Psychodynamic/Interpersonal Group Psychotherapy ... - APA PsycNET

Psychotherapy 2015, Vol. 52, No. 2, 205–217

© 2015 American Psychological Association 0033-3204/15/$12.00 http://dx.doi.org/10.1037/pst0000016

This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

Psychodynamic/Interpersonal Group Psychotherapy for Perfectionism: Evaluating the Effectiveness of a Short-Term Treatment Paul L. Hewitt

Samuel F. Mikail

University of British Columbia

The Southdown Institute, Holland Landing, Ontario, Canada

Gordon L. Flett

Giorgio A. Tasca

York University

The Ottawa Hospital, Ottawa, Ontario, Canada, and University of Ottawa

Carol A. Flynn, Xiaolei Deng, Janet Kaldas, and Chang Chen University of British Columbia This study sought to determine whether clinically significant improvement could be obtained using a psychodynamic/interpersonal group treatment based on a comprehensive conceptualization of perfectionism. A sample of 71 community-recruited perfectionistic individuals participated in the University of British Columbia Perfectionism Treatment Study. Eighteen of these participants were initially nonrandomly assigned to a waitlist control condition. All participants completed measures of perfectionism traits, perfectionistic self-presentation, and automatic perfectionistic thoughts, as well as measures of distress including depression, anxiety, and interpersonal problems at pretreatment, posttreatment, and at a 4-month follow-up. Multilevel modeling demonstrated that perfectionism levels decreased with large effect sizes and that these decreases were associated with reductions in distress measures. Clinically significant decreases were found in all perfectionism components, and posttreatment scores on most variables were significantly lower in the treatment condition versus the waitlist control condition. The findings suggest that psychodynamic/interpersonal group treatment is effective in treating components of perfectionism. Keywords: perfectionism, psychodynamic, group psychotherapy, treatment effectiveness

change. In the current article, we report on the evaluation of treatment for one purported underlying personality style relevant to psychopathology and relationship, achievement, and physical health problems—perfectionism (Flett & Hewitt, 2002; Hewitt, Flett, & Mikail, in press). We assess the effectiveness of a psychodynamic/interpersonal (PI) group treatment in reducing perfectionism and distress associated with perfectionism.

There has been considerable attention directed toward psychotherapy outcome research that focuses on treatment of specific symptoms and disorders. This work has informed treatments of particular Axis I disorders; however, by focusing solely on symptoms, important vulnerability and maintenance factors may not be addressed adequately in treatment. Indeed, Blatt, Auerbach, Zuroff, and Shahar (2006) argued persuasively that focusing on evaluating treatments of personality-related variables that underlie psychopathology is essential to effecting significant and lasting

Conceptualization of Perfectionism Although perfectionism has long been recognized as a relevant clinical variable (Horney, 1950), the past 25 years has witnessed a burgeoning of research on perfectionism and its outcomes. The picture that emerges from this work is that perfectionism is a core personality vulnerability factor contributing to several forms of maladjustment, with it even being suggested that extreme perfectionism deserves more attention in the diagnosis and treatment of personality disorders (see Ayearst, Flett, & Hewitt, 2012). Several conceptualizations of perfectionism have been posited (Frost, Marten, Lahart, & Rosenblate, 1990; Hewitt & Flett, 1991; Shafran, Cooper, & Fairburn, 2002; Slaney, Rice, Mobley, Trippi, J., & Ashby, 2001) including combinations of these conceptualizations (Blankstein & Dunkley, 2002). Over the past two decades, Hewitt and Flett have developed the comprehensive model of perfectionistic behavior (Hewitt et al., in press) that involves three major

Paul L. Hewitt, Department of Psychology, University of British Columbia; Samuel F. Mikail, The Southdown Institute, Holland Landing, Ontario, Canada; Gordon L. Flett, Department of Psychology, York University; Giorgio A. Tasca, The Ottawa Hospital, Ottawa, Ontario, Canada, and Department of Psychology, University of Ottawa; Carol A. Flynn, Xiaolei Deng, Janet Kaldas, and Chang Chen, Department of Psychology, University of British Columbia. This research was supported by a grant from the Social Sciences and Humanities Research Council of Canada (SSHRC; 410-2009-1050) awarded to the first author. We thank the anonymous reviewers for their comments and suggestions. Correspondence concerning this article should be addressed to Paul L. Hewitt, Department of Psychology, University of British Columbia, 2136 West Mall, Vancouver, BC, Canada V6T 1Z4. E-mail: phewitt@psych .ubc.ca 205

HEWITT ET AL.


206

intra- and interpersonal components including perfectionism traits, perfectionistic self-presentational facets, and perfectionistic information processing (see Flett, Hewitt, Blankstein, & Gray, 1998; Hewitt & Flett, 1991; Hewitt et al., 2003; Hewitt & Genest, 1990). Briefly, the perfectionism traits include the requirement of perfection of the self, known as self-oriented perfectionism, the requirement of perfection of others, known as other-oriented perfectionism, and the perception that others require perfection of oneself, known as socially prescribed perfectionism. Perfectionistic selfpresentation, the interpersonal expression of one’s purported perfection, involves three facets of perfectionistic self-promotion, nondisplay of imperfections, and the nondisclosure of imperfections. Lastly, the information processing component reflects the activation of an ideal self-schema and is reflected in automatic perfectionistic thoughts. Research using this conceptualization has shown that these components are independent of one another and has demonstrated a strong and consistent link between various perfectionism components and several kinds of psychopathology and maladjustment (see Antony, Purdon, Huta, & Swinson, 1998; Bastiani, Rao, Weltzin, & Kaye, 1995; Enns & Cox, 1999; Hewitt, Flett, & Ediger, 1996; O’Connor, 2007). As indicated, there is increasing recognition among researchers and clinicians of the need to consider and develop psychotherapies that focus on personality and other vulnerability factors rather than focusing solely on symptoms of specific disorders (see Beutler, 2010; Blatt, Zuroff, Hawley, Auerbach, 2010). For example, Hewitt, Habke, Lee-Baggley, Sherry, & Flett (2008) stated that “. . . the focus of psychotherapy should be on patient characteristics and personality vulnerabilities that bear directly and indirectly on the psychopathology the patient exhibits rather than on the clinical syndrome symptoms per se” (p. 116). This is especially the case for personality variables that have been shown to be associated widely with psychopathology. We argue that perfectionism should be a treatment focus for several reasons. First, perfectionistic behavior can act as an underlying vulnerability factor or potential causal or maintenance factor for a variety of disorders and problems. For example, self-oriented perfectionism acts as a vulnerability to unipolar depression (Enns & Cox, 1999; Hewitt et al., 1996) and anorexia nervosa (Bastiani et al., 1995). Likewise, socially prescribed perfectionism is associated with suicide ideation, risk, and attempts in both adolescent and adult samples (see Flett, Hewitt, & Heisel, 2014; O’Connor, 2007). Moreover, studies have shown that features of perfectionism interfere with therapeutic outcomes (Blatt, Quinlan, Pilkonis, & Shea, 1995), affect help-seeking attitudes and fears of psychotherapy (Hewitt et al., 2015), and impact negatively the therapeutic alliance (Hewitt et al., 2008; Shahar, Blatt, Zuroff, Krupnick, & Sotsky, 2004).

Treatment of Perfectionism To our knowledge, the current study is the first to focus on the treatment of perfectionism by representing perfectionism as a broad, multifaceted construct that includes perfectionistic traits, self-presentational styles, and automatic thoughts. Whereas past studies have provided some indications that perfectionism can be reduced via psychotherapy, these studies have been limited in key respects. Some have focused on the impact of treatments, such as cognitive– behavioral therapy, without an explicit and extensive focus on perfectionism. These studies have found inconsistent

change in perfectionism as a function of targeting symptoms of depression (Blatt et al., 1995; Enns, Cox, & Pidlubny, 2002), social anxiety (Ashbaugh et al., 2007; Lundh & Öst, 2001), OCD (Chik, Whittal, & O’Neill, 2008), or anorexia nervosa (Bastiani et al., 1995). Other studies found that perfectionism traits and attitudes remain elevated or unchanged in response to psychotherapy (Blatt et al., 1995; Chik et al., 2008; Rosser, Issakidis, & Peters, 2003) while others have shown moderate reductions in aspects of perfectionism (Ashbaugh et al., 2007; Enns et al., 2002; Lundh & Öst, 2001) or that perfectionism levels remain in the clinical range (Salbach-Andrae, Bohnekamp, Pfeiffer, Lehmkuhl, & Miller, 2008). Other studies have found change according to some perfectionism measures but improvements were not maintained at follow-up or attributable to the treatment (Riley, Lee, Cooper, Fairburn, & Shafran, 2007). Two studies illustrate that when treatments targeting perfectionism yield reductions, it may still be the case that posttreatment scores on key elements of the perfectionism construct remain elevated (see Egan et al., 2014; Lloyd, Fleming, Schmidt, & Tchanturia, 2014). Both of these studies obtained findings indicating that postintervention means on the Frost et al. (1990) concern over mistakes dimension exceeded 25, which is the cutoff score used in experimental work and diary studies to identify a group clearly characterized by an excessive level of concern over mistakes (see Frost et al., 1997, 1995). In the current project, we evaluated the effectiveness of a short-term intensive PI group treatment of perfectionistic behavior. We chose a PI approach because we believed that it would be most effective in dealing with deeply ingrained and more trait-like or stable aspects of perfectionism (Blatt, 1992). The treatment focused on the multifaceted nature of perfectionistic behavior using psychodynamic and interpersonal conceptualizations and interventions (Hewitt et al., in press), is consistent with other psychodynamic approaches used in the treatment of perfectionism (Greenspon, 2008; Sorotzkin, 1998), and is informed by our Group PI treatment for binge eating (Tasca, Mikail, & Hewitt, 2005). Our approach is based on the general premise that when assessing and treating personality vulnerability factors, it is imperative to both conceptualize and operationalize the personality construct comprehensively by taking into account trait levels and personality process components (for discussions, see Buss & Finn, 1987; Wachtel, 1994). Accordingly, in this instance, perfectionism is considered in terms of perfectionism traits, self-presentational facets, and automatic perfectionistic cognitions (Flett et al., 1998; 1991; Hewitt & Flett, 2004)— components of our comprehensive descriptive model of perfectionism (Hewitt et al., in press). We assessed these various aspects of the perfectionism construct as well as depression, anxiety symptoms, and interpersonal problems to determine if changes in perfectionism were associated with corresponding changes in symptoms. The current study was designed to assess whether the treatment was effective, relative to a control condition, in reducing perfectionism at the conclusion of treatment and at a 4-month follow-up. We evaluated whether treatment yielded associated reductions in distress at both time points. The follow-up assessment allowed us to determine if gains were sustained over time but, more importantly, whether the levels of perfectionism and distress continued to decrease as a function of the psychodynamic nature of the treatment (see Hawley, Ho, Zuroff, & Blatt, 2006). Other analyses were conducted to determine whether the changes in the distress

GROUP PSYCHOTHERAPY FOR PERFECTIONISM

produced by treatment were associated with changes in perfectionism, as should be the case. Finally, comparisons of participants in the treatment versus control conditions were conducted to assess whether changes in key variables were attributable to the therapy.

Method


Participants A total of 71 participants were enrolled in this study, a part of the larger University of British Columbia Perfectionism Treatment Study. Overall, 53 participants were in the initial treatment condition and 18 in the waitlist control condition. People in the control condition waited 11 weeks (i.e., the duration of a group treatment protocol) and then 17 accepted the offer of group psychotherapy. A total of 60 participants completed the treatment and the posttreatment assessment (43 from the treatment condition and 17 from the control condition) and a total of 44 participants completed the 4-month follow-up assessment.1 Demographic information for each of these groupings of participants is presented in the left panel of Table 1. To characterize the clinical nature of the sample, 48 reported previous treatment for depression (42%), anxiety (15%), or eating and psychophysiological disorders and symptoms (10%).

Measures Multidimensional Perfectionism Scale. The Multidimensional Perfectionism Scale (Hewitt & Flett, 1991) is a 45-item scale that assesses the three trait dimensions—namely, selforiented, other-oriented, and socially prescribed perfectionism. Participants make seven-point ratings on the degree of their agreement with statements such as “When I am working on something, I cannot relax until it is perfect” (self-oriented), “I have high expectations for people who are important to me” (other-oriented), and “I feel that people are too demanding of me” (socially prescribed). This measure has good reliability and validity (Hewitt & Flett, 1991, 2004; Hewitt et al., 1991). Perfectionistic Self-Presentation Scale. The Perfectionistic Self-Presentation Scale (PSPS; Hewitt et al., 2003) is a 27-item measure that assesses three facets of perfectionistic self-presentation: perfectionistic self-promotion, nondisplay of imperfection, and nondisclosure of imperfection. Participants rate their agreement, on a seven-point scale, of statements such as “It is very important that I always appear to be ‘on top of things’” (perfectionistic self-promotion), “It would be awful if I made a fool of myself in front of others” (nondisplay of imperfection), and “I should solve my own problems rather than admit them to others” (nondisclosure of imperfection). The subscales have good internal consistency and are correlated but distinct from measures of trait perfectionism. Additional validity is demonstrated by the high correlations between ratings of subjects by peers and by therapists with subjects’ self-ratings (Hewitt et al., 2003). Perfectionism Cognitions Inventory. The Perfectionism Cognitions Inventory (PCI; Flett et al., 1998) is a 25-item measure that assesses automatic thoughts that reflect perfectionistic themes. Participants rate items such as “I should be perfect,” “I have to be the best,” and “Why can’t I be perfect?” on the frequency with which they have experienced these thoughts during the past week.

207

Higher scores represent a greater frequency of perfectionistic cognitions. A growing body of evidence attests to the reliability and validity of the PCI, including its incremental validity when pitted against other perfectionism measures (see Flett & Hewitt, 2015). Beck Depression Inventory. The Beck Depression Inventory (BDI; Beck & Steer, 1987) is a 21-item inventory that assesses depression severity by inquiring about both cognitive and somatic symptoms over the past week. This well-known instrument has demonstrated strong internal consistency, concurrent, and discriminant validity (Beck, Steer, & Garbin, 1988). Beck Anxiety Inventory. The Beck Anxiety Inventory (BAI; Beck, Epstein, Brown, & Steer, 1988) is a 21-item inventory that assesses a variety of anxiety symptoms including subjective, neurophysiological, autonomic, and panic symptoms (Beck & Steer, 1991). Participants rate the degree of severity of each symptom over the past week using a four-point scale. Studies have demonstrated the test–retest reliability (Beck et al., 1988) and discriminant validity (Beck & Steer, 1991) of this scale. Interaction Anxiousness Scale. The Interaction Anxiousness Scale (IAS; Leary, 1983) is a 15-item scale that assesses social anxiety in various situations (e.g., casual get-togethers, talking to authority figures). Participants rate statements on a five-point scale to indicate how true each one is for them. Coefficients alpha reflect good internal consistency, and test– retest reliability is adequate (Leary, 1983). Validity studies suggest that the IAS measures social anxiety specifically (Leary & Kowalski, 1993). Inventory of Interpersonal Problems. The Inventory of Interpersonal Problems (IIP; Horowitz, Rosenberg, Baer, Ureño, & Villaseñor, 1988) is a 127-item inventory that measures difficulties which people report in relating to others. Factor analyses identified six primary areas of difficulty: assertiveness, sociability, intimacy, submissiveness, over responsibility, and being too controlling (Horowitz et al., 1988). This overall score, used in this study, has good internal consistency and good test–retest reliability and is sensitive to therapeutic change (Horowitz et al., 1988).

Therapy Format This PI group psychotherapy program combined knowledge of critical components of interpersonal group psychotherapy (MacKenzie, 1990; Yalom & Leszcz, 2005) and key ingredients in the psychodynamic treatment of perfectionists in individual and group psychotherapy (Hewitt et al., in press; Tasca et al., 2005). Use of this model meant that therapy focused on the interpersonal precursors, interpersonal impact, and underlying relational dynamics of perfectionism rather than focusing directly on reducing perfectionistic behaviors (e.g., negative evaluations, stringent expectations, etc.). That is, the emphasis of interventions was placed on addressing perfectionism-related relational patterns manifest in interactions among group members as well as those described by members within the context of other relationships, including one’s 1 Six participants dropped out during treatment and four dropped out after completing treatment but prior to the posttreatment assessment. Statistical tests found no difference between those who did or who did not complete this study in terms of the perfectionism and distress measures and other features such as gender, marital status, and work status (all ps ⬎ .05).

HEWITT ET AL.

208

Table 1 Demographic Data for Pretreatment, Posttreatment, Follow-Up, Treatment Comparison and Waitlist Control Groups


Change over treatment

Treatment vs. control

Demographic

Pretreatment

Posttreatment

Follow-up

Treated group

Waitlist control

N Age Gender Men Women Marital status Single Married Separated/Divorced Employment status Unemployed Full-time Part-time

71 41.55 (10.40)

60 41.33 (10.89)

44 40.21 (11.26)

43 43.40 (9.79)

18 38.39 (10.84)

24 47

20 40

15 29

13 30

6 12

27 28 16

24 23 13

18 17 9

14 19 10

8 5 5

20 35 16

17 28 15

13 20 11

9 21 13

6 10 2

relationship with self. This approach is consistent with traditional and contemporary models of psychoanalytic and interpersonal therapy (McWilliams, 2004; Sullivan, 1953) and with other psychodynamic treatments of perfectionism (Greenspon, 2008; Sorotzkin, 1998). Specifically, therapists worked to keep group discussion in the “here and now” and group members were encouraged to explore their relationships and experiences within the group. Therapists emphasized expression of affect, interpersonal feedback among members, and interpretations of group processes and transference responses within the group as a means of exploring and challenging self-limiting interpersonal dynamics. Prominent themes included members’ reactions to empathic failures, tolerance of therapists’ limitations, interpersonal feedback (often experienced as exposure of personal imperfections), and addressing the use of perfectionism as a means of creating safety or defending against abandonment, rejection, criticism, intimacy, interpersonal conflict and tension, or a lack of control over one’s relational world. Issues pertaining to termination were discussed throughout and especially in the last several sessions. All participants completed two pregroup preparation sessions aimed at enhancing their participation and the benefits they would derive from treatment (MacKenzie, 1990). The groups ran for 11 consecutive weeks. Both pregroup sessions were held during the first week and treatment sessions were held weekly for 10 weeks. Sessions were 1.5 hr in length. Each group was composed of 7 to 10 members. Each group was assigned a male and a female therapist and a total of two male and three female senior level graduate students in clinical psychology acted as cotherapists. All therapists had at least four years of supervised clinical experience and were trained extensively by the first two authors on the PI treatment. The therapists were provided weekly supervision by the first two authors that involved review of video-tapes, feedback on each therapy sessions, and adherence to the treatment model and protocol. Groups were closed; no new members were assigned once the group had commenced.

Procedure The study received ethical approval from the University of British Columbia Behavioral Research Ethics Board. Participants were recruited by media advertisements and posters at campuses and community centers in Vancouver, Canada, that advertised a

Test of difference t(59) ⫽ 1.77, ns ␹2(1) ⫽ .52, ns ␹2(2) ⫽ 1.47, ns

␹2(2) ⫽ 2.79, ns

group treatment program for adults with perfectionism. A brief description of perfectionism and common outcomes was provided as well as contact information. A total of 175 individuals took part in an initial phone screen to rule out psychosis or other severe psychopathology, nonfluency in English, and to ensure that participants’ difficulties actually involved perfectionism. The flowchart of the study is presented in Figure 1 where it can be seen that a total of 127 people attended an initial clinical interview and extensive psychological testing, including all the perfectionism and distress measures. Test results were used to determine eligibility to participate, and, for those selected, these results were used as the pretreatment measures of perfectionism and distress. Inclusion criteria included scores of a minimum of a half standard deviation above the mean on at least one form of trait or selfpresentational perfectionism. A total of 87 participants met inclusion criteria. Exclusion criteria included those with severe clinical issues (e.g., currently suicidal or psychotic), inability to disclose in the initial interview, or those who never had a close relationship. Fifteen individuals were excluded and referred elsewhere, leaving a total 72 participants. One participant dropped out prior to group assignment due to time constraints. Participants who remained were assigned to one of eight treatment groups. Consideration was given to balancing sex, age, disclosure level, type of elevated perfectionism, and ability to attend all sessions. These criteria were used to limit the number of dropouts and to facilitate group participation and cohesion (Gans & Counselman, 2010; Yalom & Leszcz, 2005). As mentioned, 18 of 71 participants served as waitlist controls comprising individuals whose characteristics (i.e., sex, age, elevated perfectionism, disclosure level) were already represented in the treatment groups or those who were unable to attend all sessions in the initial phase of the study. When two individuals with identical characteristics were considered for treatment versus control group, the decision was determined randomly. As indicated in Figure 1, following the wait list control period of 11 weeks after pretesting, each waitlist control group member completed a post control period assessment of the perfectionism and distress measures. Each of the control participants was then assigned to a treatment group. One par-


209

Assessed for eligibility/pre-treatment (n = 127) Excluded (n = 56) ♦ Not meeting inclusion criteria (n = 40) ♦ Declined to participate (n = 1) ♦ Met exclusion criteria (n = 15 )


Assigned nonrandomly (n = 71)

Allocated to intervention (n = 53) ♦ Received allocated intervention (n = 47) ♦ Did not receive intervention (n = 6; 1 dropped out due to illness, 2 work commitments, 2 relational focus not enough information, 1 did not start)

Post-treatment assessment (n = 47) Received assessment (n = 43) Did not receive assessment (n = 4; no show for assessment appointment)

Allocated to wait list control (n = 18) ♦ Received allocated intervention (n = 18)

Post wait list control period assessment after 11 weeks (n = 18 ) Received assessment (n = 18)

Analysed (n = 61) ♦ Excluded from analysis (n = 0)

Allocated to intervention (n = 18) Received intervention (n = 17) Did not receive intervention (n = 1; dropped out prior to treatment due to work concerns) Post treatment assessment (n = 17)

Follow up assessment after 4 months (n = 44) Lost to follow-up (n = 16)

Analysed -pre-treatment (n = 71), post-treatment (n = 60), -follow-up (n = 44) Excluded from analysis (n = 0)

Figure 1. Flow chart of study.

ticipant who served as a waitlist control, after completing the post waitlist assessment, chose not to participate in treatment due to work commitments. At the conclusion of each treatment group, each participant completed the posttreatment assessment. At four months posttreatment, each participant was mailed a questionnaire package and asked to complete a follow-up assessment including the perfectionism and distress measures.

Design and Data Analysis Plan We used a one-group longitudinal correlational design to assess change in perfectionism and distress over the pretreatment, posttreatment, and follow-up period, and the relationship between change in perfectionism and change in distress. To assess effectiveness of the treatment condition compared with a wait list control condition, we used a quasi-experimental nonrandomized

two-group design with repeated measurements at pre- and posttreatment. Change in perfectionism and distress. First, we assessed the treatment effects on perfectionism over the pretreatment, posttreatment, and follow-up periods. Because these data were nested within groups, we assessed if the data violated the assumption of independence by calculating the intraclass correlation coefficient (␳) for each dependent variable (Tasca, Illing, Joyce, & Ogrodniczuk, 2009). That is, we used a three-level multilevel model (MLM) in which repeated measurements (pretreatment, posttreatment, follow-up) at Level 1 were nested within individuals at Level 2, who in turn were nested within groups at Level 3. When ␳ ⬍ .05, the effect of dependency in the data can be ignored (Kenny, Kashy, & Bolger, 1998), and the data can be evaluated at the individual level in a two-level MLM. As indicated below, most ␳s were less than .05 for the variables. For those variables where ␳ ⬎ .05, we

HEWITT ET AL.


210

adjusted the Type I error rate (Kenny et al., 1998). As a result, all analyses were at the individual level using a two-level MLM, with adjustments to Type I error when necessary.2 All Level 2 predictors were centered and residuals allowed to vary. The time parameter was modeled as a log function to indicate precipitous change from pre to post treatment and a less rapid rate of change from post to follow-up.3 Effect sizes are reported as pseudo-R2 (Raudenbush & Bryk, 2002). The HLM program version 6.08 (Raudenbush, Bryk, Cheong, Congdon, & duToit, 2004) was used with the full maximum likelihood method of estimation. Change in perfectionism predicting change in distress. To assess if changes in perfectionism were associated with changes in distress, we ran a two-level MLM for the perfectionism scales and saved the OLS slope estimates of each perfectionism scale for each individual. These slopes estimated the rate of change on the perfectionism scales and were entered in the two-level MLM as predictors of change in each of the distress outcomes. Type I error rate was kept at .05 for each predictor parameter, as the maximum likelihood method in MLM minimizes estimation errors (Nezlek, 2001). For all MLMs, the addition of predictors (time or perfectionism) was assessed for model fit with the difference in deviance statistics of nested models evaluated on a chi-square distribution (Raudenbush & Bryk, 2002). Treatment versus wait list control. To test for differences at pretreatment between the conditions (treatment vs. control), we used chi-square tests and independent samples t test. To assess the effect of condition on pre- to posttreatment change, we used a two-level MLM in which individual responses at Level 1 were nested within group at Level 2. Treatment conditions and pretreatment scores were entered as predictors for posttreatment scores. We did not adjust error rate for parameter estimate, as MLM models perform partial pooling in calculating point estimates (Gelman, Hill, & Yajima, 2012). We also analyzed the data using ANCOVA (with pretreatment scores as covariates) to estimate effect size and we compared model fitting indices using Maximum Likelihood Estimation. Rather than imputing missing postdata by last observation carried forward, which introduces a number of biases (Molnar, Hutton, & Fergusson, 2008; Saha & Joines, 2009), we included only those who had complete data at posttreatment in these analyses.

Results Change in Perfectionism The means and standard deviations for the perfectionism and distress measures at pretreatment, posttreatment, and follow-up are presented in Table 2. The levels of perfectionism had means greater than typical means for clinical samples (Flett, Hewitt, Whelan, & Martin, 2007; Hewitt & Flett, 2004; Hewitt et al., 2003). The perfectionism and distress measures data were normally distributed with no univariate or multivariate outliers at any time point. Moreover, the variance components from the three level MLM indicated that the ␳ for most variables were below .05, suggesting little or no dependence in the data; however dependence was evident in some scales requiring adjustment of Type I error rate for those scales. For the Other-Oriented Perfectionism subscale, ␳ ⫽ .08, Type I error was adjusted to .035; for the Perfectionistic Self-Promotion subscale, ␳ ⫽ .23, Type I error was

adjusted to .015; for the Perfectionistic Nondisclosure scale, ␳ ⫽ .11, Type I error was adjusted to .025, and for the BAI, ␳ ⫽ .10, Type I error was adjusted to .025. In addition, according to Little (1988), the chi-square test was used to determine whether any missing data were missing at random. On the subscale level, data were missing completely at random, ␹2 ⫽ 16.90, df ⫽ 21, p ⫽ .72. We tested whether the treatment resulted in significant change over time including follow-up. Table 3 shows the fixed effects for each perfectionism scale and distress measure from the two-level MLM. All variables showed a significant change from pre- to postto four month postintervention indicating a reduction in scores. Because of the logtime modeling for change, the results indicate a precipitous reduction in scores from pre- to posttreatment, and a continued reduction but at a less accelerated rate from posttreatment to four months posttreatment. All effect sizes for the slopes were large, except for socially prescribed perfectionism which was medium. For each variable, the difference in deviance statistics indicated that the MLM that modeled logtime was a better fitting model than the base model (all ps ⬍ .001). An indication of the clinical significance of changes in perfectionism over the treatment period was obtained by calculating the Reliable Change Index (Jacobson & Truax, 1991) for each perfectionism measure for the participants who completed treatment. Of the 60 individuals who completed treatment, 55 individuals showed clinically significant improvement (i.e., RCI ⬎1.96) on at least one perfectionism measure, and 54 individuals showed clinically significant improvement on two or more perfectionism measures. However, seven participants showed deterioration (i.e., RCI ⬍ ⫺1.96) on at least one perfectionism dimension.

Change in Perfectionism Predicting Change in Distress We also examined whether changes in the perfectionism components, indicated by their OLS slope parameter values, were associated with change or slopes of each of the distress variables. Table 4 reports the fixed effects parameters for these relationships and Table 5 shows the random effects and deviance statistics for relevant models. The OLS slopes of perfectionism scales, as a group, accounted for 31% of the variance in the BDI slopes, and the model with the perfectionism OLS slopes as predictors improved the fit to the data, ␹2(13) ⫽ 23.10, p ⬍ .001. Specifically, change in the Self-Oriented subscale was associated with change 2 Some researchers (Baldwin, Murray, & Shadish, 2005) have suggested that group-level dependency may be an issue and analyses should include group as a nested third level in MLMs. We chose to analyse these data by not including group at the third level for several reasons. Most notably, we tested whether therapeutic relationship variables believed essential to effective group treatment differed across groups for all 12 sessions. Using the Group Climate Questionnaire (MacKenzie, 1983) administered after each session for each group member, mixed-design ANOVAs yielded no significant differences among the groups on levels of engagement, conflict, or avoidance, nor was there a significant interaction between group and session. 3 We also fit both a no-change after treatment model and a steady change model in addition to a log-linear change model. We used maximum likelihood estimation to compare model fittings. The steady change model was identical to the log-linear model and for SOP and IPP both models achieved significantly better model fitting than the no-change after treatment model ␹2(2) ⫽ 28.97, p ⬍ .01; ␹2(2) ⫽ 47.50, p ⬍ .01. We chose to report the log-linear model due to both better model fittings and our theoretical interest in change between posttreatment to follow-up.


211

Table 2 Means and Standard Deviations of the Pretreatment, Posttreatment, and Follow-Up Perfectionism Trait, Perfectionistic Self-Presentation, Perfectionism Cognitions, and Distress Measures Pretreatment (n ⫽ 71)


Variable Perfectionism traits Self-oriented Other-oriented Socially prescribed Perfectionism self-presentation Self-promotion Nondisplay Nondisclosure Perfectionism cognitions BDI BAI IAS IIP

Posttreatment (n ⫽ 60)

Follow-up (n ⫽ 44)

M

SD

M

SD

M

SD

87.99 72.50 69.03

9.15 13.92 16.38

71.24 63.26 56.44

17.69 17.60 17.79

62.45 60.02 62.27

10.94 4.92 10.20

52.49 55.37 30.31 51.48 17.39 15.26 47.75 1.71

9.50 9.32 8.95 18.25 8.50 10.25 12.07 0.69

44.36 46.45 24.67 37.15 10.83 10.28 43.52 1.38

11.98 11.94 9.56 19.37 8.22 7.40 12.09 0.63

42.64 46.68 24.04 37.41 9.98 9.00 41.36 1.48

12.50 11.76 8.35 23.80 8.50 7.87 11.57 0.98

Note. BDI ⫽ Beck Depression Inventory; BAI ⫽ Beck Anxiety Inventory; IAS ⫽ Interpersonal Anxiety Scale; IIP ⫽ Inventory of Interpersonal Problems.

in the BDI, p ⫽ .05, and uniquely accounted for 6% of the variance in BDI slopes. Further, change in the Perfectionistic Nondisplay subscale was associated with change in the BDI, p ⫽ .01, and uniquely accounted for 8% of variance BDI slopes. The perfectionism OLS slopes as a group accounted for 41% of variance in change in the BAI scale, and the perfectionism slopes as predictors significantly improved the model’s fit to the data, ␹2(13) ⫽ 24.92, p ⬍ .001. Specifically, change in the PCI was Table 3 Fixed and Random Effects for Slopes of Pre- to Post- to 4 Months Postintervention From the Two-Level Multilevel Models Controlling for the Individual Prescore Fixed effects

Random effects

Variable

Change/Slope (␤10)

t

p

␴2b

␴t2

R2

Self Other Social PCI Promote Nondisplay Nondisclosure BDI BAIa IAS IIP

⫺52.42 ⫺27.30 ⫺20.39 ⫺32.53 ⫺22.11 ⫺21.12 ⫺13.70 ⫺21.22 ⫺17.87 ⫺10.62 ⫺0.49

15.06 18.37 5.81 5.07 6.60 7.09 6.39 7.90 6.64 5.15 2.40

⬍.001 ⬍.001 ⬍.001 ⬍.001 ⬍.001 ⬍.001 ⬍.001 ⬍.001 ⬍.001 ⬍.001 .019

235.04 176.89 163.15 206.83 83.59 64.34 37.09 62.18 62.50 27.17 0.23

75.13 113.36 141.26 128.71 48.09 30.62 19.74 25.35 34.63 15.59 0.17

.69 .37 .14 .37 .43 .53 .47 .59 .44 .43 .27

Note. df ⫽ 69, N ⫽ 71. Self ⫽ self-oriented perfectionism; other ⫽ other-oriented perfectionism; Social ⫽ socially prescribed perfectionism; PCI ⫽ Perfectionism Cognitions Inventory; Promote ⫽ perfectionistic self-promotion; Nondisplay ⫽ nondisplay of imperfections; Nondisclosure ⫽ nondisclosure of imperfections; BDI ⫽ Beck Depression Inventory; BAI ⫽ Beck Anxiety Inventory; IAS ⫽ Interpersonal Anxiety Scale; IIP ⫽ Inventory of Interpersonal Problems; ␴2b ⫽ the within-person variance component from the base model; ␴t2 ⫽ the within-person variance component from the model with time. a Type I error for the BAI scale was set at .025.

associated with change in the BAI, p ⫽ .004, and uniquely accounted for 13% of the variance in BAI slopes. Further, change in the Perfectionistic Self-Promotion subscale was associated with change in the BAI, p ⫽ .02, and uniquely accounted for 12% of the variance in BAI slopes. The perfectionism scale OLS slopes as a group accounted for 57.5% of the variance in the change of the IAS scale, and the Perfectionism scale slopes as predictors improved the model’s fit to the data, ␹2(13) ⫽ 43.09, p ⬍ .001. Specifically, the change in the Perfectionistic Nondisplay scale was associated with change in the IAS, p ⬍ .001, and uniquely accounted for 41% of variance in IAS slopes. Lastly, the perfectionism scale OLS slopes as a group accounted for 48% of the variance in change in total IIP scores, and adding the perfectionism scales as predictors significantly improved the model’s fit to the data, ␹2(13) ⫽ 33.09, p ⬍ .001. Change in socially prescribed perfectionism scale was associated with change in the IIP total scores, p ⫽ .05, and accounted for 5% of the variance in IIP score slopes. Also, change in the PCI scale was associated with change in the IIP score slopes, p ⫽ .02, and accounted for 6% of the variance. Finally, the change in the Perfectionistic Nondisplay scale was associated with change in the IIP scores, p ⫽ .005, and accounted for 7% of the variance in IIP score slopes.

Treatment Versus Control Comparisons for Perfectionism and Distress Measures The treatment and control groups did not differ on any demographic variables including age, gender, marital, or employment status (Table 1, right hand panel). The means, standard deviations, and tests of differences on the outcomes between the treatment and control conditions pretreatment and posttreatment are presented in Table 6. The two conditions did not differ on the pretreatment measures of perfectionism or distress.

HEWITT ET AL.

212

Table 4 Fixed Effects Parameters for Change in Perfectionism Scales Predicting Change in the Outcomes BAIa


BDI

IAS

IIP

Variable

␤1j

t

p

␤1j

t

p

␤1j

t

p

␤1j

t

p

Self Other Social PCI Promote Nondisplay Nondisclosure

0.23 ⫺0.17 0.00 0.07 ⫺0.04 0.30 ⫺0.04

2.08 1.89 0.02 0.97 0.29 2.59 .24

.04 .06 .99 .34 .78 .01 .81

⫺0.13 ⫺0.03 0.08 0.19 0.27 0.16 ⫺0.09

1.59 .57 .93 2.99 2.41 1.59 .93

.12 .57 .36 .004 .02 .12 .36

⫺0.12 0.02 0.04 0.05 0.15 0.34 ⫺.04

1.89 .42 .06 .96 1.92 4.03 .41

.06 .68 .48 .34 .06 ⬍.001 .68

0.00 ⫺0.01 ⫺0.01 0.01 0.02 0.03 0.01

.23 1.90 1.97 2.49 1.75 2.93 .63

.82 .06 .05 .02 .08 .005 .53

Note. df ⫽ 69, N ⫽ 71. Self ⫽ self-oriented perfectionism; Other ⫽ other-oriented perfectionism; Social ⫽ socially prescribed perfectionism; PCI ⫽ Perfectionism Cognitions Inventory; Promote ⫽ perfectionistic self-promotion; Nondisplay ⫽ nondisplay of imperfections; Nondisclosure ⫽ nondisclosure of imperfections; BDI ⫽ Beck Depression Inventory; BAI ⫽ Beck Anxiety Inventory; IAS ⫽ Interpersonal Anxiety Scale; IIP ⫽ Inventory of Interpersonal Problems. ␤1j refers to the slope of the relationship between change in each perfectionism scale and change in the outcome. a Type I error for the BAI scale was set at .025.

A series of MLMs was conducted to determine whether the treatment condition differed from the control condition in terms of perfectionism at posttreatment/control period while controlling for prescores and accounting for group dependency (Table 6, last column). The treatment effect was significant for self-oriented perfectionism, socially prescribed perfectionism, perfectionistic cognitions, and all three PSPS facets. Only other-oriented perfectionism did not change differentially between the treatment and the control conditions. A series of ANCOVA analyses revealed very similar patterns to MLM models4 (see Table 6, column labeled “eta2” and “F(1, 58)” in right hand panel). These findings suggest that, relative to the control condition, the treatment was effective in reducing levels of these components of perfectionism and these represent large effect sizes (Cohen, 1988). Another series of MLMs was conducted to assess whether posttreatment distress scores, controlling for pretreatment distress and accounting for group dependency, were lower in the treatment group than in the control group (Table 6, lower portion). The treatment effect was significant for depression and interpersonal problems. The two measures of anxiety at posttreatment did not demonstrate significant differences between the conditions. A series of ANCOVA analyses revealed very similar patterns to MLM models.4

Discussion This article describes the evaluation of a PI group treatment for perfectionism. In a sample of 71 patients, we found that various components of perfectionism and distress were reduced at significant levels following treatment and at a 4-month follow-up. Moreover, the reductions in components of perfectionism were associated with improvement in depression and interpersonal problems but not anxiety. Finally, results suggest that reduction in components of perfectionism were greater in those receiving treatment compared with a waitlist control condition. The current work provides evidence that a PI group treatment that focuses specifically on underlying relational mechanisms of perfectionistic behavior has a significant effect on reducing perfectionistic behavior and on the attendant distress. These findings were evident both at posttreatment and at follow-up, thus suggest-

ing that this form of treatment may be effective in reducing perfectionism as well as symptoms of psychological distress. Importantly, the findings indicate that the PI group treatment was associated with changes in different perfectionism components including the state-like automatic perfectionistic cognitions but also the more deeply ingrained perfectionism traits and selfpresentational styles. This is consistent with similar psychodynamic treatments that focus on underlying mechanisms (Blatt et al., 2006). Collectively, the findings of the current study suggest that a treatment that focuses specifically on reducing perfectionistic behavior by addressing the relational underpinnings of perfectionism produces significant changes in perfectionism and that these changes may result in changes in depression and interpersonal problems. Large effect sizes were typically obtained across the various measures of perfectionism, with the exceptions of socially prescribed perfectionism, which showed a medium effect size. Moreover, the vast majority of patients who completed treatment showed clinically significant improvements, underscoring the clinical utility of this PI treatment. Overall, 54 of the 60 people who completed treatment showed clinically significant improvement on at least two perfectionism measures. Although it has been suggested that the treatment of perfectionistic behavior requires a long-term intensive intervention, and this may certainly be the case for specific individuals, the current work shows that a short-term intensive group therapy approach (12 sessions including 10 actual treatment sessions) can produce clinically significant changes in various components of perfectionistic behavior. Posttreatment comparisons revealed that the treatment participants versus controls had significantly lower levels of selforiented perfectionism, socially prescribed perfectionism, and perfectionistic automatic thoughts, as well as significantly lower scores on all three PSPS facets. However, the mean level of other-oriented perfectionism at the posttreatment and follow-up was still substantially elevated relative to clinical norms, and this 4 We constructed chi-square tests by comparing the likelihood deviance scores between models that accounted for group membership (MLM) and models that did not (ANCOVA). No statistically significant differences were found between models for any outcome variable. In other words, modeling group membership did not significantly improve model fitting.


213

Table 5 Random Effects for Slopes and Deviance Statistics of Models With All Predictors, and for Significant Predictors Removed From Full Model


BDI

BAI

IAS

IIP

Model and predictors

D

␶1

D

␶1

D

␶1

D

␶1

Unconditional model Full model Significant predictors Self slope Social slope PCI slope Promote slope Nondisplay slope Nondisclosure slope

1,112.64 1,089.54

451.66 311.94

1,123.60 1,098.68

354.45 210.66

961.56 918.46

206.03 87.47

145.74 112.65

1.85 0.97

1,094.44 1,092.97 1,089.54 1,091.16 1,093.72 1,089.60

333.39 333.94 311.96 315.41 337.23 312.26

1,100.54 1,099.07 1,099.38 1,106.46 1,099.95 1,099.03

218.96 211.87 214.59 241.79 218.83 209.43

921.32 918.72 918.91 919.33 931.65 918.56

93.27 87.73 89.08 89.04 113.27 87.57

112.83 119.03 115.88 116.00 119.54 112.83

0.97 1.09 1.04 1.02 1.13 0.98

Note. D ⫽ deviance statistic, ␶1 ⫽ variance component; Self ⫽ self-oriented; Social ⫽ socially prescribed; PCI ⫽ Perfectionism Cognitions Inventory; Promote ⫽ perfectionistic self-promotion; Nondisplay ⫽ nondisplay of imperfections; Nondisclosue ⫽ nondisclosure of imperfections; BDI ⫽ Beck Depression Inventory; BAI ⫽ Beck Anxiety Inventory; IAS ⫽ Interpersonal Anxiety Scale; IIP ⫽ Inventory of Interpersonal Problems. Degrees of freedom for D were as follows: unconditional model ⫽ 69 (see Model 2 in Appendix), full model ⫽ 22 (Model 3 in Appendix), and for significant predictors removed ⫽ 20.

is worth exploring further given that this orientation contributes to a host of interpersonal problems and relationship difficulties. Extreme other-oriented perfectionists may seem to improve but still be at risk by engaging in a style that generates stress due to their inability to accept the limitations of other people; this can extend to be highly judgmental of treatment providers in ways that can undermine the development of a positive therapeutic alliance. If viewed from this perspective, it is not surprising that there is evidence indicating that other-oriented perfectionism in the patient can predict early termination of treatment (McCown & Carlson, 2004). It may be that other-oriented perfectionists respond differently to treatment and may require longer term or more focused group treatment or individual psychotherapy. The largest treatment versus control group difference was found in the reductions in nondisclosure of imperfections. This is worth noting given that an unwillingness to disclose emotion-laden issues and the tendency to conceal problems involving the self is a

factor that tends to mitigate treatment success (Farber, 2003; Kahn, Achter, & Shambaugh, 2001). It may be particularly useful to know that participation in this type of group treatment may result in greater verbal disclosure among those high in perfectionism. Previous research showed that personal disclosures can be quite aversive to individuals with excessive levels of perfectionism and tend to elevate physiological arousal among patients in a clinical setting who are characterized by a high need to avoid verbal disclosures of imperfections to others (see Hewitt et al., 2008). Our results may suggest that perfectionism is best treated in the context of groups that are homogenous in composition (i.e., all members struggle with perfectionism in some form), as this may lessen feelings of shame that pervade the nondisplay and nondisclosures of imperfections. The inclusion of the follow-up allowed us to evaluate whether the reductions in perfectionistic behavior and symptoms of distress were maintained and continued to decrease at follow-up. This

Table 6 Means, SDs, F Tests, and Treatment Effects of Perfectionism and Distress for Treatment Versus Control Groups Pretreatment Variable Self Other Social Promote Nondisplay Nondisclosure PCI BDI BAI IAS IIP

Treatment group 87.07 (9.25) 71.99 (14.30) 65.05 (17.36) 51.20 (10.05) 53.53 (10.55) 29.00 (9.39) 50.07 (16.40) 18.00 (8.47) 15.44 (11.01) 45.81 (11.54) 1.64 (.63)

Control group 88.00 (10.17) 75.50 (11.09) 75.17 (14.70) 55.50 (9.92) 58.94 (6.08) 31.83 (8.93) 50.00 (22.70) 15.28 (8.80) 16.81 (9.83) 49.44 (11.37) 1.65 (.52)

Posttreatment Treatment group 70.68 (17.06) 61.73 (17.40) 52.75 (14.78) 43.13 (12.21) 44.79 (12.63) 21.67 (7.73) 36.88 (17.56) 9.23 (5.66) 8.81 (6.01) 41.28 (11.13) 1.30 (.63)

Control group 85.33 (11.36) 70.75 (14.01) 71.75 (15.02) 55.00 (8.77) 54.58 (7.79) 31.39 (10.20) 51.97 (20.93) 13.94 (9.42) 10.17 (6.07) 47.53 (12.96) 1.66 (.62)

Group 2

Differences

Eta

F(1, 58)

.19 .05 .20 .16 .08 .24 .18 .14 .01 .05 .13

ⴱⴱⴱ

13.73 2.99 14.64ⴱⴱⴱ 11.08ⴱⴱ 4.79 18.11ⴱⴱⴱ 12.38ⴱⴱⴱ 9.44ⴱⴱⴱ 0.49 3.03 8.68ⴱⴱⴱ

␤ 13.98ⴱⴱⴱ 5.95 14.83ⴱⴱⴱ 9.65ⴱⴱⴱ 5.63ⴱ 8.17ⴱⴱⴱ 15.78ⴱⴱⴱ 5.16ⴱⴱ .72 2.55 .35ⴱⴱ

Note. Self ⫽ self-oriented; Other ⫽ other-oriented; Social ⫽ socially prescribed; Promote ⫽ perfectionistic self-promotion; PCI ⫽ Perfectionism Cognitions Inventory; BDI Beck ⫽ Depression Inventory; BAI ⫽ Beck Anxiety Inventory; IAS ⫽ Interpersonal Anxiety Scale; IIP ⫽ Inventory of Interpersonal Problems. Alpha corrected to ⴱⴱⴱ p ⬍ .005 for F test, and uncorrected for HLM, ⴱ p ⬍ .05, ⴱⴱ p ⬍ .01, ⴱⴱⴱ p ⬍ .001. Treatment group, n ⫽ 43, and control group, n ⫽ 18. Group differences in change of perfectionism and distress were tested using multilevel modeling and ANCOVA while statistically controlling for pretreatment scores. According to Cohen (1988) partial Eta2 values greater than .14 reflect a large effect size and those between .06 and .13 reflect a medium effect size.


214

HEWITT ET AL.

result was expected because PI treatment focuses on underlying processes that produce or maintain perfectionism. Blatt et al. (2010) suggested sustained therapeutic changes should be evident when treating personality vulnerabilities and we believe that changes in perfectionism should have an impact not just on concurrent symptomatology but also should show continued reductions in distress levels. This is due to fundamental changes in the relationship between the patient and others and the patient and him/herself. Hawley et al. (2006) assessed perfectionistic attitude changes and attendant changes in depression and found that continued changes in depression could be predicted with changes in perfectionism. More generally, Shedler (2010) reviewed psychoanalytic treatment studies and also proposed that such treatments can continue to have effects beyond the treatment period as a result of changes in underlying dynamics. Indeed, our results indicated that the effects of the treatment continued to reduce perfectionism levels several months after the treatment terminated. Our analyses showed that levels of depression and interpersonal problems were lowered significantly in the treatment condition, but anxiety levels of those treated did not show significant changes when compared with levels of anxiety among participants in the control condition. This finding is consistent with Riley et al. (2007) who also found that anxiety scores did not change as a result of treatment that was directed toward perfectionistic behavior. Similarly, Arpin-Cribbie, Irvine, and Ritvo (2012) found that levels of anxiety were still elevated despite reductions in perfectionism among participants receiving a Web-based intervention. Clearly, further work is needed to gain a better understanding of the persistence of anxiety among vulnerable perfectionists. The analyses showing that changes in specific components of perfectionism were associated with reductions in specific symptoms are particularly informative in understanding the link between perfectionism and specific symptoms. For example, changes in self-oriented perfectionism and the nondisplay of imperfections were associated with decreases in depression symptoms. It has been shown that self-oriented perfectionism can act as a vulnerability factor in depression (Enns & Cox, 2005; Hewitt & Flett, 1993; Hewitt et al., 1996) and that the nondisplay of imperfections is associated with depression symptoms (Hewitt et al., 2003). The current findings add to our understanding of the importance of these two perfectionism components in the experience of depression such that reducing levels of these components of perfectionism may produce a reduction in depression. Likewise, the automatic perfectionism cognitions and nondisplay reductions were associated with decreases in anxiety and reductions of socially prescribed perfectionism and the nondisplay of imperfections were associated with decreases in interpersonal problems. These findings further underscore the specific effects of reductions in specific components of perfectionistic behavior. Although the current study yielded some important findings, a few limitations should be noted. First, patients were not randomly assigned into groups but were assigned based on the type of perfectionism, gender, and age. Although this procedure has ecological validity and reflects the way treatment groups are selected in clinical contexts, it precludes the randomization of patients into conditions. Despite this, there was no evidence that control participants differed from treatment participants at pretreatment and only differed on outcomes at posttreatment. Nevertheless, it is important to further this work with randomized control trials.

Also this study used a correlation method for some findings. Changes in perfectionism scores predicted change in measures of symptoms; however, the direction of the effect cannot be assumed with certainty. Although it is reasonable to propose that changes in perfectionism were the basis for changes in symptoms, it could be argued that changes in symptoms may result in changes in perfectionism, or that other therapeutic factors explain the changes. Future research should examine this and the PI approach in individual treatment as well as testing different treatment lengths. Moreover, longer follow-up assessments could determine if the treatment is lasting and whether by reducing the putative vulnerability factor of perfectionism there is a commensurate reduction in episodes of depression and anxiety and other forms of pathology. These are important to move our understanding of the treatment of perfectionism forward.

Conclusion Perfectionism is often noted as a powerful personality variable that has an impact on psychopathology and treatment of psychological problems. The present study is a potentially significant development for the literature on the treatment of perfectionism in that it demonstrates that a short-term intensive PI group therapy focused on the underpinnings of perfectionism may have beneficial effects for the patients. Perfectionism levels were reduced significantly and were associated with decreased levels of distress. Overall, the findings of this study are encouraging in demonstrating large and clinically significant effects in the reduction of perfectionistic behavior.

References Antony, M. M., Purdon, C. L., Huta, V., & Swinson, R. P. (1998). Dimensions of perfectionism across the anxiety disorders. Behaviour Research and Therapy, 36, 1143–1154. Arpin-Cribbie, C., Irvine, J., & Ritvo, P. (2012). Web-based cognitivebehavioral therapy for perfectionism: A randomized controlled trial. Psychotherapy Research, 22, 194 –207. Ashbaugh, A., Antony, M. M., Liss, A., Summerfeldt, L. J., McCabe, R. E., & Swinson, R. P. (2007). Changes in perfectionism following cognitivebehavioral treatment for social phobia. Depression and Anxiety, 24, 169 –177. Ayearst, L. E., Flett, G. L., & Hewitt, P. L. (2012). Where is multidimensional perfectionism in DSM-5? A question posed to the DSM-5 personality and personality disorders work group. Personality Disorders: Theory, Research, and Treatment, 3, 458 – 469. Baldwin, S. A., Murray, D. M., & Shadish, W. R. (2005). Empirically supported treatments or type I errors? Problems with the analysis of data from group-administered treatments. Journal of Consulting and Clinical Psychology, 73, 924 –935. Bastiani, A. M., Rao, R., Weltzin, T., & Kaye, W. H. (1995). Perfectionism in anorexia nervosa. International Journal of Eating Disorders, 17, 147–152. Beck, A. T., Epstein, N., Brown, G., & Steer, R. A. (1988). An inventory for measuring clinical anxiety: Psychometric properties. Journal of Consulting and Clinical Psychology, 56, 893– 897. Beck, A. T., & Steer, R. A. (1987). Beck depression inventory - Manual. Toronto: The Psychological Corporation, Harcourt Brace Jovanovich, Inc. Beck, A. T., & Steer, R. (1991). Relationships between the Beck Anxiety Inventory and the Ham-ilton Anxiety Rating Scale with outpatients. Journal of Anxiety Disorders, 5, 213–223.


GROUP PSYCHOTHERAPY FOR PERFECTIONISM Beck, A. T., Steer, R., & Garbin, M. (1988). Psychometric properties of the Beck Depression Inventory: Twenty-five years of evaluation. Clinical Psychology Review, 8, 77–100. Beutler, L. E. (2010). Predictors of sustained therapeutic change: Some thoughts about conceptualizations. Psychotherapy Research, 20, 55–59. Blankstein, K. R., & Dunkley, D. M. (2002). Evaluative concerns, selfcritical, and personal standards perfectionism: A structural equation modeling strategy. In G. L. Flett & P. L. Hewitt (Eds.), Perfectionism: Theory, research, and treatment (pp. 285–315). Washington, DC: American Psychological Association. Blatt, S. J. (1992). The differential effect of psychotherapy and psychoanalysis with anaclitic and introjective patients: The Menninger Psychotherapy Research Project revisited. Journal of the American Psychoanalytic Association, 40, 691–724. Blatt, S. J., Auerbach, J. S., Zuroff, D. C., & Shahar, G. (2006). Evaluating efficacy, effectiveness, and mutative factors in psychodynamic psychotherapies. In P. D. M. Task Force (Ed.), Psychodynamic diagnostic manual (pp. 537–572). Silver Spring, MD: Alliance of Psychoanalytic Organizations. Blatt, S. J., Quinlan, D. M., Pilkonis, P. A., & Shea, M. T. (1995). Impact of perfectionism and need for approval on the brief treatment of depression: The National Institute of Mental Health Treatment of Depression Collaborative Research Program revisited. Journal of Consulting and Clinical Psychology, 63, 125–132. Blatt, S. J., Zuroff, D. C., Hawley, L. L., & Auerbach, J. S. (2010). Predictors of sustained therapeutic change. Psychotherapy Research, 20, 37–54. Buss, A. H., & Finn, S. E. (1987). Classification of personality traits. Journal of Personality and Social Psychology, 52, 432– 444. Chik, H. M., Whittal, M. L., & O’Neill, M. L. (2008). Perfectionism and treatment outcome in obsessive-compulsive disorder. Cognitive Therapy and Research, 32, 676 – 688. Cohen, J. (1988). Statistical power analysis for the behavioural sciences (2nd ed.). Hillsdale, NJ: Erlbaum. Egan, S. J., van Noort, E., Chee, A., Kane, R. T., Hoiles, K. J., Shafran, R., & Wade, T. D. (2014). A randomised controlled trial of face to face versus pure online self-help cognitive behavioural treatment for perfectionism. Behaviour Research and Therapy, 63C, 107–113. Enns, M. W., & Cox, B. J. (1999). Perfectionism and depression symptom severity in major depressive disorder. Behaviour Research and Therapy, 37, 783–794. Enns, M. W., & Cox, B. J. (2005). Perfectionism, stressful events, and the one year outcome of depression. Cognitive Therapy and Research, 29, 541–553. Enns, M. W., Cox, B. J., & Pidlubny, S. R. (2002). Group CBT for residual depression: Effectiveness and predictors of response. Cognitive Behaviour Therapy, 31, 31– 40. Farber, B. A. (2003). Patient self-disclosure: A review of the research. Journal of Clinical Psychology, 59, 589 – 600. Flett, G. L., & Hewitt, P. L. (2002). Perfectionism: Theory, Research, and Treatment. Washington, DC: American Psychological Association. Flett, G. L., & Hewitt, P. L. (2015). Measures of perfectionism. In G. Boyle, D. Saklofske, & G. Matthews (Eds.), Measures of personality and social psychological constructs (Vol. 2, pp. 595– 618). London: Academic Press. Flett, G. L., Hewitt, P. L., Blankstein, K. R., & Gray, L. (1998). Psychological distress and the frequency of perfectionistic thinking. Journal of Personality and Social Psychology, 75, 1363–1381. Flett, G. L., Hewitt, P. L., & Heisel, M. J. (2014). The destructiveness of perfectionism revisited: Implications for the assessment of suicide risk. Review of General Psychology, 18, 156 –172. Flett, G. L., Hewitt, P. L., Whelan, T., & Martin, T. R. (2007). The Perfectionism Cognitions Inventory: Psychometric properties and asso-

215

ciations with distress and deficits in cognitive self-management. Journal of Rational-Emotive and Cognitive-Behavior Therapy, 25, 255–277. Frost, R. O., Marten, P., Lahart, C., & Rosenblate, R. (1990). The dimensions of perfectionism. Cognitive Therapy and Research, 14, 449 – 468. Frost, R. O., Trepanier, K. L., Brown, E. J., Heimberg, R. G., Juster, H. R., Makris, G. S., & Leung, A. W. (1997). Self-monitoring of mistakes among subjects high and low in perfectionistic concern over mistakes. Cognitive Therapy and Research, 21, 209 –222. Frost, R. O., Turcotte, T. A., Heimberg, R. G., Mattia, J. I., Holt, C. S., & Hope, D. A. (1995). Reactions to mistakes among subjects high and low in perfectionistic concern over mistakes. Cognitive Therapy and Research, 19, 195–205. Gans, J. S., & Counselman, E. F. (2010). Patient selection for psychodynamic group psychotherapy: Practical and dynamic considerations. International Journal of Group Psychotherapy, 60, 197–220. Gelman, A., Hill, J., & Yajima, M. (2012). Why we (usually) don’t have to worry about multiple comparisons. Journal of Research on Educational Effectiveness, 5, 189 –211. Greenspon, T. S. (2008). Making sense of error: A view of the origins and treatment of perfectionism. American Journal of Psychotherapy, 62, 263–282. Hawley, L. L., Ho, M. H., Zuroff, D. C., & Blatt, S. J. (2006). The relationship of perfectionism, depression, and therapeutic alliance during treatment for depression: Latent difference score analysis. Journal of Consulting and Clinical Psychology, 74, 930 –942. Hewitt, P. L., & Flett, G. L. (1991). Perfectionism in the self and social contexts: Conceptualization, assessment, and association with psychopathology. Journal of Personality and Social Psychology, 60, 456 – 470. Hewitt, P. L., & Flett, G. L. (1993). Dimensions of perfectionism, daily stress, and depression: A test of the specific vulnerability hypothesis. Journal of Abnormal Psychology, 102, 58 – 65. Hewitt, P. L., & Flett, G. L. (2004). The Multidimensional Perfectionism Scale: Technical Manual. Toronto: Multihealth Systems Inc. Hewitt, P. L., Flett, G. L., & Ediger, E. (1996). Perfectionism and depression: Longitudinal assessment of a specific vulnerability hypothesis. Journal of Abnormal Psychology, 105, 276 –280. Hewitt, P. L., Flett, G. L., & Mikail, S. F. (in press). Perfectionism: Conceptualization, assessment, and treatment. New York, NY: Guilford Press. Hewitt, P. L., Flett, G. L., Sherry, S. B., Habke, M., Parkin, M., Lam, R. W., . . . Stein, M. B. (2003). The interpersonal expression of perfection: Perfectionistic self-presentation and psychological distress. Journal of Personality and Social Psychology, 84, 1303–1325. Hewitt, P. L., Flett, G. L., Turnbull-Donovan, W., & Mikail, S. (1991). The Multidimensional Perfectionism Scale: Reliability, validity, and psychometric properties in psychiatric samples. Psychological Assessment: A Journal of Consulting and Clinical Psychology, 3, 464 – 468. Hewitt, P. L., & Genest, M. (1990). The ideal self: Schematic processing of perfectionistic content in dysphoric university students. Journal of Personality and Social Psychology, 59, 802– 808. Hewitt, P. L., Habke, A. M., Lee-Baggley, D. L., Sherry, S. B., & Flett, G. L. (2008). The impact of perfectionistic self-presentation on the cognitive, affective, and physiological experience of a clinical interview. Psychiatry: Interpersonal and Biological Processes, 71, 93–122. Hewitt, P. L., Sherry, D., Kaldas, J., Deng, X., Bae, V., & Flett, G. L. (2015). Perfectionism, help-seeking and fears of psychotherapy. Manuscript in preparation. Horney, K. (1950). Neurosis and human growth. New York, NY: Norton. Horowitz, L. M., Rosenberg, S. E., Baer, B. A., Ureño, G., & Villaseñor, V. S. (1988). Inventory of interpersonal problems: Psychometric properties and clinical applications. Journal of Consulting and Clinical Psychology, 56, 885– 892.


216

HEWITT ET AL.

Jacobson, N. S., & Truax, P. (1991). Clinical significance: A statistical approach to defining meaningful change in psychotherapy research. Journal of Consulting and Clinical Psychology, 59, 12–19. Kahn, J. H., Achter, J., & Shambaugh, E. (2001). Client distress disclosure, characteristics at intake, and outcome in brief counseling. Journal of Counseling Psychology, 48, 203–211. Kenny, D. A., Kashy, D. A., & Bolger, N. (1998). Data analysis in social psychology. In D. T. Gilbert, S. T. Fiske, & G. Lindzey (Eds.), The handbook of social psychology (4th ed., Vol. 1 and 2, pp. 233–265). New York, NY: McGraw-Hill. Leary, M. R. (1983). Social anxiousness: The construct and its measurement. Journal of Personality Assessment, 47, 66 –75. Leary, M. R., & Kowalski, R. M. (1993). The Interaction Anxiousness Scale: Construct and criterion-related validity. Journal of Personality Assessment, 61, 136 –146. Little, R. J. A. (1988). A test of missing completely at random for multivariate data with missing values. Journal of the American Statistical Association, 83, 1198 –1202. Lloyd, S., Fleming, C., Schmidt, U., & Tchanturia, K. (2014). Targeting perfectionism in anorexia nervosa using a group-based cognitive behavioural approach: A pilot study. European Eating Disorders Review, 22, 366 –372. Lundh, L.-G., & Öst, L.-G. (2001). Attentional bias, self-consciousness and perfectionism in social phobia before and after cognitive-behaviour therapy. Scandinavian Journal of Behaviour Therapy, 30, 4 –16. MacKenzie, K. R. (1983). The clinical application of a group climate measure. In R. R. Dies & K. R. MacKenzie (Eds.), Advances in group psychotherapy: Integrating research and practice (pp. 159 –170). New York, NY: International Universities Press. MacKenzie, K. R. (1990). Introduction to time-limited group psychotherapy. Washington, DC: American Psychiatric Press, Inc. McCown, W., & Carlson, G. (2004). Narcissism, perfectionism, and selftermination from treatment in outpatient cocaine users. Journal of Rational-Emotive and Cognitive-Behavior Therapy, 22, 329 –340. McWilliams, N. (2004). Psychoanalytic psychotherapy: A practitioner’s guide. New York, NY: Guilford Press. Molnar, F. J., Hutton, B., & Fergusson, D. (2008). Does analysis using “last observation carried forward” introduce bias in dementia research? Canadian Medical Association Journal, 179, 751–753. Nezlek, J. B. (2001). Multilevel random coefficient analyses of event- and interval-contingent data in social and personality psychology research. Personality and Social Psychology Bulletin, 27, 771–785. O’Connor, R. C. (2007). The relations between perfectionism and suicidality: A systematic review. Suicide and Life-Threatening Behavior, 37, 698 –714. Raudenbush, S., & Bryk, A. (2002). Hierarchical linear models. Lincolnwood, IL: Scientific Software International.

Raudenbush, S., Bryk, A., Cheong, Y. F., Congdon, R., & du Toit, M. (2004). HLM6: Hierarchical linear and non-linear modeling. Lincolnwood, IL: Scientific Software. Riley, C., Lee, M., Cooper, Z., Fairburn, C. G., & Shafran, R. (2007). A randomised controlled trial of cognitive-behaviour therapy for clinical perfectionism: A preliminary study. Behaviour Research and Therapy, 45, 2221–2231. Rosser, S., Issakidis, C., & Peters, L. (2003). Perfectionism and social phobia: Relationships between the constructs and impact on cognitive behavior therapy. Cognitive Therapy and Research, 27, 143–151. Saha, C., & Joines, M. P. (2009). Bias in the last observation carried forward method under informative dropout. Journal of Statistical Planning and Inference, 139, 246 –255. Salbach-Andrae, H., Bohnekamp, I., Pfeiffer, E., Lehmkuhl, C., & Miller, A. (2008). Dialectical behavior therapy of anorexia and bulimia nervosa among adolescents: A case series. Cognitive and Behavioral Practice, 15, 415– 425. Shafran, R., Cooper, Z., & Fairburn, C. G. (2002). Clinical perfectionism: A cognitive-behavioural analysis. Behaviour Research and Therapy, 40, 773–791. Shahar, G., Blatt, S. J., Zuroff, D. C., Krupnick, J. L., & Sotsky, S. M. (2004). Perfectionism impedes social relations and response to brief treatment for depression. Journal of Social and Clinical Psychology, 23, 140 –154. Shedler, J. (2010). The efficacy of psychodynamic psychotherapy. American Psychologist, 65, 98 –109. Slaney, R. B., Rice, K., Mobley, M., Trippi, J., & Ashby, J. (2001). The revised almost perfect scale. Measurement and Evaluation in Counseling and Development, 34, 130 –145. Sorotzkin, B. (1998). Understanding and treating perfectionism in religious adolescents. Psychotherapy: Theory, Research, Practice, Training, 35, 87–95. Sullivan, H. S. (1953). The interpersonal theory of psychiatry. New York, NY: Norton & Co. Tasca, G. A., Illing, V., Joyce, A. S., & Ogrodniczuk, J. S. (2009). Three-level multilevel growth models for nested change data: A guide for group treatment researchers. Psychotherapy Research, 19, 453– 461. Tasca, G. A., Mikail, S. F., & Hewitt, P. L. (2005). Group psychodynamic interpersonal psychotherapy: Summary of a treatment model and outcomes for depressive symptoms. In M. E. Abelian (Ed.), Focus on psychotherapy research (pp. 159 –188). Hauppauge, NY: Nova Science Publishers. Wachtel, P. L. (1994). Cyclical processes in personality and psychopathology. Journal of Abnormal Psychology, 103, 51–54. Yalom, I. D., & Leszcz, M. (2005). The theory and practice of group psychotherapy (5th ed.). New York, NY: Basic Books Inc.


217

Appendix Multilevel Models Model 1: Three-Level Models to Assess Intraclass Correlation Coefficients Level 1: Ytij ⫽ ␲0ij ⫹ ␲1ij(logtime) ⫹ etij Level 2: ␲ 0ij ⫽ ␤00j ⫹ ␤01j(individual prescore) ⫹ r0ij ␲1ij ⫽ ␤10j ⫹ ␤11j(individual prescore) ⫹ r1ij Level 3: ␤00j ⫽ ␥000 ⫹ ␥001(group prescore) ⫹ ␥002(condition) ⫹ u00j This document is copyrighted by the American Psychological Association or one of its allied publishers. This article is intended solely for the personal use of the individual user and is not to be disseminated broadly.

␤01j ⫽ ␥010 ⫹ u01j ␤10j ⫽ ␥100 ⫹ ␥101(group prescore ⫹ ␥102(condition) ⫹ u10j ␤11j ⫽ ␥110 ⫹ u11j Individual prescores were group mean centered, and group prescores were grand mean centered. Model 2: Two-Level Multilevel Models to Assess Change in the Outcome Level 1: Yti ⫽ ␲0i ⫹ ␲1i(logtime) ⫹ eti Level 2: ␲0i ⫽ ␤00 ⫹ ␤01(individual prescore) ⫹ r0i ␲1i ⫽ ␤10 ⫹ ␤11(individual prescore) ⫹ r1i Model 3: Two-Level Multilevel Models to Assess Perfectionism Scales as Predictors of Outcomes Level 1: Yti ⫽ ␲0i ⫹ ␲1i(logtime) ⫹ eti Level 2: ␲0i ⫽ ␤00 ⫹ ␤01(individual prescore) ⫹ ␤02(Disclose slope) ⫹ ␤03(Display slope) ⫹ ␤04(Other slope) ⫹ ␤05(Perfectionism Cognitions slope) ⫹ ␤06(Promote slope) ⫹ ␤07(Self slope) ⫹ ␤08(Social slope) ⫹ r0i ␲1i ⫽ ␤10 ⫹ ␤11(individual prescore) ⫹ ␤12(Disclose slope) ⫹ ␤13(Display slope) ⫹ ␤14(Other slope) ⫹ ␤15(PCI slope) ⫹ ␤16(Promote slope) ⫹ ␤17(Self slope) ⫹ ␤18(Social slope) ⫹ r1i Individual prescores in these two-level models were centered. Received February 10, 2015 Accepted February 14, 2015 䡲