A Randomised Controlled CBT Intervention for Maladaptive Perfectionism: Outcome and Predictors

: Maladaptive perfectionism has been associated with certain mental health problems. Moreover, studies suggest that the development of perfectionism can be attributed to childhood experiences and, more speciﬁcally, to parenting styles. (1) Background: The aims of the present study were ﬁrst to examine the relationship of perfectionism to perceived parenting and current symptoms of depression and anxiety and, secondly, to study the effectiveness of a group CBT intervention program for high perfectionism; (2) Methods: Participants were 81 young Greek adults with relatively high scores in perfectionism. They were randomly allocated to two conditions: either the intervention group (IG; n = 40) or a non-active control group (CG; n = 41). The intervention group received a CBT intervention of 10 weekly sessions, while assessments were made in two time points for both groups; (3) Results: There was signiﬁcant correlation of perfectionism with perceived parental indifference, abuse, and overcontrol as well as measures of anxiety and depression. Results showed signiﬁcant decrease in perfectionism and other symptoms in the IG. Although perceived paternal indifference and abuse could predict perfectionism at pre-intervention, no pre-intervention variable could predict the perfectionism intervention outcome; (4) Conclusions: Overall, the study suggests that maladaptive perfectionism is associated with perceived negative parenting and current anxiety and depression and that a group CBT intervention can address speciﬁc dimensions of maladaptive perfectionism.


Introduction
There is a consensus that perfectionism is best understood as a multi-dimensional construct [1][2][3][4] and that the core aspects of maladaptive perfectionism generally involve high standards or striving and excessive concerns about those standards [1,5]. Perfectionist strivings refer to setting high standards for oneself as well as one's tendency to search for perfection. Furthermore, perfectionism concerns refer to the fear of making mistakes, doubts about actions, fear of being negatively evaluated by others, and discrepancy between expectations and self-performance [6].
The importance of family factors in the development of maladaptive perfectionism has been supported in various models about perfectionism [7]. Parents may have an important role in the development of perfectionism in various ways, such as psychopathology symptoms [8] and parenting style [7,[9][10][11][12][13].
There is a variety of theoretical perspectives and methodological approaches used to identify the major dimensions and characteristics of parental styles. The most empirical work in the area of parenting was done by Baumrind [14,15] and was extended by Maccoby [16]. According to this approach, there are four conceptual types of parenting (authoritative, authoritarian, permissive, neglectful) that vary according to two dimensions: demandingness (standards and demands set by parents) and responsiveness (parents' responses to children) [14][15][16]. An authoritative style involves high levels of demandingness and responsiveness. Authoritative parents try to direct their children, make reasonable Similar results were also reported by Steele et al. [53], who studied the effectiveness of group therapy of perfectionism and psychoeducation based, again, on the same Shafran et al. [54] protocol. Steele's results showed a statistically significant improvement in perfectionism, depression, and stress at the end of the trial, which was also evident at the three-month follow-up [53]. Additionally, a review of CBT-based interventions (individual, guided self-help, web-based, and group) showed that perfectionism may significantly reduce in individuals with clinical disorders and that such interventions are associated with decreases in anxiety, depression, eating disorder, and obsessive-compulsive symptoms [57]. A brief (five weekly sessions) cognitive-behavioural group intervention for maladaptive perfectionism in Argentinian students showed that most participants had statistically and clinically meaningful reductions in perfectionism and other distressing symptoms, such as anxiety and depressive symptoms, at the end of the intervention [58]. These studies provide convincing evidence for the efficacy of cognitive-behavioural therapies in addressing maladaptive perfectionism in both clinical and non-clinical samples.
Although relevant studies have shown promising results on the efficacy of CBTmanualized programs addressing perfectionism, only a few studies examined the association of the trait of perfectionism in adults with perceived negative parenting styles in childhood and early adolescence. Consequently, the present study aims to examine the above association in young adults as well as the association of perfectionism with symptoms of anxiety and depression. Having studied the above relationships, a resulting important goal of this study is to examine the effectiveness of a group intervention program for maladaptive perfectionism and coexisting symptomatology. The final and last aim of the present study is to look into the influence of the pre-intervention variables (e.g., parenting styles, anxiety, and depression) to the post-intervention outcome.

Materials and Methods
A cross-sectional, descriptive-correlational design was performed at the initial phase of the study that was followed by a parallel group trial design where a perfectionismfocused CBT group intervention was compared with a no intervention concurrent control. The allocation ratio was 1:1.

Participants
Participants were recruited through posters displayed in two universities of the city of Thessaloniki, Greece, and also through posts in social media pages of various departments of both universities. Two-hundred and fifteen participants responded and completed the questionnaires (Figure 1). Inclusion criterion was the degree of perfectionism as it is reflected by the total score in the Frost Multi-Dimensional Perfection Scale (without the FMPS subscale of Organisation). Mean value of the FMPS plus one standard deviation was the cut-off point [59][60][61][62]. Students with FMPS scores higher that the cut-off point were considered for participation in the study. Appropriate candidates that were invited for the study were 135, and 60% of them (n = 81) consented to participate in the second phase of the study. Mean age was 21.95 (SD= 3.62, range . Participants were randomly allocated to the intervention group (IG; n = 40) and the control group (CG; n = 41). Thirtyfour participants completed the second phase of the study. Intervention completers were 19 participants (M age = 22.47 SD = 2.55, 73.7% females). Control group completers consisted of 15 participants (M age = 19.80 SD = 0.86, 86.7% females).
Participation in the study was on a voluntary basis, and no money or credits were given to participants. The research was approved by the ethics committee of the university. Psychiatry Int. 2021, 2, FOR PEER REVIEW 4 Figure 1. Flow chart of participants through the study.

Measures
Frost Multi-Dimensional Perfection Scale (FMPS) [1]: Frost et al. [1] hypothesised that perfectionism is based on six dimensions. In their model, the first dimension was concern over mistakes (CM). The second dimension was (Personal Standards, PS) setting high and relatively inflexible standards for personal performance since perfectionists tend to set high goals that are almost impossible to meet adequately. The third (Parental Expectations, PE) and fourth (Parental Criticism, PC) dimensions were the perception of the degree to which individuals perceive their parents as exceptionally demanding or critical. The fifth dimension (Doubts about Action, D) of perfectionism was the tendency to question the quality of one's performance, emphasizing the overall feeling that it is not satisfactory. Finally, the sixth dimension (Organisation, O) concerns Organisation, as people with high perfectionism seem to accentuate order and accuracy [1]. The first five subscales represent the core dimensions of the FMPS, so the Organisation subscale is excluded when calculating the FMPS total score. The scale consists of 35 items that assess perfectionism in the aforementioned six dimensions. High scores on the Organization subscale are not intrinsically problematic but, combined with high scores on the other factors, may exacerbate dysfunction. The questionnaire is scored in 5-point Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree). Good internal consistency and convergent validity with other measures of perfectionism have been demonstrated within community and clinical samples of adults [1,63]. In the present study, the Cronbach's alpha for the total scale was 0.89, and Cronbach's alpha for the subscales ranged from 0.76 to 0.88.

Measures
Frost Multi-Dimensional Perfection Scale (FMPS) [1]: Frost et al. [1] hypothesised that perfectionism is based on six dimensions. In their model, the first dimension was concern over mistakes (CM). The second dimension was (Personal Standards, PS) setting high and relatively inflexible standards for personal performance since perfectionists tend to set high goals that are almost impossible to meet adequately. The third (Parental Expectations, PE) and fourth (Parental Criticism, PC) dimensions were the perception of the degree to which individuals perceive their parents as exceptionally demanding or critical. The fifth dimension (Doubts about Action, D) of perfectionism was the tendency to question the quality of one's performance, emphasizing the overall feeling that it is not satisfactory. Finally, the sixth dimension (Organisation, O) concerns Organisation, as people with high perfectionism seem to accentuate order and accuracy [1]. The first five subscales represent the core dimensions of the FMPS, so the Organisation subscale is excluded when calculating the FMPS total score. The scale consists of 35 items that assess perfectionism in the aforementioned six dimensions. High scores on the Organization subscale are not intrinsically problematic but, combined with high scores on the other factors, may exacerbate dysfunction. The questionnaire is scored in 5-point Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree). Good internal consistency and convergent validity with other measures of perfectionism have been demonstrated within community and clinical samples of adults [1,63]. In the present study, the Cronbach's alpha for the total scale was 0.89, and Cronbach's alpha for the subscales ranged from 0.76 to 0.88.
Depression Anxiety Stress Scale-21, (DASS-21) [64]: The scale assesses negative feelings and the degree to which the individual experiences these feelings. It consists of 21 items grouped in three subscales: (i) Depression, (ii) Anxiety, and (iii) Stress. There is evidence for high internal consistency for the subscales and good fit for the three-factor structure of the measure [65]. In the present study, we used the Greek version of DASS-21 [66]. In the present study, Cronbach's alpha for Depression, Anxiety, and Stress subscales were 0.90, 0.84, and 0.86, respectively.
Spielberger State-Trait Anxiety Inventory (STAI) [67]: The scale is a 40-item questionnaire assessing two types of anxiety-(i) state anxiety, or anxiety about an event and (ii) trait anxiety, or anxiety level as a personal characteristic anxiety. Participants answered on 4-point Likert scale from 1 (never) to 4 (always). The State-Trait Anxiety Inventory (STAI) is a reliable and valid measure that has been extensively used in research and clinical practice. It has been translated and validated in the Greek language [68]. In the present study, the State subscale had a Cronbach's alpha of 0.90, and the trait subscale had 0.80.
Measure of Parental Style (MOPS) [69]: This scale consists of 15 items that assess the perceived parenting styles. There are different subscales for Mother and Father. The MOPS focuses on three parenting styles of Indifference, Abuse, and Overcontrol. The questionnaire is scored in a 4-point Likert scale ranging from 0 (not true at all) to 3 (extremely true).
The sum of each subscale shows the degree to which that parental style was experienced by an individual. In the present study, the Cronbach's alpha for MOPS subscales ranged from 0.66 to 0.93.

Procedures
Call for participation in the study was announced through posters displayed in universities of Thessaloniki and through posts in social media pages of various schools of both universities. Participants were instructed to follow a link to a page that included information about the research, consent forms, and the questionnaires. Two-hundred and fifteen adults completed the questionnaires. Adults that reported high perfectionism in the self-reported FMPS questionnaire (at least one standard deviation above the mean score of the present sample) were invited to participate in the intervention study. From this initial sample, 135 (63%) participants presented high perfectionism, but only 81 (60%) of them opted to continue and were selected for the program. Participants were allocated either to the intervention or the wait-list group. Researchers used a schedule from a random number generator to assign participants to intervention (IG) or control group (CG).
The intervention lasted 10 weeks and consisted of two-hour weekly sessions. Consequently, the second measurements in both groups took place after 10 weeks. Nineteen participants out of forty from the intervention group completed the intervention program and the post-intervention measurement (Time 2). Participants in the IG who were absent for two consecutive sessions were excluded. If a participant was not able to attend a session, there was a filler session in between. The drop-out rate was 52.5%. Concerning the control group, 26 out of the original 41 failed to respond to the 2nd assessment.
The intervention protocol included the following sections: (i) The identification of dysfunctional/unhealthy perfectionism and those factors and behaviours that contribute to its maintenance; (ii) the recognition of dysfunctional thinking and the modification and reduction of dysfunctional behaviours; (iii) the application of behavioural experiments to challenge dysfunctional thoughts and to expose the individual to conditions they previously avoided; (iv) the application of cognitive techniques to reshape inelastic patterns and reduce self-criticism when high expectations are unmet; (v) the modification of cognitive distortions; and (vi) the application of cognitive-behavioural techniques to reduce related ineffective behaviours, such as procrastination [27,54,55].
Ethics approval was granted for this study by the Research Ethics and Deontology Committee of the University of Macedonia. The study was funded by Greece and the European Union (European Social Fund-ESF) through the Operational Programme «Human Resources Development, Education and Lifelong Learning 2014-2020». Table 1 shows the means and standard deviation of all measures in Time 1 before the randomization into two groups.  Table 2 shows all the correlations among FMPS subscales and the other measures. There was a significant correlation among most variables.

Regression Analysis in Time 1
A hierarchical regression was conducted to examine which factors could predict perfectionism in Time 1 measurements. The dependent variable was set as the perfectionism total score (FMPS total) in Time 1 measurement. All the variables of the Time 1 measurements were used as independent variables. Results showed that Father's Indifference and Abuse could predict the FMPS total score, R 2 = 0.50, F(10, 65) = 6.41, p = 0.000.

Analysis of Intervention Effect
The results of the two-way repeated measures ANOVA (Table 3)

Regression Analysis in Time 2
A hierarchical regression was conducted to examine which factors could predict the treatment outcome in the Intervention Group. The difference between the Time 1 measurement and the Time 2 measurement on the Frost Multi-Dimensional Perfection Scale was set as a dependent variable. All the variables of the Time 1 measurements were used as independent variables. There were no statistically significant findings.

Discussion
The purpose of the current study was to examine the relationship between perfectionism with the perceived parental styles and aspects of current psychopathology. Additional aims were to examine the effectiveness of a CBT group intervention to address perfectionism as well as possible predictive factors of the intervention outcome. Overall, the results replicate those of previous studies, suggesting that there is an association between perfectionism with specific parental styles as well as with depressive and anxious symptomatology. Furthermore, results showed that a CBT protocol targeting perfectionism can yield significant within-group changes on perfectionism and additional but more moderate benefits on other measures of symptomatology [70,71].
Concerning parental styles, our study supported that the Father's Indifference, Abuse, and over-control presented a low but statistically significant correlation with overall (total) perfectionism, and also Concern over Mistakes. Similarly, positive correlations were found between overall perfectionism and the subscale of CM with all the subscales of Mother's Parenting style. MOPS subscales assess negative parenting experiences [72]. To our knowledge, there is no study that examines the association between the MOPS parenting styles with Baumrind's parenting styles. However, we may assume that the dimensions of MOPS Abuse and Overcontrol resemble an authoritarian style characterized by high demandingness and low responsiveness. Previous research also supports the correlation between authoritarian style and perfectionism (especially socially prescribed perfectionism) [10,20,25,73,74]. Our study also found that MOPS subscale Father's Abuse, meaning violent behaviour that causes feelings of unsafety, predicted perfectionism. A recent study that examined possible causal relationships between negative parenting and psychopathology symptoms supported that pathogenic beliefs mediated the relationship between perceived negative parenting style and psychopathology in adults [72]. Pathogenic beliefs were defined as negative cognitions about self and others that cause emotional distress, for example, "I believe that I must be perfect in order to feel good about myself" [72] (p. 263). Concerning the aspect of Fear over Mistakes, Miller stated that "conditional acceptance by parents results in a child's thinking pattern of 'I am acceptable [to my parents] as long as I can perform well [make good grades, win awards, etc.]'" [20] (p. 357). This may explain the relationship between the perceived parental over-control and the concern over mistakes. Young adults may fear making mistakes because they experienced harsh parental criticism whenever they did not meet the standards. In the long run, they may have internalized their parents' criticism and transform it to their own harsh self-criticism [20]. The absence of correlations among the perfectionism aspects of Parental Expectations and Parental Criticism on the one hand and MOPS subscales on the other hand may support the existence of the over-critical inner voice. However, it has to be noted that the reports in our study on parenting were retrospective perceptions of the parenting style; thus, maybe participants may recall more easily incidences when they failed and their parents responded over-critically [25].
The subscale of indifference in MOPS appears to share common characteristics with the neglectful parenting style, as it is referred in theories about parenting [14][15][16]. In congruence with other studies [19], the present study showed that parental indifference correlated with Concern over Mistakes and the overall perfectionism but with no other aspects of perfectionism. Additionally, the present study found that Father's Indifference predicted total perfectionism. A possible explanation for this finding may be the argument that neglectful parents give so little structure, control, and involvement in their child's life that the child tries to take responsibilities and adopt perfectionistic tendencies in order to gain some control and sense of security in their life [7].
Furthermore, our study found that there was a strong relationship between perfectionism total score and other symptoms of psychopathology, such as stress, depression, and anxiety, which is in line with previous findings [32,33,40,75]. However, there was no correlation among the subscales of parental expectations and criticism and the other variables. An explanation may be that these subscales are not core features of perfectionism [76] and therefore may not contribute actively to symptoms of distress; they are rather viewed as precursors of perfectionism. Likewise, as indicated by Frost [1], organization is a correlate of perfectionism, not a central characteristic; therefore, we did not expect great correlations with these subscales.
Concerning our intervention, there was a significant time × group interaction in the FMPS Total and FMPS Concern over Mistakes subscale. The IG had significantly lower scores in Time 2 measures in the aforementioned variables compared with CG. Similarly, significant interactions were shown with DASS-21. Such results appear to support the growing amount of evidence suggesting that CBT can be effective in treating perfectionism [57,70] and that addressing perfectionism in a therapy setting may be beneficial in reducing other symptoms of distress [49]. Parental expectations and parental criticism remained largely unaffected after the intervention. As Rozental et al [70] suggested, parental criticism and expectations are less susceptible to change during treatment, as they measure perfectionism expressed by the parents of participants and are more related to its cause. Other aspects, such as Concern over Mistakes and Personal Standards, capture key symptoms associated with perfectionism [77]. Interestingly, there was no interaction in the subscale of personal standards, a finding that is not in line with other studies that supported significant changes [57,58]. However, in Lloyd's [57] review, Personal Standards appear to have smaller effect sizes than the Concerns over Mistakes subscale, which shows a trend of slighter changes in the dimension of Personal Standards even after different types of interventions (for example individual therapy, internet-based intervention, psychoeducation). A reason for this may be that Personal Standards are not only associated with negative outcomes but with positives outcomes as well [1]. High personal standards are associated with less procrastination and higher achievements [1,78]. It is possible that when concerns over mistakes decrease, then Personal Standards alone may not cause such a distress, and the person may not feel that they need to change them.
Lastly, a regression analysis was conducted, in the present study, to examine factors that may predict or contribute to the intervention outcome, but no such factors were identified. Similarly, Rozental et al. [70], in their study, attempted to identify such factors from the selected variables, but they did not find any significant factors that could predict treatment outcome.
A number of limitations are important to address when reviewing our results. The first limitation was that there was a considerable rate of participant drop-out, resulting in a small active group of intervention participants. Secondly, the design of the study did not include a second condition in order to address the effectiveness of the particular protocol. Third, the majority of our participants mostly consisted of highly educated women, a sample which is not considered representative of the general population. Also, participants were recruited through announcements that may appeal to a certain demographic that identifies themselves as perfectionists or experience problems with perfectionism to a greater degree than the general population. The use of exclusively self-reported questionnaires is another obvious limitation of the study. Finally, the current study did not have a follow-up assessment, making it impossible to explore the medium-or long-term results of the intervention. Nevertheless, a strength of this study is that the results could possibly evidence the effectiveness of a protocol addressing perfectionism in different cultures.
Future research should focus on examining the effectiveness of CBT programs compared with other active treatments. Moreover, follow-up assessments and longitudinal research designs are needed to examine long-term improvements or changes. Also, causal relationships concerning the development of perfectionism may shed light on factors that may affect treatment as well. Although perfectionism is not a diagnostic disorder per se, it may have a role in the core of other disorders, and thus the results of such an intervention may be useful in transdiagnostic treatments settings.

Conclusions
Overall, the present results suggest that parenting style is associated with the development of specific, maladaptive aspects of perfectionism. In addition, it shows clearly that a ten-session, cognitive-behavioural group intervention can be effective in the management of maladaptive perfectionism. Furthermore, the CBT intervention addressing perfectionism may also be beneficial for individuals with other difficulties related to clinical perfectionism, such as anxiety or other symptoms. More research is, however, needed to determine its efficacy in the long-run. Further research in the drop-out rate will be useful in the future to examine why some people commit to the therapeutic process while others do not. Finally, further research on related variables, as potential predictors of treatment outcome, will be important in order to improve therapies. As Egan et al. [27] suggested, an important clinical implication is that, if perfectionism is elevated in a patient to the degree that interferes with their quality of life, then it should be addressed directly in therapy, particularly if it appears to undermine change.