A randomised controlled trial of group cognitive behavioural therapy for perfectionism

Behaviour Research and Therapy 68 (2015) 37e47

Contents lists available at ScienceDirect

Behaviour Research and Therapy journal homepage: www.elsevier.com/locate/brat

A randomised controlled trial of group cognitive behavioural therapy for perfectionism Alicia K. Handley*, Sarah J. Egan, Robert T. Kane, Clare S. Rees School of Psychology and Speech Pathology, Faculty of Health Sciences, Curtin University, Australia

a r t i c l e i n f o

a b s t r a c t

Article history: Received 22 December 2013 Received in revised form 5 February 2015 Accepted 21 February 2015 Available online 25 February 2015

Perfectionism is associated with symptoms of anxiety disorders, eating disorders and mood disorders. Treatments targeting perfectionism may reduce the symptoms of these disorders (Egan, Wade, & Shafran, 2011). This study is the first randomised controlled trial to investigate the efficacy of group cognitive behavioural therapy (CBT) for perfectionism. Forty-two participants with elevated perfectionism and a range of anxiety, eating and mood disorders were randomised to group CBT for perfectionism or a waitlist control. The treatment group reported significantly greater pre-post reductions in perfectionism, symptoms of depression, eating disorders, social anxiety, anxiety sensitivity and rumination, as well as significantly greater pre-post increases in self-esteem and quality of life compared to the waitlist control group. The impact of treatment on most of these outcomes was mediated by pre-post change in perfectionism (Concern over Mistakes). Treatment gains were reliable and clinically significant, and were maintained at 6-month follow-up. Findings support group CBT for perfectionism being an efficacious treatment for perfectionism and related psychopathology, as well as increasing self-esteem and quality of life. © 2015 Elsevier Ltd. All rights reserved.

Keywords: Perfectionism Group Cognitive behavioural therapy

Perfectionism involves setting demanding standards and having a significant concern over mistakes (Frost, Marten, Lahart, & Rosenblate, 1990). This definition has primarily arisen from the results of factor analytic studies of multidimensional perfectionism measures (Frost et al., 1990; Hewitt & Flett, 1991) as opposed to being based on theoretical descriptions of the construct (Shafran, Cooper, & Fairburn, 2002). Perfectionism has been identified as an important factor in the onset, severity and maintenance of anxiety, mood and eating disorders (Egan, Wade, & Shafran, 2011). It can account for the co-morbidity of disorders (Bieling, Summerfeldt, Israeli, & Antony, 2004) and can negatively impact treatment (e.g., Blatt, Quinlan, Pilkonis, & Shea, 1995). Consequently, perfectionism has been argued to be a transdiagnostic process (Egan et al., 2011). This implies that treatments targeting perfectionism may not only reduce perfectionism but also the symptoms of related disorders (Egan et al., 2011). Several studies have examined the efficacy of cognitive behavioural therapy (CBT) for perfectionism (CBT-P) and the results so far are promising. A recent meta-analysis examined the outcomes of

* Corresponding author. School of Psychology and Speech Pathology, Curtin University, GPO Box U1987, Perth, Western Australia, 6845, Australia. E-mail address: [email protected] (A.K. Handley). http://dx.doi.org/10.1016/j.brat.2015.02.006 0005-7967/© 2015 Elsevier Ltd. All rights reserved.

eight studies of CBT-P and found large pooled effect sizes for prepost treatment reductions in perfectionism and medium pooled effect sizes for pre-post treatment reductions in depression and anxiety (Lloyd, Schmidt, Khondoker, & Tchanturia, 2014). Only two of the studies included in the meta-analysis were RCTs examining the efficacy of CBT-P in clinical samples. Riley, Lee, Cooper, Fairburn, and Shafran (2007) found in a sample of participants, of whom 70% met diagnoses of anxiety and depression, that CBT-P resulted in significant reductions in Clinical Perfectionism Questionnaire (CPQ; Fairburn, Cooper, & Shafran, 2003) scores relative to a waitlist control condition. Post-treatment reductions in perfectionism, anxiety and depression were maintained at 4-month follow-up. Clinically significant change in perfectionism occurred in 75% of participants and the number of anxiety and depression diagnoses halved at post-treatment. Steele and Wade (2008) examined the efficacy of guided self-help CBT-P relative to standard CBT for Bulimia Nervosa (BN; Cooper, 1993, in Steele & Wade, 2008) and mindfulness (Segal, Williams, & Teasdale, 2002) in individuals with BN and Eating Disorder Not Otherwise Specified (EDNOS). Participants in all conditions demonstrated significant decreases in perfectionism, depression and eating disorder symptoms and significant increases in self-esteem between pre- and post-treatment. CBT-P tended to produce larger effect sizes for co-morbid anxiety

38

A.K. Handley et al. / Behaviour Research and Therapy 68 (2015) 37e47

and depression compared to the other conditions. These trends have clinical relevance as they are consistent with perfectionism having a transdiagnostic role (Egan et al., 2011); however, the findings are limited by the absence of a pure control condition (Steele & Wade, 2008). Only two studies have explored the effectiveness of group CBT-P in individuals with elevated perfectionism and psychological disorders and neither study was an RCT (Egan & Stout, 2007; Steele et al., 2013). In Egan and Stout's (2007) single case series design, downward trends were observed in perfectionism, depression and anxiety, with one of the three participants demonstrating clinically significant decreases in anxiety and depression. Steele et al.'s (2013) case series design investigated the effectiveness of psychoeducation and group CBT-P in 21 individuals with elevated perfectionism as well as anxiety disorders or current or past depression. Psycho-education did not produce any significant reductions. Following group CBT-P there were statistically significant decreases in perfectionism, stress, anxiety and depression that were maintained at 3-month follow-up, and clinically significant change in perfectionism was observed in 32% of participants; however, there was no separate waitlist control group. Group therapy has some advantages over individual therapy such as time efficiency, decreased cost (Himle, Van Etten, & Fischer, 2003) and therapeutic benefits (Bieling, McCabe, & Antony, 2013); hence it would be useful to determine the efficacy of group CBT-P in an RCT. Furthermore, it is important to determine if pre-post changes in perfectionism are responsible for pre-post changes in symptoms as there has been no analysis of mediators of change in the perfectionism treatment literature to date. As reviewed, there is some promising preliminary evidence for the efficacy of CBT-P, with the majority of studies having investigated the intervention when delivered in an individual format. Only two uncontrolled studies have investigated CBT-P delivered via a group therapy approach and the results are promising (Egan & Stout, 2007; Steele et al., 2013). Currently, no RCTs of group CBT-P have been conducted. Therefore, the aim of the current study is to compare group CBT-P with a waitlist control group. It is predicted that participants who receive group CBT-P will show statistically and clinically significant reductions in perfectionism from pre-to post-test, whereas the control group will show only negligible pre-post changes. As perfectionism is proposed to be a transdiagnostic process (Egan et al., 2011), it is also predicted that participants receiving group CBT-P will show statistically and clinically significant pre-post reductions in symptoms of eating disorders, depression and anxiety, as well as statistically significant and reliable pre-post increases in self-esteem and quality of life, whereas the control group will show only negligible pre-post changes. It is also hypothesised that pre-post changes in perfectionism will account for pre-post changes in symptoms. Finally, it is predicted that all changes will be maintained at 6-month follow-up.

substance abuse or substance dependence, as determined by the Mini International Neuropsychiatric Interview (MINI, Sheehan et al., 1998). Participants were required to abstain from external psychological treatment between baseline and follow-up. Those on medication were to be on a stable dose for one month before the study and throughout the trial. Using G*Power (Version 3.1; Faul, Erdfelder, Lang, & Buchner, 2007), it was estimated that approximately 40 participants (20 per group) would be required for the Generalised Linear Mixed Model (GLMM; Holden, Kelley, & Agarwal, 2008) procedure to have an 80 percent chance of detecting moderate to large interaction effects at a per-test alpha level of 0.05. Forty-three participants (79% females; 21% males) were eligible.

Method

Treatment adherence and collaborative structure

Participants

Two independent Clinical Psychologists rated five randomly selected videotapes of group therapy. Consistent with an RCT of group CBT for obsessive-compulsive disorder (Anderson & Rees, 2007), adherence to session objectives was rated on 7-point Likert-type scales. Inter-rater reliability of the adherence measure was high (r ¼ .89), and mean adherence to protocol was high (M ¼ 6.71/7; SD ¼ 0.98). Therapist behaviour (e.g., warmth) was rated using nine items from the Collaborative Study Psychotherapy Rating Scale-6 (CSPRS-6; Evans, Piasecki, Kriss, & Hollon, 1984) and mean scores were high: warmth: 6.1/7 (SD ¼ 0.87); supportive encouragement: 6.6/7 (SD ¼ 0.52); empathy: 5.8/7 (SD ¼ 0.92); rapport: 5.5/7 (SD ¼ 1.08), conveyance of expertise: 5.6/7

Participants self-referred in response to advertisements distributed to universities, GPs, psychologists, psychiatrists and workplaces throughout Perth, Australia. Participants were required to have elevated perfectionism, defined by a score greater than 24.7 on the Concern over Mistakes (CM) subscale of the Frost Multidimensional Perfectionism Scale (FMPS; Frost et al., 1990). This is the average CM score derived from previous research examining perfectionism in anxiety disorder samples included in Egan et al.'s (2011) review. Exclusion criteria included self-harm, moderate or severe suicidality, a body mass index below 17.5, psychosis, and

Design and procedure After baseline assessment, participants were randomly allocated to treatment or control conditions. Randomisation was conducted using randomised number lists generated by Saghaei's (2004) Random Allocation Software Version 1.0. The treatment group commenced therapy while the control group received no therapy. After eight weeks all participants were re-assessed. Participants from the control group received group CBT-P followed by their post-treatment assessment (the treated control group). Treatment and treated control groups were assessed at 3-months and 6months post-treatment. This RCT complied with CONSORT guidelines (Moher et al., 2010) and was registered as a clinical trial with the Australian and New Zealand Clinical Trials Registry (2007). Ethics approval was received from the Curtin University Human Research Ethics Committee. Treatment Treatment consisted of eight 2-hour group therapy sessions over eight weeks. Treatment was adapted from Shafran, Egan, and Wade (2010). This protocol was evaluated in the pilot trial of group CBT-P in a different sample (Steele et al., 2013). Sessions included understanding perfectionism, motivation to change, challenging perfectionist beliefs through behavioural experiments and thought diaries, decreasing procrastination and self-criticism, and balancing self-esteem (Shafran et al., 2010). Therapists Treatment was administered by a Clinical Psychologist Registrar (Handley) who conducted four groups, as well as two Master of Clinical Psychology students who each co-facilitated two groups. Students were trained by a Clinical Psychologist (Egan) who is a coauthor of the CBT-P treatment. Egan also provided weekly supervision to ensure treatment adherence.

A.K. Handley et al. / Behaviour Research and Therapy 68 (2015) 37e47

39

Table 1 Description of the measures used grouped according to conceptual category. Conceptual Category

Measure

Description

Perfectionism

Frost Multidimensional Perfectionism Scale (FMPS; Frost et al., 1990).  Concern over Mistakes (CM)  Doubts about Actions (DA)  Personal Standards (PS)

The CM subscale measures concern over making performance errors. The DA subscale assesses doubts in regard to performance. The PS subscale assesses the setting of very high standards of performance. These subscales have high internal consistency (Frost et al., 1990) and validity (Antony, Purden, Huta, & Swinson, 1998). In the current study a ¼ .89 (CM); .72 (DA); and .81 (PS). The CPQ assesses clinical perfectionism and has adequate test-retest reliability, internal consistency and validity (Egan et al., 2015; Steele, O'Shea, Murdock, Karney, & Wade, 2011). In the present study a ¼ .77. The DAS-SC assesses the perception one has of themselves and others when standards are not attained. It has high internal consistency and validity (Dunkley, Sanislow, Grilo, & McGlashan, 2006; Imber et al., 1990). In this study a ¼ .93. This scale was adapted from the Anxiety Disorders Interview Schedule-adult version (ADIS-adult; Brown et al., 1994) and assesses interference associated with perfectionism. This scale was adapted from the ADIS-adult (Brown et al., 1994) and assesses distress associated with perfectionism. The BDI-II assesses depression and has high internal consistency and validity (Beck et al., 1996). In this study a ¼ .93. The EDE-Q measures concerns about shape, weight and eating as well as eating behaviours. This scale has high internal consistency and validity (Mond, Hay, Rodgers, Owen, & Beumont, 2004a; 2004b). In the present study a ¼ .96 (EDE-Q total). This subscale assesses anxiety and has high internal consistency and validity (Antony, Bieling, Cox, Enns, & Swinson, 1998). In the current study a ¼ .86. The FNE-B examines social-evaluative anxiety. This scale has high internal consistency and validity (Collins et al., 2005). In this study a ¼ .90 The PSWQ measures the uncontrollability and excessiveness of worry. It has high internal consistency and validity (Meyer et al., 1990). In the current study a ¼ .90. The ASI-3 assesses fear of the somatic sensations evoked by anxiety. It has high internal consistency and validity (Taylor et al., 2007). In the present study a ¼ .88. The Q-LES-Q-18 measures life satisfaction and enjoyment. It has high internal consistency and validity (Ritsner et al., 2005). In this study a ¼ .91 The RSES assesses global self-esteem and has high internal consistency (Steele & Wade, 2008) and validity (Gray-little, Williams, & Hancock, 1997). In the current study a ¼ .88. The RNTS assesses the extent of repetitive negative thinking in response to a distressing event. It has high internal consistency and validity (McEvoy et al., 2010). In this study a ¼ .92

Clinical Perfectionism Questionnaire (CPQ; Fairburn et al., 2003).

Dysfunctional Attitudes Scale-Self Criticism (DAS-SC; Weissman & Beck, 1978; Imber et al., 1990).

Interference Scale (Brown, DiNardo, & Barlow, 1994). Distress Scale (Brown et al., 1994).

Depression

Beck Depression Inventory-II (BDI-II; Beck, Steer, & Brown, 1996).

Eating Disorder Pathology

Eating Disorder Examination Questionnaire (EDE-Q: total score; Fairburn & Beglin, 1994).

General Anxiety

Anxiety subscale of the Depression, Anxiety and Stress Scale-21 (Lovibond & Lovibond, 1995).

Social Anxiety

Fear of Negative Evaluation Scale-brief version (FNE-B; Collins, Westra, Dozois, & Stewart, 2005).

Pathological Worry

Penn State Worry Questionnaire (PSWQ; Meyer, Miller, Metzger, & Borkovec, 1990).

Anxiety Sensitivity

Anxiety Sensitivity Index-3 (ASI-3; Taylor et al., 2007).

Quality of Life

Quality of Life, Enjoyment and Satisfaction Questionnaire-18 (Q-LES-Q-18; Ritsner, Kurs, Gibel, Ratner, & Endicott, 2005).

Self-Esteem

Rosenberg Self-Esteem Scale (RSES; Rosenberg, 1965).

Rumination

Repetitive Negative Thinking Scale (RNTS; McEvoy, Mahoney, & Moulds, 2010).

(SD ¼ 0.97); formality: 5.8/7 (SD ¼ 0.63); collaboration: 4.7/7 (SD ¼ 1.83); encouragement of independence: 5.4/7 (SD ¼ 1.07); and relating improvement to cognitive change: 5.5/7 (SD ¼ 1.43).

Measures The self-report measures used in this study are described in Table 1. Conceptually-related outcome measures have been grouped together to highlight the categories to which Bonferroni corrections were applied (see statistical significance section).

Mini International Neuropsychiatric Interview, Version 5.0 (MINI; Sheehan et al., 1998) The MINI was used to determine diagnoses. All interviews were conducted by a single interviewer (Handley) with four years of experience using this interview. Consequently, it was not possible to determine inter-rater reliability; however, the interviewer discussed diagnoses with an experienced Clinical Psychologist (Egan). The MINI has high validity, internal consistency and test-retest reliability (Sheehan et al., 1997).

Client adherence to treatment The number of sessions attended was recorded. Reading and homework compliance were examined using items adapted from Thiels, Schmidt, Troop, Treasure, and Garthe (2001) and Troop et al. (1996). The homework measure has moderately high internal consistency (Steele et al., 2013). In the current study a ¼ .73. Statistical analyses Statistical significance GLMM, as implemented through SPSS Version 20.0, was used to test for pre-post changes and their maintenance (Holden et al., 2008). As GLMM is a full information estimation procedure, it utilises all of the data available at each assessment period to obtain the parameter estimates. Consequently, use of GLMM decreases sampling bias and the need for missing data to be replaced (Elobeid et al., 2009; Holden et al., 2008). Time, group and the Group  Time interaction were treated as fixed effects; participant and therapy group were treated as random effects. The data was analysed within the context of a hierarchical structure in which participants were nested in therapy groups (Holden et al., 2008).

40

A.K. Handley et al. / Behaviour Research and Therapy 68 (2015) 37e47

To control for the inflation of the family-wise error rate, Bonferroni corrections were applied according to the conceptual categories specified in Table 1. The perfectionism category consisted of seven outcome measures (.05/7, a ¼ .007). The remaining categories in Table 1 each contained one outcome measure, as each of these outcome measures assessed a different form of psychopathology. Thus, the standard alpha level was applied to these outcome measures (.05/1, a ¼ .05). Partial eta-squared statistics were calculated to determine the effect size of the interaction effects (Richardson, 2011). A partial eta-squared value of .01 signifies a small effect, .06 a moderate effect and .15þ a large effect (Richardson, 2011). For the significant Group  Time interactions, post-hoc Least Significant Difference (LSD) contrasts were conducted across the simple main effects of time (Keppel & Wickens, 2004). Cohen's d statistics were used to estimate the effect sizes of the LSD contrasts. The formula used to calculate Cohen's d h .rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi  . ffii. pffiffiffiffiffiffiffiffi was Mean1  Mean2 SD21 þ SD22 2 ½1  ð r1;2 Þ, where Mean1 is the pre-test mean, Mean2 is the post-test mean, SD1 is the pre-test standard deviation, SD2 is the post-test standard deviation and r1,2 is the correlation between the outcome measure at pre-test and post-test (Cohen, 1988). Post hoc mediational analyses For outcomes showing post-treatment effects, post hoc mediational analyses were conducted to explore whether pre-post change in perfectionism mediated the impact of group CBT-P on pre-post changes in the outcomes. Each outcome was analysed in a separate mediation model; Bonferroni correction procedures were subsequently applied to accommodate the number of mediation models being tested. Concern over Mistakes was selected as the mediator as it has been frequently reported in the perfectionism literature and is clinically relevant due to its relationships with numerous psychological disorders (Egan et al., 2011). Standardised path estimates and standard errors for the indirect pathways passing from the binary group variable (intervention, control) through pre-post change in Concern over Mistakes to the pre-post changes in outcomes were estimated with a bootstrapping procedure based on 1000 draws as implemented by Mplus (Version 5.2; n & Muthe n, 2007). Muthe Reliable and clinically significant change For outcomes showing intervention effects, reliable change was calculated for each participant. A Reliable Change Index greater than an absolute value of 1.96 is indicative of reliable change (Jacobson & Truax, 1991). If a participant demonstrated reliable change in perfectionism (Concern over Mistakes) and his/her posttest Concern over Mistakes score fell below a pre-defined clinical cut-off, this participant had experienced clinically significant change in Concern over Mistakes (Jacobson & Truax, 1991; Riley et al., 2007). Fisher's 1-sided tests were used to test for between-group differences in the proportion of individuals showing reliable change in outcomes, clinically significant change in Concern over Mistakes and recovery from psychological disorders. Results Participant flow and sample characteristics There were 154 individuals who expressed interest; 43 were eligible and 42 were randomised, as seen in the CONSORT diagram in Fig. 1 (Moher et al., 2010). As reported in Table 2, the intervention

and control groups did not significantly differ on any outcome measure at baseline. Treatment effects on perfectionism As seen in Table 3, there were significant Group  Time interactions for perfectionism (CM, PS, DA, CPQ, DAS-SC, distress and interference) at the Bonferroni-corrected a-level of .007. Effect sizes for the interactions ranged from moderate to large. Post-hoc analyses revealed that group CBT-P resulted in significant pre-post reductions in all measures of perfectionism. Effect sizes were all large except for PS, which was moderate. There were no significant changes in perfectionism for the control group. Treatment effects on psychopathology, self-esteem and quality of life There were significant Group  Time interactions for depression and eating pathology at the conventional alpha level of .05. Effect sizes for these interactions were moderate to large. There was a large significant Group  Time interaction for social anxiety, a moderate significant Group  Time interaction for anxiety sensitivity and a small significant Group  Time interaction for rumination. There was a large significant Group  Time interaction for self-esteem and a moderate significant Group  Time interaction for quality of life. The interactions for the remaining variables were not significant. Post-hoc analyses indicated that the treatment group displayed a moderate significant pre-post reduction in depression and a small significant pre-post reduction in eating disorder pathology. The treatment group also exhibited large significant pre-post reductions in social anxiety and rumination, small significant prepost reductions in anxiety sensitivity, large significant pre-post increases in self-esteem and small significant pre-post increases in quality of life. For the control group, there were no significant changes in eating disorder pathology, social anxiety, rumination, anxiety sensitivity, self-esteem or quality of life; however, there was a small significant decrease in depression. Post hoc mediational analyses Post hoc mediational analyses were conducted to explore whether pre-post change in Concern over Mistakes significantly mediated the impact of group CBT-P on pre-post changes in seven outcomes: depression, eating pathology, social anxiety, anxiety sensitivity, rumination, self-esteem and quality of life. Six of the seven indirect pathways were significant at the conventional pertest alpha-level of .05. Five of the seven indirect pathways were significant at the Bonferroni-adjusted alpha-level of .007, which indicated that pre-post change in Concern over Mistakes significantly mediated pre-post changes in these outcomes. The bootstrapped path estimates, standard errors and two-tailed p-values for the indirect pathways are reported in Table 4. Maintenance of change at follow-up There was no significant difference between the treatment and treated control groups in group therapy session attendance (treatment group: M ¼ 5.85, SD ¼ 1.93; treated control group: (M ¼ 5.39, SD ¼ 2.30; p ¼ .510). As seen in Table 5, when the treatment group was compared to the treated control group across pre-treatment, post-treatment and follow-up, the interaction effects for all outcomes except rumination were non-significant. This suggested that the treated control group made comparable gains to the treatment group after receiving group CBT-P. For rumination, the treatment group exhibited a large significant pre-post

A.K. Handley et al. / Behaviour Research and Therapy 68 (2015) 37e47

41

Fig. 1. CONSORT diagram displaying the flow of participants through the study.

reduction (t (72) ¼ 3.48, p ¼ .001), whereas the treated control group did not demonstrate significant change between their preand post-treatment period; however, they did exhibit a moderate significant change between their pre-treatment and 3-month follow-up period. For the outcomes associated with non-significant interactions, significant main effects of time were observed for all perfectionism outcomes, depression, eating disorder pathology, social anxiety,

anxiety sensitivity, self-esteem and quality of life. Treatment gains were maintained at 3-month and 6-month follow-ups. Reliable and clinically significant change As demonstrated in Table 6, Fisher's exact 1-sided tests revealed that a significantly greater proportion of participants in the treatment group exhibited pre-post reliable change in perfectionism (CM,

42

A.K. Handley et al. / Behaviour Research and Therapy 68 (2015) 37e47

Table 2 Means and (standard errors) of baseline outcome variables for treatment and control conditions.

Age Sex (female) On medication English as first language Occupation Psychological Disorder Number of disorders CM PS DA CPQ DAS-SC Distress Interference DASS-anx EDE-Q-total RSES Q-LES-Q-18 RNTSb PSWQ BDI-II FNE-Bb ASI-3b

Treatment (n ¼ 21)

Control (n ¼ 21)

t(df), p

28.86 (8.27) 17 8 (38.10%) 17 11 students, 10 employed 21 with current diagnosis 2.10 (0.23) 33.14 (1.32) 28.86 (1.04) 15.61 (0.64) 32.10 (1.17) 68.43 (2.74) 6.62 (0.36) 6.90 (0.35) 12.10 (2.38) 1.56 (0.27) 25.05 (0.75) 3.20 (0.16) 34.33 (2.18) 65.10 (1.65) 17.45 (2.61) 49.93 (2.22) 25.62 (3.15)

33.00 (12.30) 17 6 (28.60%) 19 11 students, 10 employed 17 with current diagnosis 1.95 (1.47) 33.14 (1.35) 28.57 (0.78) 15.66 (0.61) 31.24 (1.26) 71.29 (4.05) 6.10 (0.41) 6.19 (0.43) 11.62 (2.16) 2.11 (0.34) 25.81(1.30) 3.36 (0.17) 31.10 (2.48) 66.90 (2.51) 18.59 (2.67) 50.79 (1.99) 27.33 (3.95)

t(40) ¼ 1.28, p ¼ .208 a Fisher's exact p ¼ 1.000 a Fisher's exact p ¼ .744 a Fisher's exact p ¼ .663 a Fisher's exact p ¼ 1.000 a Fisher's exact p ¼ .107 t(40) ¼ 3.64, p ¼ .718 t(40) ¼ 0.00, p ¼ 1.000 t(40) ¼ 0.22, p ¼ .827 t(40) ¼ -0.05, p ¼ .957 t(40) ¼ 0.50, p ¼ .620 t(40) ¼ -0.59, p ¼ .562 t(40) ¼ 0.95, p ¼ .346 t(40) ¼ 1.28, p ¼ .208 t(40) ¼ 0.15, p ¼ .883 t(40) ¼ 1.24, p ¼ .222 t(40) ¼ 0.51, p ¼ .614 t(40) ¼ 0.60, p ¼ .555 t(40) ¼ 0.98, p ¼ .332 t(40) ¼ 0.60, p ¼ .551 t(36) ¼ 0.18, p ¼ .856 t(26) ¼ 0.29, p ¼ .776 t(26) ¼ 0.33, p ¼ .742

CM ¼ Concern over Mistakes; PS ¼ Personal Standards; DA ¼ Doubts about Actions; CPQ ¼ Clinical Perfectionism Questionnaire; DAS-SC ¼ Dysfunctional Attitudes Scale-SelfCriticism; Distress ¼ Distress scale; Interference ¼ Interference scale; DASS-anx ¼ Anxiety subscale of the Depression, Anxiety and Stress Scale (DASS-21); EDE-Qtotal ¼ Eating Disorder Examination Questionnaire (EDE-Q) total score; RSES ¼ Rosenberg Self-esteem Scale; Q-LES-Q- 18 ¼ Quality of Life Enjoyment and Satisfaction Questionnaire-18; RNTS ¼ Repetitive Negative Thinking Scale; PSWQ ¼ Penn State Worry Questionnaire; BDI-II ¼ Beck Depression Inventory-II; FNE-B ¼ Brief Fear of Negative Evaluation Scale; ASI-3 ¼ Anxiety Sensitivity Index-3. a Fisher's exact (2-sided) test value reported for nominal outcomes. b RNTS: n ¼ 18 in treatment condition, n ¼ 20 in control condition, n ¼ 38 total; FNE-B: n ¼ 14 in treatment condition, n ¼ 14 in control condition, n ¼ 28 total; ASI-3: n ¼ 13 in treatment condition, n ¼ 15 in control condition, n ¼ 28 total.

DA, DAS-SC), self-esteem and quality of life relative to the control group. After the control group received group CBT-P there were no longer significant differences between the treatment group and treated control group in the proportion of participants experiencing pre-post reliable change, suggesting that the treated control group made similar gains to the treatment group following group CBT-P. Fisher's exact 1-sided tests revealed that a significantly greater proportion of participants in the treatment group demonstrated pre-post clinically significant change in perfectionism (CM) compared to the control group (p ¼ .008). Fisher's exact 1-sided tests also indicated that a significantly greater proportion of individuals in the treatment group recovered from depression between pre- and post-treatment compared to the control group (p ¼ .035). There were no significant group differences in the proportions of individuals who recovered from the other disorders. After the control group received group CBT-P there were no longer significant differences between the treatment group and treated control group in the proportion of participants exhibiting pre-post clinically significant change in perfectionism (CM) or pre-post recovery from depression, suggesting that the treated control group made comparable gains to the treatment group following group CBT-P. Once the entire sample had received group CBT-P, there were high percentages of recovery from psychological disorders at 6month follow-up. Specifically, over 80 percent of participants who had a pre-treatment diagnosis of depression, generalised anxiety disorder, obsessive-compulsive disorder or an eating disorder no longer had these diagnoses at 6-month follow-up. Sixtyseven percent of participants with a pre-treatment diagnosis of social phobia no longer met this diagnosis at 6-month follow-up. Discussion This study provides evidence that group CBT-P effectively reduces perfectionism and this effect is maintained at 6-month

follow-up. This RCT found that CBT-P reduces Concern over Mistakes, Personal Standards, Doubts about Actions and Dysfunctional Attitude Scale-Self Criticism (DAS-SC) scores in addition to Clinical Perfectionism Questionnaire (CPQ) scores to a significantly greater extent than a control group. While Riley et al. (2007) found significant Group  Time interaction effects for the CPQ, previous RCTs of perfectionism interventions (Pleva & Wade, 2006; Riley et al. 2007; Steele & Wade, 2008) have not found significant interaction effects for perfectionism as measured by the multidimensional scales and have not measured perfectionism using DAS-SC. It would be useful for future research to examine group CBT-P versus individual CBT-P to determine if group treatment is more efficacious given these encouraging results. Additionally, this is the first RCT to find that a perfectionism treatment decreases distress and interference associated with perfectionism to a significantly greater extent than a control group. These findings build upon those of Steele et al.'s (2013) case series design; however, the greater internal validity of the current RCT enables one to better attribute these findings to the efficacy of group CBT-P. The current findings also indicate that group CBT-P significantly reduces social anxiety, anxiety sensitivity, rumination, depression and eating disorder pathology, and significantly increases selfesteem and quality of life, which are maintained at 6-month follow-up. Furthermore, this is the first RCT to demonstrate that the effects of CBT-P on many of these outcomes are mediated by pre-post changes in perfectionism, specifically pre-post change in Concern over Mistakes. Previous perfectionism treatment trials have not included mediational analyses (Pleva & Wade, 2006; Riley et al. 2007; Steele & Wade, 2008). Such findings are promising for transdiagnostic treatments as they highlight that CBT-P significantly decreases a wide range of psychological symptoms through its impact on perfectionism. This supports arguments of perfectionism being a transdiagnostic process (Egan et al., 2011). Moreover, this is the first RCT to demonstrate that CBT-P significantly

a

a

a

a

a

a

a

a

a

a

a

a

a a

a

a

a

a

a

a

a

a

a

a

a

a

Note. Change in subscript denotes a significant change between baseline and post-treatment/post-waitlist; Partial h ¼ partial eta squared; CM ¼ Concern over Mistakes; PS ¼ Personal Standards; DA ¼ Doubts about Actions; CPQ ¼ Clinical Perfectionism Questionnaire; DAS-SC ¼ Dysfunctional Attitudes Scale-Self-Criticism; Distress ¼ Distress scale; Interference ¼ Interference scale; EDEQ-total ¼ EDE-Q total score; BDI-II ¼ Beck Depression Inventory-II; DASS-anx ¼ DASS-21anxiety subscale; FNE-B ¼ Brief Fear of Negative Evaluation Scale; PSWQ ¼ Penn State Worry Questionnaire; ASI-3 ¼ Anxiety Sensitivity Index-3; RSES ¼ Rosenberg Self-esteem Scale; Q-LES-Q18 ¼ Quality of Life Enjoyment and Satisfaction Questionnaire-18; RNTS ¼ Repetitive Negative Thinking Scale. 2

0.04 0.06 0.08 0.31 0.05 0.24 0.00 0.06 0.22 0.25 0.25 0.21 0.09 0.01 0.13 0.08 (1.38) (0.96) (0.62) (1.17) (3.60) (0.38) (0.41) (0.28) (0.72) (0.03) (4.07) (0.25) (0.42) (0.30) (0.16) (2.37) 33.42 28.34 a 15.90 a 29.57 a 70.32 a 6.50 a 6.19 a 2.21 a 16.19 b 9.54 a 49.44 a 64.59 b 24.32 a 25.73 a 3.25 a 30.20 a (1.36) (0.95) (0.61) (1.16) (3.55) (0.37) (0.40) (0.17) (1.40) (1.90) (4.07) (0.88) (1.78) (0.80) (0.16) (2.32) a

a

33.14 28.57 15.67 31.24 71.29 6.10 6.19 2.11 18.59 11.62 51.65 66.91 25.64 25.81 3.36 31.10 1.23 0.69 0.81 1.20 1.48 0.95 1.27 0.30 0.74 0.56 1.38 0.57 0.43 1.03 0.43 0.90 (1.40) (0.97) (0.64) (1.19) (3.65) (0.38) (0.41) (0.08) (0.82) (1.51) (4.03) (1.24) (5.75) (0.36) (0.16) (2.48) 25.66 25.64 b 13.38 b 26.04 b 50.18 b 4.98 b 4.65 b 1.23 b 10.40 b 7.33 b 40.18 b 59.95 b 22.37b 28.44 b 3.48 b 25.65 b (1.36) (0.95) (0.61) (1.16) (3.55) (0.37) (0.40) (0.14) (1.14) (1.01) (4.04) (2.69) (4.72) (0.44) (0.16) (2.32) a

a

33.14 28.86 15.62 32.10 68.43 6.62 6.91 1.56 17.45 12.10 50.41 65.10 27.49 25.05 3.20 34.33 .20 .12 .10 .11 .19 .24 .28 .13 .34 .03 .12 .03 .09 .15 .07 .05 F(1,77) ¼ 19.03 p < .001* F(1,77) ¼ 10.04 p ¼ .002* F(1,77) ¼ 8.42 p ¼ .005* F(1,77) ¼ 9.33 p ¼ .003* F(1,77) ¼ 18.40 p < .001* F(1,77) ¼ 23.89 p < .001* F(1,77) ¼ 30.31 p < .001* F(1,77) ¼ 11.66 p ¼ .001* F(1,73) ¼ 38.00 p < .001* F(1,77) ¼ 1.94 p ¼ .168 F(1,52) ¼ 7.12 p ¼ .010* F(1,77) ¼ 2.35 p ¼ .129 F(1,53) ¼ 5.03 p ¼ .029* F(1,77)¼13.46 p < .001* F(1,77) ¼ 5.70 p ¼ .019* F(1,76) ¼ 4.19 p ¼ .044* CM PS DA CPQ DAS-SC Distress Int EDE-Q-total BDI-II DASS-anx FNE-B PSWQ ASI-3 RSES Q-LES-Q-18 RNTS

F(1,77) F(1,77) F(1,77) F(1,77) F(1,77) F(1,77) F(1,77) F(1,77) F(1,73) F(1,77) F(1,52) F(1,77) F(1,53) F(1,77) F(1,77) F(1,76)

¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼

5.04 0.90 2.80 0.81 6.10 1.03 0.59 9.38 5.73 0.30 0.90 2.63 0.00 3.10 0.03 0.06

p p p p p p p p p p p p p p p p

¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼

.028* .347 .098 .372 .016 .314 .445 .003* .019* .585 .347 .109 .992 .082 .868 .813

F(1,77) ¼ 16.42 p < .001* F(1,77) ¼ 13.49 p < .001* F(1,77) ¼ 5.54 p ¼ .021 F(1,77) ¼ 28.83 p < .001* F(1,77) ¼ 22.73 p < .001* F(1,77) ¼ 8.78 p ¼ .004* F(1,77) ¼ 30.19 p < .001* F(1,77) ¼ 3.42 p ¼ .068 F(1,73)¼157.48 p < .001* F(1,77) ¼ 12.58 p ¼ .001* F(1,52) ¼ 17.13 p < .001* F(1,77) ¼ 16.36 p < .001* F(1,53) ¼ 14.35 p < .001* F(1,77) ¼ 12.32 p ¼ .001* F(1,77)¼1.19 p ¼ .279 F(1,76) ¼ 6.33 p ¼ .014*

Post-waitlist M (SE)

a a b

Control condition

Baseline M(SE) Cohen's d Post-treatment M(SE) Baseline M(SE)

Intervention condition Partial h2 Group*Time effect Time effect Group effect Measure

Table 3 Adjusted means and (standard errors) of outcome variables at baseline and post-treatment/post-waitlist for the intervention and control conditions.

a

Cohen's d

A.K. Handley et al. / Behaviour Research and Therapy 68 (2015) 37e47

43

Table 4 Bootstrapped standardised path estimates, standard errors, and p-values for the indirect pathways from the binary group variable through pre-post change in concern over mistakes to pre-post change in each of nine outcomes. Outcome

Bootstrapped standardised indirect path estimate

Standard error

Two-tailed p-values

BDI-II EDE-Q-tot FNE-B ASI-3 RNTS RSES Q-LES-Q-18

.205 .193 .409 .369 .245 .379 .259

.103 .099 .098 .100 .075 .081 .075

.046a .052 <.001b <.001b .001b <.001b .001b

Note. BDI-II ¼ Beck Depression Inventory-II; EDE-Q-total ¼ Eating Disorder Examination-Questionnaire (EDE-Q) total score; FNE-B ¼ Brief Fear of Negative Evaluation Scale; ASI-3 ¼ Anxiety Sensitivity Index-3; RNTS ¼ Repetitive Negative Thinking Scale; RSES ¼ Rosenberg Self-esteem Scale; Q-LES-Q-18 ¼ Quality of Life, Enjoyment and Satisfaction Questionnaire-18. a Significant at alpha ¼ .05. b Significant at alpha ¼ .007.

increases quality of life and that pre-post change in Concern over Mistakes mediates this effect. This is important as it demonstrates the multidimensional impact of CBT-P. The current study advances the literature as previous RCTs of CBT-P have not examined the impact of this treatment on social anxiety or rumination. The finding of group CBT-P reducing anxiety sensitivity is in accord with Radhu et al.'s (2012) finding. As prepost change in Concern over Mistakes mediated the impact of group CBT-P on social anxiety, rumination and anxiety sensitivity at the Bonferroni-adjusted alpha level of .007, these findings are consistent with perfectionism maintaining social anxiety (Egan et al., 2011) and suggest that perfectionism also maintains rumination and anxiety sensitivity. The findings of group CBT-P reducing depression are congruent with those of Riley et al. (2007) and concur with previous research highlighting the role of perfectionism in depression (Egan et al., 2011). Pre-post change in Concern over Mistakes mediated the impact of CBT-P on depression at the conventional alpha level; however, this was no longer significant at the Bonferroni-adjusted alpha level of .007. This finding suggests that future examination of mediation of pre-post changes in depression is required in a larger sample. The current study also adds to the literature by demonstrating the benefit of group CBT-P in reducing eating disorder pathology and increasing self-esteem. Previous RCTs of CBT-P have not found greater improvements in eating disorder or self-esteem outcomes compared to a comparison or control group (Steele & Wade, 2008). In Steele and Wade's (2008) study, individual CBT-P was comparable to other treatments in the trial in reducing eating disorder symptoms and increasing self-esteem. In the present study, prepost change in Concern over Mistakes tended toward mediating the impact of group CBT-P on eating pathology; however, this was of borderline significance, likely due to a Type II error. This finding is still consistent with perfectionism maintaining eating disorder pathology (Egan et al., 2011) but requires further investigation in a larger sample. The finding that pre-post change in Concern over Mistakes significantly mediated the impact of CBT-P on self-esteem is consistent with perfectionism maintaining low self-esteem (Shafran et al., 2002). While there was a tendency for the treatment group to have greater reductions in DASS-anxiety than the control group, the Group  Time interaction effects were not significant. Previous RCTs have also found non-significant Group  Time interaction effects for anxiety (Riley et al., 2007; Steele & Wade, 2008). This finding is interesting given the large significant interaction effect for social anxiety and the moderate significant interaction effect for anxiety

44

Table 5 Table of effects, adjusted means (standard errors) and effect sizes for all participants at their pre-treatment, post-treatment, 3-month and 6-month follow-ups. Measure

Group effect

Group*Time effect

Partial Pre-tx mean (SE) h2

F(3,139) ¼ 25.27 p < .001* F(3,139) ¼ 14.22 p < .001* F(3,139) ¼ 17.80 p < .001* F(3,139) ¼ 20.52 p < .001* F(3,139) ¼ 32.75 p < .001* F(3,139) ¼ 32.55 p < .001* F(3,139) ¼ 31.48 p < .001* F(3,139) ¼ 14.51 p < .001* F(3,139) ¼ 8.12 p < .001* F(1,93) ¼ 11.17 p < .001* F (3,97) ¼ 9.43 p < .001* F(3,139)¼20.57 p < .001* F(3,139) ¼ 12.25 p < .001* F(3,139) ¼ 6.55 p < .001*

F(3,139)¼0.27 p ¼ .844 F(3,139)¼0.33 p ¼ .807 F(3,139) ¼ 0.68 p ¼ .566 F(3,139)¼0.95 p ¼ .420 F(3,139)¼0.79 p ¼ .500 F(3,139) ¼ 1.22 p ¼ .304 F(3,139) ¼ 2.47 p ¼ .064 F(3,139) ¼ 2.38 p ¼ .072 F(3,139)¼1.27 p ¼ .288 F(3,93) ¼ 0.67 p ¼ .571 F (3,97) ¼ 0.72 p ¼ .540 F(3,139)¼2.53 p ¼ .060 F(3,139) ¼ 0.21 p ¼ .890 F(3,139)¼3.82 p ¼ .012*

.01 .01 .01 .02 .02 .03 .05 .05 .03 .02 .02 .05 .00 .08

Postetx mean (SE)

3mo Mean (SE)

6mo Mean (SE)

Pre-post

t(139) ¼ 6.78 p < .001* d ¼ 1.13 28.70 (0.71) 25.71 (0.74) 25.51 (0.74) 25.67 (0.75) t(139) ¼ 5.17 p < .001* d ¼ 0.66 15.82 (0.51) 13.75 (0.53) 12.90 (0.54) 12.57 (0.55) t(139) ¼ 4.43 p < .001* d ¼ 0.76 30.82 (0.85) 25.97 (0.89) 24.98 (0.89) 24.93 (0.91) t(139) ¼ 5.53 p < .001* d ¼ 0.99 70.11 (2.77) 53.28 (2.85) 53.28 (2.85) 51.72 (2.89) t(139) ¼ 7.81 p < .001* d ¼ 1.07 6.56 (0.25) 4.98 (0.27) 4.17 (0.27) 4.37 (0.27) t(139) ¼ 5.88 p < .001* d ¼ 0.98 6.55 (0.28) 4.91 (0.30) 3.78 (0.30) 4.13 (0.31) t(139) ¼ 5.28 p < .001* d ¼ 0.89 16.81 (1.70) 10.46 (1.59) 8.32 (1.29) 8.73 (1.40) t(136) ¼ 3.70 p < .001* d ¼ 0.63 1.93 (0.22) 1.57 (0.22) 1.31 (0.21) 1.18 (0.16) t(139) ¼ 2.81 p ¼ .006* d ¼ 0.25 49.41 (2.55) 41.82 (2.60) 40.75 (2.62) 41.08 (2.62) t(93) ¼ 4.36 p < .001* d ¼ 0.84 24.90 (2.39) 17.91 (2.12) 16.55 (1.61) 15.20 (1.68) t(97)¼3.89 p < .001* d ¼ 0.55 25.42 (0.61) 27.48 (0.62) 28.82 (0.69) 29.44 (0.78) t(139) ¼ -3.38 p ¼ .001* d ¼ 0.52 3.23 (0.13) 3.53 (0.13) 3.73 (0.13) 3.70 (0.14) t(139) ¼ -3.17 p ¼ .002* d ¼ 0.40 Tx: 34.33 (2.38) Tx: 25.85 (2.15) Tx: 28.38 (2.51) Tx: 25.22 (2.47) Tx: t(137) ¼ 3.39 TC: 30.15 (2.43) TC: 31.16 (2.59) TC: 22.40 (2.59) TC: 25.60 (2.70) p ¼ .001* d ¼ 0.88 TC: t(137) ¼ -0.39 p ¼ .696 d ¼ 0.08 33.51 (1.23)

26.42 (1.27)

26.47 (1.27)

25.54 (1.29)

Pre-3mo

Pre-6mo

t(139) ¼ 6.73 p < .001* d ¼ 1.08 t(139) ¼ 5.51 p < .001* d ¼ 0.67 t(139) ¼ 5.98 p < .001* d ¼ 0.97 t(139) ¼ 6.65 p < .001* d ¼ 1.07 t(139) ¼ 7.81 p < .001* d ¼ 1.02 t(139) ¼ 8.89 p < .001* d ¼ 1.55 t(139) ¼ 8.90 p < .001* d ¼ 1.56 t(136) ¼ 5.93 p < .001* d ¼ 0.88 t(139) ¼ 3.86 p < .001* d ¼ 0.44 t(93) ¼ 4.89 p < .001* d ¼ 0.83 t(97)¼4.16 p < .001* d ¼ 0.72 t(139)¼-6.88 p < .001* d ¼ 0.82 t(139)¼-5.43 p < .001* d ¼ 0.68 Tx: t(137) ¼ 2.38 p ¼ .019* d ¼ 0.58 TC: t(137) ¼ 3.01 p ¼ .003* d ¼ 0.73

t(139) ¼ 7.45 p < .001* d ¼ 1.13 t(139) ¼ 5 .11 p < .001* d ¼ 0.66 t(139) ¼ 6.51 p < .001* d ¼ 1.10 t(139) ¼ 6.56 p < .001* d ¼ 1.02 t(139) ¼ 8.34 p < .001* d ¼ 1.07 t(139) ¼ 7.97 p < .001* d ¼ 1.32 t(139) ¼ 7.63 p < .001* d ¼ 1.31 t(136) ¼ 5.80 p < .001* d ¼ 0.85 t(139) ¼ 4.93 p < .001* d ¼ 0.62 t(93) ¼ 4.70 p < .001* d ¼ 0.94 t(97)¼5.23 p < .001* d ¼ 0.82 t(139)¼-6.86 p < .001* d ¼ 0.92 t(139)¼4.96 p < .001* d ¼ 0.58 Tx: t(137) ¼ 3.71 p < .001* d ¼ 0.91 TC: t(137) ¼ 1.69 p ¼ .093 d ¼ 0.35

* Denotes a significant effect. Note. Partial h2 ¼ partial eta squared; Tx ¼ treatment condition; TC: treated control; CM ¼ Concern over Mistakes; PS ¼ Personal Standards; DA ¼ Doubts about Actions; CPQ ¼ Clinical Perfectionism Questionnaire; DASSC ¼ Dysfunctional Attitudes Scale-Self-Criticism; Distress ¼ Distress scale; Interference ¼ Interference scale; EDE-Q-total ¼ Eating Disorder Examination-Questionnaire (EDE-Q) total score; BDI-II ¼ Beck Depression InventoryII; FNE-B ¼ Brief Fear of Negative Evaluation Scale; ASI-3 ¼ Anxiety Sensitivity Index-3; RSES ¼ Rosenberg Self-esteem Scale; Q-LES- Q- 18 ¼ Quality of Life, Enjoyment and Satisfaction Questionnaire-18; RNTS ¼ Repetitive Negative Thinking Scale.

A.K. Handley et al. / Behaviour Research and Therapy 68 (2015) 37e47

F(1,139) ¼ 0.07 p ¼ .798 PS F(1,139) ¼ 0.02 p ¼ .901 DA F(1,139) ¼ 0.04 p ¼ .835 CPQ F(1,139) ¼ 1.19 p ¼ .278 DAS-SC F(1,139) ¼ 0.14 p ¼ .713 Dis F(1,139) ¼ 0.84 p ¼ .362 Int F(1,139) ¼ 0.67 p ¼ .413 BDI-II F(1,139) ¼ 0.30 p ¼ .583 EDE-Q-tot F(1,139) ¼ 2.37 p ¼ .126 FNE-B F(1,93) ¼ 0.00 p ¼ .988 ASI-3 F (1,97) ¼ 0.92 p ¼ .340 RSES F(1,139) ¼ 0.00 p ¼ .986 Q-LES-Q-18 F(1,139) ¼ 0.33 p ¼ .569 RNTS F(1,139) ¼ 0.16 p ¼ .688 CM

Time effect

A.K. Handley et al. / Behaviour Research and Therapy 68 (2015) 37e47

45

Table 6 The number of participants in the treatment and control conditions who experienced reliable change on outcome variables between baseline and post-treatment/post-waitlist, and the number of participants in the treated control condition who experienced reliable pre-post change. Measure

CM PS DA CPQ DAS-SC BDI-II EDE-Qtotal FNE-Bb ASI-3b RNTSb Q-LES-Qc RSESc

Treatment (n ¼ 19)

Control (n ¼ 20)

Y n, %

[ n, %

Yn, %

[ n, %

0 0 0 0 0 1 2 0 0 2 5 8

1 3 1 3 2 3 1 5 3 4 2 1

1 0 3 0 1 2 0 3 2 4 0 0

12 7 6 7 12 6 5 9 2 6 1 1

(63%)a (37%) (32%) (37%) (63%) (35%) (26%) (69%) (17%) (33%) (5%) (5%)

(0%) (0%) (0%) (0%) (0%) (6%) (11%) (0)% (0%) (11%) (26%) (42%)

(5%) (15%) (5%) (15%) (10%) (17%) (5%) (38%) (20%) (20%) (10%) (5%)

(5%) (0%) (15%) (0%) (5%) (11%) (0%) (23%) (13%) (20%) (0%) (0%)

Fisher's exact test (1-sided)

p p p p p p p p p p p p

< .001* ¼ .116 ¼ .044* ¼ .155 ¼ .001* ¼ .192 ¼ .080 ¼ .119 ¼ .612 ¼ .468 ¼ .020* ¼ .001*

Treated Control(n ¼ 17) Yn, %

[ n, %

9 7 6 4 11 5 4 4 3 3 2 1

0 1 1 1 1 1 0 0 0 2 4 2

(53%) (41%) (35%) (24%) (65%) (29%) (24%) (36%) (23%) (18%) (12%) (6%)

(0%) (6%) (6%) (6%) (6%) (6%) (0%) (0%) (0%) (12%) (24%) (12%)

Fisher's exact test (1-sided)

p p p p p p p p p p p p

¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼ ¼

.736 1.00 1.00 .481 1.00 1.00 .577 .115 .541 .443 1.00 .065

* Denotes a significant effect. Note. Y n, % ¼ the number and percentage of participants experiencing a reliable decrease on an outcome variable; [ n, % ¼ the number and percentage of participants experiencing a reliable increase on an outcome variable; CM ¼ Concern over Mistakes subscale; PS ¼ Personal Standards subscale; DA ¼ Doubts about Actions subscale; CPQ ¼ Clinical Perfectionism Questionnaire; DAS-SC ¼ Dysfunctional Attitudes Scale- Self-Criticism; BDI-II ¼ Beck Depression Inventory-II; EDE-Q-total ¼ EDE-Q total score; FNE-B ¼ Brief Fear of Negative Evaluation Scale; ASI-3 ¼ Anxiety Sensitivity Index-3; RNTS ¼ Repetitive Negative Thinking Scale; Q-LES-Q ¼ Quality of Life, Enjoyment and Satisfaction Questionnaire-18; RSES ¼ Rosenberg Self-esteem Scale. a Percentages rounded. b FNE-B: n ¼ 13 in treatment condition, n ¼ 13 in control condition; n ¼ 11 in treated control condition; n ¼ 24 total who received treatment; ASI-3: n ¼ 12 in treatment condition, n ¼ 15 in control condition; n ¼ 13 in treated control condition; n ¼ 25 total who received treatment; RNTS: n ¼ 18 in treatment condition, n ¼ 20 in control condition; n ¼ 17 in treated control condition; n ¼ 35 total who received treatment. c An increase in score suggests an improvement.

sensitivity. It is possible that CBT-P produces smaller effects for the type of anxiety measured by the DASS-anxiety scale relative to social anxiety and anxiety sensitivity, and that the current and previous RCTs had insufficient power to detect these effects. Although further research with larger samples is needed to ascertain the impact of CBT-P on different aspects of anxiety, the current results suggest that group CBT-P does reduce social anxiety and anxiety sensitivity. The reliable change demonstrated in perfectionism (DAS-SC), self-esteem and quality of life contribute to the literature as previous RCTs of CBT-P have not examined reliable change in these constructs (Pleva & Wade, 2006; Riley et al., 2007; Steele & Wade, 2008). Group CBT-P also produced clinically significant change in perfectionism as measured by Concern over Mistakes. This adds to the literature as previous RCTs of perfectionism treatments did not calculate clinically significant change in a measure of multidimensional perfectionism (Pleva & Wade, 2006; Riley et al., 2007; Steele & Wade, 2008). A significantly greater proportion of individuals in the group CBT-P condition showed a post-treatment recovery from a diagnosis of depression relative to the control condition. While there were non-significant differences in recovery from eating and anxiety disorders, this was likely an artefact of the diagnostic criteria for eating disorders and generalised anxiety disorder, which makes it difficult for participants to no longer meet a diagnosis after eight weeks (DSM-IV-TR, APA, 2000). The non-significant differences may have also been due to the small number of participants with social phobia, obsessive-compulsive disorder and panic disorder per condition, which may have reduced the power for detecting these effects. Nevertheless, once the entire sample received treatment, high percentages of individuals recovered from their disorders at 6-month follow-up, which provided indirect support for participants no longer meeting anxiety, mood and eating disorder diagnoses after receiving group CBT-P. Four limitations require discussion. First, this study utilised a design where participants who had been in the control condition were given group CBT-P after eight weeks. While this was important for ethical reasons, it prevented group CBT-P from being compared to a control condition at 6-month follow-up. The second limitation

was the preponderance of individuals with a principal diagnosis of generalised anxiety disorder and the lower numbers of individuals with diagnoses of social phobia, obsessive-compulsive disorder and panic disorder. Future studies should use samples with a greater representation of psychological disorders to ensure that findings can be generalised. The third limitation arose from the absence of established population norms for perfectionism variables, which meant that a very conservative clinical significance cut-off was utilised (Jacobson & Truax, 1991; Riley et al., 2007). Fourth, the primary purpose of this study was to examine the efficacy of group CBT-P as opposed to testing a mediational model. Thus, while the sample size allowed for sufficient power to detect moderate to large Group  Time interaction effects (Faul et al., 2007), the study was underpowered for the mediational analyses. This likely explains why pre-post change in Concern over Mistakes did not significantly mediate the impact of CBT-P on depression at the Bonferroniadjusted alpha level, and why pre-post change in Concern over Mistakes did not significantly mediate the impact of CBT-P on eating pathology at the conventional alpha level. Future studies need to conduct mediational analyses using larger sample sizes to ensure that there is sufficient power to test mediational models. Collectively, these findings contribute toward establishing the efficacy of group CBT-P. Given that the intervention can decrease perfectionism and related psychopathology, as well as increase selfesteem and quality of life, this has important clinical implications. In view of the high rates of diagnostic co-morbidity (Kessler, Chiu, Demler, & Walters, 2005), implementing this transdiagnostic intervention rather than sequentially focussing on each disorder may create shorter time in therapy and lower cost for the client (Craske, 2012; Egan, Wade, & Shafran, 2012). The group nature of the intervention creates additional time efficiency for the therapist and cost reduction for the client (Himle et al., 2003).

Acknowledgements The individuals who contributed to this manuscript all met the criteria for authorship. All authors acknowledge the School of Psychology and Speech Pathology at Curtin University, Perth,

46

A.K. Handley et al. / Behaviour Research and Therapy 68 (2015) 37e47

Australia who provided the funding for the data collection materials. References American Psychiatric Association. (2000). Diagnostic and statistical manual of mental disorders. text rev (4th ed.). Washington DC: American Psychiatric Association. Anderson, R. A., & Rees, C. S. (2007). Group versus individual cognitive behavioural therapy for obsessive-compulsive disorder: a randomised controlled trial. Behaviour Research and Therapy, 45, 123e137. http://dx.doi.org/10.1016/j.brat. 2006.01.016. Antony, M., Bieling, P., Cox, B., Enns, M., & Swinson, R. (1998). Psychometric properties of the 42-item and 21-item versions of the Depression, Anxiety and Stress Scales in clinical groups and a community sample. Psychological Assessment, 10(2), 176e181. http://dx.doi.org/10.1037//1040-3590.10.2.176. Antony, M., Purden, C., Huta, V., & Swinson, R. (1998). Dimensions of perfectionism across the anxiety disorders. Behaviour Research and Therapy, 36, 1143e1154. http://dx.doi.org/10.1016/S0005-7967(98)00083-7. Beck, A., Steer, R., & Brown, M. (1996). Beck depression inventory manual (2nd ed.). San Antonio, Texas: Psychological Corporation. Bieling, P. J., McCabe, R. E., & Antony, M. M. (2013). Cognitive behavioural therapy in groups. New York: The Guilford Press. Bieling, P., Summerfeldt, L., Israeli, A., & Antony, M. (2004). Perfectionism as an explanatory construct in co-morbidity of axis 1 disorders. Journal of Psychopathology and Behavioural Assessment, 26(3), 193e201. http://dx.doi.org/10.1023/ B:JOBA.0000022112.27186.98. Blatt, S., Quinlan, D., Pilkonis, P., & Shea, M. (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, 125e132. http://dx. doi.org/10.1037//0022-006X.63.1.125. Brown, T., DiNardo, P., & Barlow, D. (1994). Anxiety disorders interview schedule for DSM-IV-Adult version- client interview schedule. New York: Oxford University Press. Cohen, J. (1988). Statistical power for the behavioural sciences (2nd ed.). Hillsdale, NJ: Erlbaum. Collins, K., Westra, H., Dozois, D., & Stewart, S. (2005). The validity of the brief version of the Fear of Negative Evaluation Scale. Journal of Anxiety Disorders, 19, 345e359. http://dx.doi.org/10.1016/S0887-6185(04)00022e2. Craske, M. G. (2012). The cutting edge: transdiagnostic treatment for anxiety and depression. Depression and Anxiety, 29, 749e753. http://dx.doi.org/10.1002/da. 21992. Dunkley, D., Sanislow, C., Grilo, C., & McGlashan, T. (2006). Perfectionism and depressive symptoms three years later: negative social interactions, avoidant coping and perceived social support as mediators. Comprehensive Psychiatry, 47, 106e115. http://dx.doi.org/10.1016/j.comppsych.2005.06.003. Egan, S. J., Shafran, R., Lee, M., Fairburn, C. G., Cooper, Z., Doll, H. A., et al. (2015). The reliability and validity of the clinical perfectionism questionnaire in eating disorder and community samples. Manuscript under review. Egan, S., & Stout, S. (2007). Group cognitive behavioural therapy for clinical perfectionism: A preliminary investigation. Poster Presented at the American Association for Behavioural and Cognitive Therapy 41st Annual Convention, Philadelphia, USA. Egan, S., Wade, T., & Shafran, R. (2011). Perfectionism as a transdiagnostic process: a clinical review. Clinical Psychology Review, 31(2), 203e212. http://dx.doi.org/10. 1016/j.cpr.2010.04.009. Egan, S. J., Wade, T., & Shafran, R. (2012). The transdiagnostic process of perfectionism. Journal of Psychopathology and Clinical Psychology, 17(3), 279e294. Evans, M. D., Piasecki, J. M., Kris, M. R., & Hollon, S. D. (1984). Raters' manual for the collaborative study Psychotherapy Rating Scale-Form 6. Minneapolis, U.S.A: University of Minnesota and the St Paul Ramsey Medical Center. Fairburn, C. G., & Beglin, S. J. (1994). The assessment of eating disorders: interview or self-report questionnaire? International Journal of Eating Disorders, 16, 363e370. http://dx.doi.org/10.1002/1098-108X(199412). Fairburn, C. G., Cooper, Z., & Shafran, R. (2003). Clinical perfectionism questionnaire. Unpublished scale. Oxford, UK: University of Oxford. Faul, F., Erdfelder, E., Lang, A., & Buchner, A. (2007). G*Power 3: a flexible statistical power analysis program for the social, behavioural, and biomedical sciences. Behaviour Research Methods, 35, 175. http://dx.doi.org/10.3758/BF03193146. Frost, R., Marten, P., Lahart, C., & Rosenblate, R. (1990). The dimensions of perfectionism. Cognitive Therapy and Research, 14, 449e468. http://dx.doi.org/10.1007/ BF01172967. Gray-little, B., Williams, V., & Hancock, T. (1997). Item-response theory analysis of the Rosenberg Self-Esteem Scale. Personality and Social Psychology Bulletin, 23, 443e451. http://dx.doi.org/10.1177/0146167297235001. 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, 456e470. http://dx.doi.org/10.1037//00223514.60.3.456. Himle, J. A., Van Etten, M., & Fischer, D. J. (2003). Group cognitive behavioural therapy for obsessiveecompulsive disorder (OCD): a controlled study. International Journal of Psychiatry in Clinical Practice, 9, 257e263. http://dx.doi.org/10. 1002/da.10088.

Holden, J. E., Kelley, K., & Agarwal, R. (2008). Analysing change: a primer on multilevel models with applications to nephrology. American Journal of Nephrology, 28(5), 792e801. http://dx.doi.org/10.1159/000131102. Imber, S. D., Pilkonis, P. A., Sotsky, S. M., Elkin, I., Watkins, J. T., Collins, J. F., et al. (1990). Mode-specific effects among three treatments for depression. Journal of Consulting and Clinical Psychology, 58(3), 352e359. http://dx.doi.org/10.1037// 0022-006X.58.3.352. 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(1), 12e19. http://dx.doi.org/10.1037//0022-006X.59. 1.12. Keppel, G., & Wickens, T. D. (2004). Design and analysis: A researcher's handbook (4th ed.). U.S.A: Pearson Higher Education. Kessler, R., Chiu, W., Demler, O., & Walters, E. (2005). Prevalence, severity, and comorbidity of 12-month DSM-IV disorders in the national co-morbidity survey replication. Archives of General Psychiatry, 62, 617e627. http://dx.doi.org/10. 1001/archpsyc.62.6.617. Lloyd, S., Schmidt, U., Khondoker, M., & Tchanturia, K. (2014). Can psychological interventions reduce perfectionism? A systematic review and meta-analysis. Behavioural and Cognitive Psychotherapy. Advance online publication. http://dx. doi.org/10.1017/S1352465814000162 Lovibond, S., & Lovibond, P. (1995). Manual for the Depression, Anxiety and Stress Scales (2nd ed.). Sydney: Psychology Foundation. McEvoy, P. M., Mahoney, A. E., & Moulds, M. L. (2010). Are worry, rumination and post-event processing one and the same? Development of the repetitive thinking questionnaire. Journal of Anxiety Disorders, 24, 509e519. http://dx.doi. org/10.1016/j.janxdis.2010.03.008. Meyer, T., Miller, M., Metzger, R., & Borkovec, T. (1990). Development and validation of the Penn State worry questionnaire. Behaviour Research and Therapy, 28(6), 487e495. http://dx.doi.org/10.1016/0005-7967(90)90135-6. Moher, D., Hopewell, S., Schulz, K. F., Montori, V., Gotzsche, P. C., Devereaux, P. J., et al. (2010). CONSORT 2010 explanation and elaboration: updated guidelines for reporting parallel group randomised trials. British Medical Journal, 340, c869. http://dx.doi.org/10.1136/bmj.c869. Mond, J., Hay, P., Rodgers, B., Owen, C., & Beumont, P. (2004a). Temporal stability of the eating disorder examination questionnaire. International Journal of Eating Disorders, 36(2), 195e203. http://dx.doi.org/10.1002/eat.20017. Mond, J., Hay, P., Rodgers, B., Owen, C., & Beumont, P. (2004b). Validity of the Eating Disorder Examination Questionnaire (EDE-Q) in screening for eating disorders in community samples. Behaviour Research and Therapy, 42, 551e567. http://dx. doi.org/10.1016/S0005-7967(03)00161-X. n, L. K., & Muthe n, B. O. (1998 e 2007). Mplus version 5.2. Los Angeles, CA: Muthe n & Muthe n. Muthe Pleva, J., & Wade, T. (2006). Guided self-help versus pure self-help for perfectionism: a randomised controlled trial. Behaviour Research and Therapy, 45, 849e861. http://dx.doi.org/10.1016/j.brat.2006.08.009. Radhu, N., Daskalakis, Z. J., Guglietti, C. L., Farzan, F., Barr, M. S., Arpin-Cribbie, C. A., et al. (2012). Cognitive behavioural therapy-related increases in cortical inhibition in problematic perfectionists. Brain Stimulation, 5, 44e54. http://dx.doi. org/10.1016/j.brs.2011.01.006. Richardson, J. T. (2011). Eta squared and partial eta squared as measures of effect size in educational research. Educational Research Review, 6, 135e147. http://dx. doi.org/10.1016/j.edurev.2010.12.001. Riley, C., Lee, M., Cooper, Z., Fairburn, C., & Shafran, R. (2007). A randomised controlled trial of cognitive-behaviour therapy for clinical perfectionism: a preliminary study. Behaviour Research and Therapy, 45, 2221e2231. http://dx. doi.org/10.1016/j.brat.2006.12.003. Ritsner, M., Kurs, R., Gibel, A., Ratner, Y., & Endicott, J. (2005). Validity of an abbreviated quality of life, enjoyment and satisfaction questionnaire (QLES-Q-18) for schizophrenia, schizoaffective and mood disorder patients. Quality of Life Research, 14, 1693e1703. http://dx.doi.org/10.1007/s11136005-2816-9. Rosenberg, M. (1965). Society and the self-image. Princeton, NJ: Princeton University Press. Saghaei, M. (2004). Random allocation software for parallel group randomized trials. Biomed Central Medical Research Methodology, 4(1), 26. http://dx.doi.org/ 10.1186/1471-2288-4-26. Segal, S., Williams, M., & Teasdale, J. (2002). Mindfulness-based cognitive therapy for depression: a new approach to preventing relapse. New York: Guildford Press. Shafran, R., Cooper, Z., & Fairburn, C. (2002). Clinical perfectionism: a cognitive behavioural analysis. Behaviour Research and Therapy, 40, 773e791. http://dx. doi.org/10.1016/S0005-7967(01)00059-6. Shafran, R., Egan, S., & Wade, T. (2010). Overcoming perfectionism: A self-help manual using cognitive-behavioural techniques. London: Robinson. Sheehan, D. V., Janavs, J., Baker, R., Sheehan, K. H., Knapp, E., & Sheehan, M. (1998). Mini international neuropsychiatric interview 5.0.0. Tampa: University of South Florida. Sheehan, D. V., Lecrubier, Y., Sheehan, K. H., Janavs, J., Weiller, E., Keskiner, A., et al. (1997). The validity of the Mini International Neuropsychiatric Interview (MINI) according to the SCID-P and its reliability. European Psychiatry, 12(5), 232e241. http://dx.doi.org/10.1016/S0924-9338(97)83297-X. Steele, A., O'Shea, A., Murdock, A., Karney, A., & Wade, T. (2011). Perfectionism and its relation to overvaluation of weight and shape and depression in an eating disorder sample. International Journal of Eating Disorders, 44(5), 459e464. http://dx.doi.org/10.1002/eat.20817.

A.K. Handley et al. / Behaviour Research and Therapy 68 (2015) 37e47 Steele, A., & Wade, T. (2008). A randomised trial investigating guided self-help to reduce perfectionism and its impact on bulimia nervosa: a pilot study. Behaviour Research and Therapy, 46, 1316e1323. http://dx.doi.org/10.1016/j.brat.2008. 09.006. Steele, A., Waite, S., Egan, S., Finnigan, J., Handley, A., & Wade, T. (2013). Psychoeducation and group cognitive-behavioural therapy for clinical perfectionism: a case series evaluation. Behavioural and Cognitive Psychotherapy, 41(2), 129e143. http://dx.doi.org/10.1017/S1352465812000628. Taylor, S., Zvolensky, M., Cox, B., Deacon, B., Heimberg, R., Ledley, B., et al. (2007). Robust dimensions of anxiety sensitivity: development and initial validation of the Anxiety Sensitivity Index-3. Psychological Assessment, 19(2), 176e188. http:// dx.doi.org/10.1037/1040-3590.19.2.176.

47

Thiels, C., Schmidt, U., Troop, N., Treasure, J., & Garthe, R. (2001). Compliance with a self-care manual in guided self-change for bulimia nervosa. European Eating Disorders Review, 9, 115e122. http://dx.doi.org/10.1002/erv.370. Troop, N. A., Schmidt, U. H., Tiller, J. M., Todd, G., Keilen, M., & Treasure, J. L. (1996). Compliance with a self-directed treatment manual for bulimia nervosa: predictors and outcome. British Journal of Clinical Psychology, 35, 435e438. http:// dx.doi.org/10.1111/j.2044-8260.1996.tb01197.x. Weissman, A., & Beck, A. (1978). Development and validation of the Dysfunctional Attitudes Scale: A preliminary investigation. Paper presented at the Annual Meeting of the American Research association, Ontario, Canada, March 27e31, 1978.