symptom severity/functioning
GSI 72.00 76.00 67.00 67.00 63.00* 57.00* 58.00* 54.00* 1.80
GAF 54.00 58.00 64.00 63.50 63.00 78.50* 79.00* 2.50
High functioning 56.20 53.85 56.85 62.30 64.80* 74.75* 75.50* 1.93
Personality disorder
Avoidant PD 60.50 50.75 53.90 57.30 55.50 48.20 37.45* 2.30
Object relations
Complexity 3.00 4.00 4.50 5.00* 6.00* 6.00* 6.00* 3.00
Affect-tone 3.50 4.00 4.50 5.00* 5.00* 5.50* 6.50* 3.00
Investment/relations 4.50 4.00 5.50 5.00 5.50 5.50 5.50 1.00
Investment/morals 4.00 4.50 5.00 5.00 5.50* 6.00* 5.00 1.00
Social causality 4.00 4.00 4.00 4.50 5.00 6.50* 6.00* 2.00
Aggression 4.00 3.50 5.50 5.00 4.50 6.00* 6.50* 2.50
Self-esteem 2.50 3.50 4.50* 4.00* 4.50* 5.00* 6.00* 3.50
Identity 4.00 4.00 5.00 4.00 4.00 6.50* 6.50* 2.50
Note. GSI Global Severity Index (Brief Symptom Inventory); Avoidant PD SWAP-200 Avoidant Personality
Disorder Scale; GAF Global Assessment of Functioning Scale (DSM-IV); high functioning SWAP-200 High
Functioning Scale; object relations Social Cognition and Object Relations Global Scale. Effect sizes were calculated
between intake and at 1-year follow-up.
*
Clinically significant change according to Jacobsen & Truax (1991): RCI 1.96 and return to a functional distribution or out
of a dysfunctional distribution.



more intensive psychodynamic therapy for feelings of guilt associated with them, social
treatment-resistant patients with AVPD with co- and work inhibitions, and feelings of inade-
morbid Axis I disorders. This naturalistic system- quacy. As indicated by the relationships be-
atic case study was conducted to assess the ef- tween actual clinical processes, as rated with
fectiveness of psychoanalysis across multiple the Psychotherapy Q-Set, the treatment adhered
domains of patient functioning, including symp- to psychodynamic principles throughout with
toms, global functioning, personality disorder, some use of cognitive­ behavioral and interper-
object relations, and psychological strengths. sonal principles.
Assessments were conducted through both pa- Psychodynamic processes included contribu-
tient self-report and through ratings by indepen- tions from the patient, the therapist, and interac-
dent raters, thus reducing therapist bias. And al- tions between the two. Patient contributions in-
though process ratings were done by the therapist cluded (in rank order) bringing up significant
and an independent rater, interrater reliability co- issues and material (Item 88), being committed to
efficients were excellent, and final process ratings the work of therapy (Item 73), an understanding
were based only upon ratings derived from the of the nature of therapy and what is expected
independent rater. (Item 72), experiencing ambivalent and con-
Findings from this case study warrant further flicted feelings about the therapist (Item 49), be-
investigation of psychoanalysis for patients with ing able to talk about feeling close to or needing
AVPD and comorbid Axis I conditions, espe- others (Item 33), and being concerned about what
cially those who have not benefited from other the therapist thinks of him (Item 53). Therapist
therapies. Mr. A reported reductions in symptoms contributions to the process included (in rank
during his prior treatments but was symptomatic order) the therapist drawing attention to feelings
shortly thereafter. Personality or character dif- regarded by the patient as unacceptable (Item
ficulties remained throughout his adult life and 50); interpreting warded off or unconscious
remitted only during psychoanalysis. In the wishes, feelings, or ideas (Item 67); emphasizing
present treatment, Mr. A needed an intensive, the patient's feelings in order to help him expe-
long-term treatment to allow the emergence of rience them more deeply (Item 81); conveying a
conflicts over aggression and sexual desires, sense of nonjudgmental acceptance (Item 18);



9
Porcerelli et al.


and accurately perceiving the therapeutic process lytic treatment, which at times evidenced pro-
(Item 28). Contribution of the patient/therapist cesses that could be classified as cognitive and
interaction included (in rank order) dreams and interpersonal therapy techniques, was associated
fantasies are discussed (Item 90) and the therapy with recovery from Axis I and Axis II disorders,
relationship is a focus of discussion (Item 98). significant decreases in symptomatic distress and
These findings support the view of many psycho- relational pathology, and an improvement in gen-
therapy researchers (e.g., Ablon & Jones, 1999; eral functioning and psychological strengths.
Blatt, Quinlan, Pilkonis, & Shea, 1995; Krupnick Psychotherapy outcome research has consis-
et al., 1996) who have reported that an analysis of tently demonstrated that symptoms, functioning,
patient characteristics in combination with thera- and personality pathology respond at different
pist and relationship variables is crucial in under- rates during psychotherapy. Studies assessing the
standing therapeutic outcome. dose­ effect relationship of psychotherapy have
The CBT and IPT processes reported at the end demonstrated that the most rapid changes occur
of Year 3 deserve some explanation. Mr. A had in the areas of self-reported subjective well-being
extremely strong resistances to acknowledging and Axis I symptoms (Howard et al., 1993, 1996;
and working with his conflicts over aggression Kopta et al., 1994; Lueger et al., 2000). Changes
toward his wife. Through the psychoanalytic in interpersonal and personality functioning tend
method, the therapist and patient came to under- to require longer interventions (Kopta et al.,
stand his intense and frightening unconscious 1994; Perry, Banon, & Ianni, 1999) and thus
rage toward his wife (and associated guilt), change at a slower rate than Axis I symptoms--
whose deteriorating physical health threatened to this three-phase sequence of recovery in known
deplete his emotional and financial resources. It as the phase model (Howard et al., 1993, 1996).
was during the third year that much work was Our hypothesis regarding the phase model of
accomplished in this area. During this time, de- change received partial support. Although the
tailed information was gathered about Mr. A's patient no longer met DSM­IV criteria for spe-
current thoughts, fantasies, feelings, and behav- cific phobia, situational type, after Year 1, he did
iors toward his wife. An emphasis on current not evidence clinically significant change in
relationships is a central focus of IPT, and thus it symptom severity until Year 4. However, these
is not surprising that a significant correlation be- clinically significant changes did occur prior to
tween session process at the end of Year 3 and the clinically significant changes in PD, thus support-
IPT ideal prototype emerged. During this same ing our hypothesis. The patient no longer met
period, Mr. A often commented on his efforts to SWAP-200 criteria for AVPD at the end of the
"behave differently" with his wife, and he and the first year of treatment; he did, however, meet
therapist would at times talk about ways of im- criteria for AVPD features at the end of the third
proving his behavior toward her. These interac- and fourth years of treatment and did not evi-
tions could be construed as a type of homework dence clinically significant change in AVPD
and as advice-giving, which are common in CBT SWAP-200 scores until 1-year follow-up. Object
(for a review of the use of homework in psy- relations are considered to be enduring intrapsy-
chodynamic psychotherapy, see Stricker, 2006). chic structures and thus are dimensions of per-
Because of the limitations of effectiveness re- sonality. Inconsistent with the phase model, as-
search, the investigators cannot say that the pects of object relations showed clinically
changes brought about during the course of the significant changes prior to clinically significant
treatment were due to psychodynamic processes symptom changes. These dimensions included
alone. Research by Ablon and Jones (1998) has Self-esteem, Complexity, and Affect­tone. The
demonstrated how therapists using relatively change process for some patients may entail a
small amounts of technique from one brand of greater degree of complexity than the current
therapy can have significant effects on treatment version of the phase model can account for. We
outcome. Thus, without a larger sample of pa- speculate that changes in internal representations
tients undergoing psychoanalysis for AVPD, we of self and other in treatment-resistant patients
cannot at this point speak to the relative effect of with AVPD may relate to trial behaviors and
psychodynamic, cognitive­ behavioral, or inter- relationships in fantasy that prepare a patient for
personal processes on treatment outcome. What more enduring relational and personality
can be said is that in this particular psychoana- changes.



10
Avoidant Personality Disorder


We also speculate that Mr. A's anxiety/distress number of sessions available for process coding
remained high for so long for several reasons. would allow for analyses of patient­therapist in-
Mr. A was genetically predisposed to anxiety, teraction processes as outlined by Jones (2000).
lived with a highly anxious/phobic mother, had We are in agreement with Jones (2000) that sys-
traumatic experiences both in childhood (i.e., tematic case studies are key to understanding the
hospitalization for a life-threatening illness, sep- mysteries of therapeutic action.
aration due to hospitalization in an isolation room
for 2 weeks) and young adulthood (i.e., emer- References
gency landing of a plane), and sexual fantasies
ABLON, J. S., & JONES, E. E. (1998). How expert clini-
and conflicts that were met by an extremely harsh cians' prototypes of an ideal treatment correlate with
and punitive conscience that resulted in distress outcome in psychodynamic and cognitive­ behavioral
(in part as a punishment for his desires). In ad- therapy. Psychotherapy Research, 8(1), 71­ 83.
dition, he was the primary caretaker for his wife, ABLON, J. S., & JONES, E. E. (1999). Psychotherapy pro-
who suffered from a deteriorating neuromuscular cess in the NIMH Collaborative Research Program.
Journal of Consulting and Clinical Psychology, 67(1),
condition. He obsessively worried about her suf- 64 ­75.
fering an exacerbation of her condition that ABLON, J. S., & JONES, E. E. (2002). Validity of con-
would warrant around-the-clock nursing care and trolled clinical trials of psychotherapy: Findings from
deplete his emotional and financial resources. It the NIMH treatment of depression collaborative re-
search program. American Journal of Psychiatry,
took time to help him to become aware of, toler- 159(5), 775­783.
ate, and work through his enormous feelings of American Psychiatric Association. (1994). Diagnostic
fear, guilt, anger, and sexual conflicts. and statistical manual of mental disorders (4th ed.).
With regard to Mr. A's driving phobia, Fon- Washington, DC: Author.
agy, Roth, and Higgett (2005) indicate that there BARBER, J. P., MORSE, J. Q., KRAKAUER, I. D., CHIT-
TAMS, J., & CRITS-CHRISTOPH, K. (1997). Change in
are no effectiveness or randomized control trials obsessive­ compulsive and avoidant personality disor-
of short- or long-term psychoanalytic therapy for ders following time-limited supportive­ expressive ther-
specific phobia in the literature, despite the fact apy. Psychotherapy, 34, 133­143.
that specific phobias are one of the most common BLATT, S. J., & AUERBACH, J. S. (2003). Psychodynamic
measures of therapeutic change. Psychoanalytic In-
anxiety disorders. However, there is data to sug- quiry, 23, 268 ­307.
gest that short-term psychodynamic therapy is BLATT, S. J., QUINLAN, D. M., PILKONIS, P. A., & SHEA,
effective for a variety of other Axis I disorders T. M. (1995). Impact of perfectionism and need for
(Leichsenring, Rabing, & Leibing, 2004), includ- approval on the brief treatment of depression: The
ing panic disorder (e.g., Milrod et al., 2001). National Institute of Mental Health Treatment of De-
pression Collaborative Program revisited. Journal of
Given the fact that Mr. A's phobia remitted after Consulting and Clinical Psychology, 63, 125­132.
1 year of psychoanalysis (combined with behav- COGAN, R., & PORCERELLI, J. H. (2006). The longitudinal
ioral technique) suggests that short-term therapy study of psychoanalysis for personality pathology. Un-
that integrates dynamic and behavioral technique published raw data.
COID, J., YANG, H., TYRER, P., ROBERTS, A., & ULLRICH,
may be useful in the treatment of specific phobia. S. (2006). Prevalence and correlates of personality dis-
It should be noted that Mr. A had a good working order in Great Britain. British Journal of Psychiatry,
knowledge of behavioral treatment for phobias. It 188, 423­ 431.
is beyond the scope of this article to discuss the DEROGATIS, L. R. (1993). BSI: Administration, scoring,
details of the therapist's approach in working and procedures manual. Minneapolis, MN: National
Computer Systems.
with Mr. A's phobic avoidance of driving on FONAGY, P. (2002). An open door review of outcome
expressways. Suffice it to say, however, that the studies in psychoanalysis (2nd ed.). London: Interna-
therapist did explore and interpret the dynamics tional Psychoanalytic Association.
of Mr. A's avoidance and helped him to make use FONAGY, P., ROTH, A., & HIGGITT, A. (2005). Psychody-
namic psychotherapies: Evidence- based practice and
of behavioral techniques while driving. clinical wisdom. Bulletin of the Menninger Clinic, 69,
One of the major drawbacks of the present 1­58.
study was the limited number of sessions during GABBARD, G. O., GUNDERSON, J. G., & FONAGY, P.
each year that were audio recorded. Future sys- (2002). The place of psychoanalytic treatments within
tematic case studies of psychoanalysis should psychiatry. Archives of General Psychiatry, 59, 505­510.
GOLDMAN, H. H., SKODOL, A. E., & LAVE, T. R. (1992).
include assessments at 6 or even 3 months apart Revising Axis V for DSM­IV: A review of measures of
for outcome measures to gain a precise under- social functioning. American Journal of Psychiatry, 149,
standing of change processes. Likewise, a greater 1148 ­1156.




11
Porcerelli et al.

GRANT, B. F., HASIN, D. S., STINSON, F. S., DAWSON, MARIN-AVELLAN, L. E., MCGAULEY, G., CAMPBELL, C.,
D. A., CHOU, S. P., RUAN, W. J., ET AL. (2004). Preva- & FONAGY, P. (2005). Using the SWAP-200 in a
lence, correlates, and disability of personality disorder personality-disordered forensic population: Is it valid,
in the United States: Results from the national epide- reliable and useful? Criminal Behaviour and Mental
miologic survey on alcohol and related conditions. Health, 15, 28 ­ 45.
Journal of Clinical Psychiatry, 65, 948 ­958. MILROD, B., BUSCH, F., LEON, A. C., ARONSON, A.,
HARDY, G. E., BARKHAM, M., SHAPIRO, D. A., STILES, ROIPHE, J., RUDDEN, M., ET AL. (2001). A pilot open
W. B., REES, A., & REYNOLDS, S. (1995). Impact of trial of brief psychodynamic psychotherapy for panic
cluster C personality disorders on outcomes of contrast- disorder. Journal of Psychotherapy Practice and Re-
ing brief psychotherapies for depression. Journal of search, 10, 239 ­245.
Consulting and Clinical Psychology, 63, 997­1004. MONSEN, J. T., ODLAND, T., FAUGLI, A., DAAE, E., &
HILSENROTH, M. J., ACKERMAN, S. J., BLAGYS, M. D., E ILERTSEN , D. E. (1995). Personality disorders:
BAUMANN, B. D., BAITY, M. R., SMITH, R., ET AL. Changes and stability after intensive psychotherapy fo-
(2000). Reliability and validity of DSM­IV Axis V. cusing on affect consciousness. Psychotherapy Re-
American Journal of Psychiatry, 157, 1858 ­1863. search, 5, 33­ 48.
HILSENROTH, M. J., STEIN, M., & PINSKER, J. (2004). PERRY, J. C., BANON, E., & IANNI, F. (1999). Effective-
Social Cognition and Object Relations scale: Global ness of psychotherapy for personality disorders. Amer-
rating method (SCORS-G). Unpublished manuscript. ican Journal of Psychiatry, 156, 1312­1321.
HOGLEND, P. (1993). Personality disorders and long-term PETERS, E. J., HILSENROTH, M. J., EUDELL, E. M.,
outcome after brief psychodynamic therapy. Journal of BLAGYS, M. D., & HANDLER, L. (2006). Reliability
Personality Disorders, 7, 168 ­181. and validity of the Social Cognition and Object Re-
HOWARD, K. I., LUEGER, R. J., MALING, M. S., & MAR- lations Scale in clinical use. Psychotherapy Research,
TINOVICH, Z. (1993). A phase model of psychotherapy 16, 606 ­ 614.
outcome: Causal mediation of change. Journal of Con- PORCERELLI, J. H., COGAN, R., & HIBBARD, S. (2004).
sulting and Clinical Psychology, 61, 678 ­ 685. Personality characteristics of partner violent men: A
HOWARD, K. I., MORAS, K., BRILL, P. L., MARTINOVICH, Q-sort approach. Journal of Personality Disorders, 18,
Z., & LUTZ, W. (1996). Evaluation of psychotherapy: 151­162.
Efficacy, effectiveness, and patient progress. American REICH, J. H. (2003). The effect of Axis II disorders on the
Psychologist, 51, 1059 ­1064. outcome of treatment of anxiety and unipolar depres-
HUPRICH, S. K., & GREENBERG, R. P. (2003). Advances sive disorders: A review. Journal of Personality Disor-
in the assessment of object relations in the 1990s. Clin- ders, 17, 387­ 405.
ical Psychology Review, 23, 665­ 698. SAMUELS, J., EATON, W. W., BIENVENU, J., BROWN, C. H.,
JACOBSON, N. S., & TRUAX, P. (1991). Clinical signifi- COSTA, P. T., & NESTADT, G. (2002). Prevalence and
cance: A statistical approach to defining meaningful correlates of personality disorders in a community sam-
change in psychotherapy research. Journal of Consult- ple. British Journal of Psychiatry, 180, 536 ­542.
ing and Clinical Psychology, 59, 12­19. SHEA, M. T., STOUT, R. L., YEN, S., PAGANO, M. E.,
JONES, E. E. (2000). Therapeutic Action. Northvale, NJ: SKODAL, A. E., MOREY, L. C., ET AL. (2004). Associa-
Jason Aronson. tions in the course of personality disorders and Axis I
KOPTA, S. M., HOWARD, K. I., LOWRY, J. L., & BEUTLER, disorders over time. Journal of Abnormal Psychology,
L. E. (1994). Patterns of symptomatic recovery in time- 113, 499 ­508.
limited psychotherapy. Journal of Consulting and Clin- SHEDLER, J., & WESTEN, D. (1998). Refining the mea-
ical Psychology, 62, 1009 ­1016. surement of Axis II: A Q-sort procedure for assessing
KRUPNICK, J. H., SOTSKY, S. M., ELKIN, I., WATKINS, J., personality pathology. Assessment, 5, 333­353.
& PILKONIS, P. A. (1996). The role of therapeutic STRICKER, G. (2006). Using homework in psychodynamic
alliance in psychotherapy and pharmacotherapy out- psychotherapy. Journal of Psychotherapy Integration,
come: Findings in the Na