Understanding eating disorder symptoms in adolescents: testing a cognitive-behavioural model of eating disorders in a community sample

ABSTRACT

Clinical perfectionism, self-esteem, mood intolerance, and interpersonal difficulties are associated with eating disorder symptoms in clinical samples. The aim of the current study was to test a model including clinical perfectionism, self-esteem, mood intolerance, and interpersonal difficulties to understand eating disorder symptoms in an adolescent community sample. Adolescents (N = 446, M age = 16.25 years, SD = 1.64; 74.2% female) completed measures of clinical perfectionism, self-esteem, mood intolerance, interpersonal difficulties, and eating disorder symptoms. Path analysis indicated clinical perfectionism, self-esteem, mood intolerance, and interpersonal difficulties were all directly associated with symptoms of eating disorders, and that clinical perfectionism was indirectly associated with eating disorders through self-esteem, mood intolerance, and interpersonal difficulties. The results indicate the cognitive-behavioural model of eating disorders can be applied to adolescents in the community with symptoms of eating disorders. Directional causality between constructs should be established to understand whether increased clinical perfectionism, and reduced self-esteem, mood intolerance, and interpersonal difficulties are vulnerabilities to the development and maintenance of eating disorders.

Clinical implications

• Transdiagnostic maintaining processes are linked to eating disorder symptoms.

• The transdiagnostic model of eating disorders has indirect support in adolescents.

• Clinical perfectionism may be an important factor to target in treatment.

• Future research should examine transdiagnostic preventive interventions.

Introduction

Eating disorders have their peak onset in adolescence and a severe impact on physical and mental health (Qian et al., 2022). At the core of eating disorders is concern over eating, shape, weight and their control (Fairburn et al., 2003b). Disordered eating behaviours may consist of overeating or insufficient food intake and be associated with compensatory behaviours such as purging or excessive exercise (Schlegl et al., 2018). Rates of eating disorder symptoms in the community are high, with 63% of adolescents in the community reporting fear of gaining weight, and up to 29% reporting distress about weight or shape (Micali et al., 2016).

Fairburn et al. (2003b) described a transdiagnostic model of eating disorders, which includes cognitive and behavioural processes that maintain symptoms across eating disorders. The core mechanisms that maintain eating disorder symptoms include a dysfunctional scheme for self-evaluation based on eating, shape, weight, and their control, interacting with other factors e.g., strict dieting, binge eating, and weight control behaviour. Fairburn et al. (2003b) outlined that in certain individuals, one or more of four maintaining mechanisms may interact with the core eating disorder maintaining mechanisms. The four maintaining mechanisms are clinical perfectionism, mood intolerance, core low self-esteem, and interpersonal difficulties (Fairburn et al., 2003b). Clinical perfectionism refers to self-worth based on achievement despite negative effects. Mood intolerance involves the intolerance of adverse mood states. Interpersonal difficulties are the problems that an individual experiences in relationships with others. Core low self-esteem refers to a global, pervasive negative view of self.

The four maintaining factors have been associated with eating disorders in adolescents (Bills et al., 2023; Drieberg et al., 2019; Pelletier Brochu et al., 2018; Zamani Sani et al., 2021). However, no studies have examined the relationship between all four maintaining mechanisms and eating disorder symptoms in a community sample of adolescents. It is important to understand whether proposed relationships in Fairburn et al. (2003b) transdiagnostic model remain true for adolescents from a dimensional perspective, given it would be expected that relationships may hold between community and clinical samples. This is important to inform targets for prevention programs in adolescents who have elevated symptoms from developing an eating disorder.

The cognitive behavioural model of eating disorders (Fairburn et al., 2003b) has been examined in several adult community samples, with findings providing support for clinical perfectionism, low self-esteem, mood intolerance, and interpersonal difficulties in eating disorder symptoms (Dakanalis et al., 2014; Hoiles et al., 2012; Lampard et al., 2013). There are limitations when generalising the findings from previous research to adolescent populations given developmental differences. It should not be assumed that findings which apply to adults will be the same for adolescents (Bills et al., 2023; Lunn et al., 2023). There are changes in the relationships among eating disorder symptoms between adolescents and adults (Christian et al., 2020). Hence, we specifically aimed to examine the transdiagnostic model in adolescents.

It would also be useful to further examine the transdiagnostic model of eating disorders in adolescents given some inconsistencies to date. Curzio et al. (2018) found in a sample of children and adolescents with eating disorders support for the association of low self-esteem, clinical perfectionism, interpersonal problems and mood intolerance. However, in a sample of adolescents with eating disorders, while E. Jones et al. (2020) found that self-esteem and mood intolerance were directly associated with eating disorder symptoms, a direct association between perfectionism and eating disorder symptoms was not found, in contrast to previous studies (Drieberg et al., 2019; Johnston et al., 2018; Morgan-Lowes et al., 2019).

A large base of evidence has been accumulated demonstrating that perfectionism is associated with symptoms of eating disorders (Bills et al., 2023; Limburg et al., 2017; Livet et al., 2023; Stackpole et al., 2023). Multidimensional perfectionism is one definition of perfectionism, based on the Multidimensional Perfectionism Scales (FMPS; Frost et al., 1990; HMPS; Hewitt & Flett, 1991). A consistent two-factor solution demonstrates multidimensional perfectionism consists of perfectionistic strivings; striving towards high standards, and perfectionistic concerns; worry over mistakes and the belief that others expect perfection (Smith & Saklofske, 2017). Clinical perfectionism is another definition, where self-worth is based on striving to meet high standards despite negative consequences (Shafran et al., 2002). The definition of clinical perfectionism is central to cognitive behaviour therapy (CBT) for perfectionism, which has been demonstrated to reduce symptoms of eating disorders (see Shafran et al., 2023 for a review). Perfectionism is important to understand as when adolescents apply rigid standards, for example “I must weigh 50 kg to be the perfect weight” they engage in eating disorder behaviours i.e., dieting to achieve their strict standard. When an adolescent believes they have not met their standard e.g., they weigh 51 kg instead of 50 kg, this reinforces their self-esteem as based on achievement and they conclude they are “not good enough”, which reinforces both clinical perfectionism and eating disorder symptoms (see Egan et al., 2023).

Given the consistent association between perfectionism and eating disorders, clinical perfectionism may play a particularly important role in the transdiagnostic model (Fairburn et al., 2003b). Perfectionism is also important to examine given evidence for increasing rates in adolescents (Curran & Hill, 2019). We aimed to clarify the inconsistent findings in E. Jones et al. (2020) by using a specific measure of clinical perfectionism rather than multidimensional perfectionism as used in that study. Further, we aimed to determine whether clinical perfectionism plays a pertinent role in the association of maintaining factors in Fairburn et al. (2003b) model given the extensive evidence for the association between perfectionism and symptoms of eating disorders (Bills et al., 2023; Limburg et al., 2017; Livet et al., 2023; Stackpole et al., 2023).

The aim of the current study was to examine the associations between mechanisms in Fairburn et al. (2003b) transdiagnostic model and eating disorder symptoms in a community-based adolescent sample. We hypothesised there would be direct associations between maintaining mechanisms and symptoms of eating disorders, and indirect associations between clinical perfectionism and eating disorder symptoms through self-esteem, mood intolerance, and interpersonal difficulties.

Method

Participants

The inclusion criteria for the study were (1) age 13 to 18 years, and (2) residing in Australia. The exclusion criteria were age range outside of 13–18 years and residing in another country than Australia. Adolescents residing in Australia were recruited through social media and the Curtin University research participation pool. Recruitment material informed potential participants that they were being asked to take part in a study investigating factors associated with psychological distress. Participants were 446 adolescents aged 13 to 18 years (M = 16.25 years, SD = 1.64; 74.2% female). A third of these participants (32%) self-reported a psychological disorder, for example, an anxiety disorder or depression, of whom 80% reported it was a current diagnosis.

On average, participants were in the non-clinical range (M = 2.14) on the Eating Disorder Examination Questionnaire (EDEQ; Fairburn & Beglin, 1994), however an independent samples t-test revealed that the current sample had significantly higher eating disorder symptoms (M = 2.14, SD = 1.52) compared to a community sample (M =1.59, SD = 1.32) used in the establishment of norms for the EDE-Q (t[1,630] = 7.19, p < .001; Mond et al., 2006). Based on the number of parameters in the model and a 10:1 ratio, the minimum number of participants required was 280 (Kline, 2005).

Measures

Clinical perfectionism questionnaire

The 12-item self-report Clinical Perfectionism Questionnaire (CPQ; Fairburn et al., 2003a) was used to measure clinical perfectionism over the past month. Participants respond on a 4-point Likert scale ranging from 1 (not at all) to 4 (all the time). The CPQ demonstrates good reliability and validity in adolescents and adults, including non-clinical and clinical eating disorder samples (Egan et al., 2016; Howell et al., 2020; Prior et al., 2018; Shu et al., 2020). In the current study, the CPQ demonstrated good internal consistency (α = .85).

Rosenberg self-esteem scale

The 10-item self-report Rosenberg Self-Esteem Scale (RSES; Rosenberg, 1965) was used to measure self-esteem. Participants responded on a 4-point Likert scale ranging from 1 (strongly disagree) to 4 (strongly agree). The RSES has good reliability and validity (Rosenberg, 1965), including in adolescents (Schmitt & Allik, 2005). The RSES demonstrated excellent reliability in the current study (α = .92).

Tolerance of mood states scale

The 11-item self-report Tolerance of Mood States Scale (TOMS; Allen et al., 2012) was used to assess mood intolerance. Participants respond using a 5-point Likert scale ranging from 1 (never) to 5 (always). The TOMS has acceptable reliability and validity in community samples (α = .77; Allen et al., 2012; Hoiles et al., 2012). Reliability for the TOMS in the present study was good (α = .80).

Inventory of interpersonal problems

The 32-item self-report Inventory of Interpersonal Problems (IIP; Barkham et al., 1996) was used to measure interpersonal problems. Participants respond using a 5-point Likert scale ranging from 0 (not at all) to 4 (extremely). The IIP has previously been used in adolescent populations and in individuals with eating disorders (McEvoy et al., 2013), and has acceptable validity and reliability (Barkham et al., 1996), and excellent reliability in the present study (α = .90).

Eating disorder examination questionnaire

The Eating Disorder Examination Questionnaire (EDE-Q; Fairburn & Beglin, 1994) is a 28-item self-report measure that assesses eating disorder symptoms over the past 28 days. The EDE-Q is scored on a 7-point Likert scale ranging from 0 (no days) to 6 (every day) and is divided into four subscales: eating concerns, restraint, shape concerns, and weight concerns. The EDE-Q has good reliability and validity in adolescent samples (Berg et al., 2012). Reliability in the present study was excellent (α = .91).

Procedure

Following ethics approval from the Curtin University Human Research Ethics Committee (HE2019–0024) the questionnaires were completed on the Internet via the Qualtrics survey platform. University students recruited through a student participation research pool were awarded points for their participation which contributed towards study requirements. Participants recruited through social media had the option to place their name in a draw to win a gift card valued at AUD$20.

Results

Results from the missing data analysis will be presented followed by the data analysis plan. Results from the first configuration of the model will then be presented, where direct associations between maintaining mechanisms and symptoms of eating disorders were examined. Results from the second configuration of the model will follow, where indirect associations between clinical perfectionism and eating disorder symptoms through self-esteem, mood intolerance, and interpersonal difficulties were examined.

There were 591 adolescents who opened the survey. However, 145 adolescents either did not consent, or had completed fewer than two of the measures, hence the final sample comprised 446 adolescents. Percentage of missingness on each measure ranged from 21% to 25%. The measure with the most missing data was the IIP. Missing values analysis on the remaining 446 participants revealed a non-significant Little’s MCAR test, χ² (df = 532) = 568.50, p = .132, indicating data were missing completely at random. Missing data were replaced using expectation-maximisation.

Path analysis in MPlus (Múthen & Múthen, 2012) was used to examine the associations between the transdiagnostic maintaining mechanisms and eating disorder symptoms. Direct pathways between the four maintaining processes outlined by Fairburn et al. (2003b) and eating disorder symptoms were tested. In a second configuration of the model, indirect pathways between clinical perfectionism and eating disorder symptoms through self-esteem, mood intolerance, and interpersonal difficulties were tested.

Modification indices (Mis >20, Hu & Bentler, 1999) were examined to check for improvements of the model, and theoretically defensible paths were freed. No modification indices were identified over 20, nor was there a theoretical justification for freeing parameters in either model for those below 20. The significance values for direct pathways were estimated with a 95% confidence interval using a bootstrapping procedure based on 1,000 draws from the data.

Goodness-of-fit was assessed using the chi-square statistic and degrees of freedom (chi-square/df), comparative fit index (CFI; values should be ≥ 0.95), root mean square error of approximation (RMSEA; values should be ≤ 0.06), Tucker-Lewis index (TLI; values should be ≥ 0.95), and standardised root mean square residual (SRMR; values should be ≤ 0.08, Hu & Bentler, 1999). Age and gender were controlled for in analyses as they were significantly associated with the independent and dependent variables (see Table 1).

Table 1. Correlation matrix between clinical perfectionism, self-esteem, mood intolerance, interpersonal difficulties, eating disorder symptoms, age, and gender.

	Clinical Perfectionism	Self-Esteem	Mood Intolerance	Interpersonal Difficulties	Eating Disorder Symptoms	Age	Gender
Clinical Perfectionism	-
Self-Esteem	−.46***	-
Mood Intolerance	.49***	−.61***	-
Interpersonal Difficulties	.44***	−.63***	.62***	-
Eating Disorder Symptoms	.45***	−.51***	.48***	.49***	-
Age	−.12*	.16**	−.21***	−.14**	.23***	-
Gender	.23***	−.14**	.07	.03	.07	.06	-

Note. N = 446. For Gender 0 = Male, 1 = Female.

*p < .05, **p < .01, ***p < .001.

Model 1

The path analysis revealed good model fit χ²/df = 1.45, CFI = .998, RMSEA = .032 (90% CI < .001; .135), TLI = .972, SRMR = .011. In combination, clinical perfectionism, self-esteem, mood intolerance, and interpersonal difficulties explained a significant 34.9% of variance in eating disorder symptoms. There were significant direct associations between clinical perfectionism (b = .04, 95% CI [.03, .06], p < .001), self-esteem (b = −.05, 95% CI [−.07, −.03], p < .001), mood intolerance (b = .23, 95% CI [.04, .43], p = .046), and interpersonal difficulties (b = .46, 95% CI [.25, .68], p < .001) and eating disorder symptoms (Figure 1).

Figure 1. Direct pathways from clinical perfectionism, self-esteem, mood intolerance, and interpersonal difficulties to eating disorder Symptoms. Standardised estimates and standard error reported. *p < .05, **p < .01, ***p < .001.

Model 2

The second model tested direct pathways between clinical perfectionism and the core psychopathology to eating disorders, in addition to indirect pathways through the four mechanisms. The path analysis revealed good model fit χ²/df = 1.448, CFI = 1.000, RMSEA = .032 (90% CI = .000; .135), TLI = .990, SRMR = .011. There was a significant direct association between clinical perfectionism and eating disorder symptoms (b = .04, 95% CI [.03, .06], p < .001). There was a significant negative indirect association between clinical perfectionism and eating disorder symptoms through self-esteem (b = .02, 95% CI [.01, .03], p = .001). There was also a significant indirect association between clinical perfectionism and eating disorder symptoms through mood intolerance (b = .01, 95% CI [.00, .02], p = .039). There was a significant indirect association between clinical perfectionism and eating disorder symptoms through interpersonal difficulties (b = .02, 95% CI [.01, .03], p = .001). The four mechanisms explained a significant 34.7% of variance in core psychopathology (Figure 2).

Figure 2. Indirect pathways from clinical perfectionism to eating disorder symptoms through self-esteem, mood intolerance, and interpersonal difficulties.

Standardised estimates and standard error reported. *p < .05, **p < .01, ***p < .001.

Discussion

The results revealed significant associations between symptoms of eating disorders and all four maintaining mechanisms in Fairburn et al. (2003b) transdiagnostic model of eating disorders in a community sample of adolescents. The findings are consistent with previous research in a clinical sample of adolescents with eating disorders (Curzio et al., 2018).

The findings supported the hypothesis that clinical perfectionism is directly associated with eating disorder symptoms, in line with previous research (Drieberg et al., 2019; Johnston et al., 2018; Morgan-Lowes et al., 2019). While this finding is consistent with literature highlighting the important role of perfectionism in adolescents with symptoms of eating disorders (Bills et al., 2023; Dahlenburg et al., 2019; Livet et al., 2023), it contrasts with the findings of E. Jones et al. (2020), in which the direct association between perfectionism and eating disorder symptoms was not significant in adolescents with eating disorders. However, it is unclear why the E. Jones et al. (2020) study was an outlier compared to the numerous studies which have found a direct association between perfectionism and eating disorder symptoms (Bills et al., 2023).

The association between clinical perfectionism and eating disorder symptoms in community samples of adolescents indicates future research is warranted in understanding these relationships in a longitudinal study and examining causality. For example, similar to existing longitudinal evidence for the relationship between perfectionism and drive for thinness (Dickie et al., 2012), perfectionism and emotion regulation (Vois & Damian, 2020), and perfectionism and bulimic symptoms (Keyhayes et al., 2019). If appropriate, research may then examine the efficacy of internet delivered Cognitive Behaviour Therapy for clinical perfectionism (CBT-P) interventions (e.g., Egan et al., 2023; Galloway et al., 2022; Shafran et al., 2023; Shu et al., 2019) for adolescents in the community at risk of developing eating disorders.

The results also revealed significant indirect associations between clinical perfectionism and eating disorder symptoms via the other three transdiagnostic maintaining mechanisms. As expected, and in line with previous research (Puttevils et al., 2019), clinical perfectionism was indirectly associated with eating disorder symptoms through core low self-esteem, mood intolerance, and interpersonal difficulties. This suggests clinical perfectionism may be a central driving factor in the transdiagnostic cognitive-behavioural model of eating disorders and is consistent with data demonstrating the consistent association between perfectionism and eating disorder symptoms (Bills et al., 2023; Limburg et al., 2017; Stackpole et al., 2023). Future research should examine these complex relationships with the use of longitudinal data and investigate whether a reduction in clinical perfectionism also reduces the other maintaining mechanisms, which may have a positive impact on eating disorder symptoms.

Consistent with previous findings (E. Jones et al., 2020; Smink et al., 2018; Zamani Sani et al., 2021), low self-esteem was associated with symptoms of eating disorders. This association is therefore not specific to adolescents with an eating disorder but also applies to adolescents in the community without an eating disorder diagnosis. Self-esteem has been shown to change across the lifespan, with generally lower levels during adolescence, and gradually increasing into adulthood (Orth et al., 2010). To gain a more in-depth understanding of the relationship between self-esteem and eating disorder symptoms, future research may adopt a longitudinal approach examining eating disorder symptoms alongside a prediction of increasing self-esteem across the lifespan (Orth et al., 2010). This longitudinal approach will give a stronger indication as to whether self-esteem is a worthy target for eating disorder prevention in community samples of adolescents.

Consistent with previous research, mood intolerance (Curzio et al., 2018; E. Jones et al., 2020) and interpersonal difficulties (Curzio et al., 2018) were directly associated with eating disorder symptoms. The inconsistency in findings in the current study and E. Jones et al. (2020) in which the direct relationship between interpersonal difficulties and eating disorder symptoms was not significant may be explained by differences in measures used. Despite evidence that the IIP-32 that was used in the current study is a valid and reliable measure in individuals with an eating disorder (McEvoy et al., 2013), general interpersonal difficulty measures may not adequately address eating-specific interpersonal problems, such as avoiding others’ judgments about one’s appearance (S. Jones et al., 2019). Future research should examine the model using measures of interpersonal difficulties that are specific to eating disorders such as the Interpersonal Relationships in Eating Disorders Scale (S. Jones et al., 2019).

Results from the current study may be useful for guiding future eating disorder prevention research in non-clinical samples of adolescents. The present research has found that the same mechanisms that are both theoretically and empirically linked to eating disorder symptoms in clinical populations, are also associated with eating disorder symptoms in a sample of youth in the community. Given eating disorder symptoms are common among adolescents, yet often remain undetected in the community (Keski-Rahkonen et al., 2009), this may have important implications in terms of the prevention of eating disorders, for example examining the utility of targeting the transdiagnostic maintaining mechanism of clinical perfectionism in a targeted prevention program (see O’Brien et al., 2022).

There were numerous limitations of the current study. The cross-sectional nature of the data is a major limitation. With the use of cross-sectional data, it is not possible to examine causality and the dynamic nature of the relationships found. Future research should examine the predictive utility of the maintaining processes outlined by Fairburn et al. (2003b) using longitudinal data. This will enable a clearer picture of whether the maintaining mechanisms lead to, and maintain, eating disorder symptoms over time. A further limitation was that specific information regarding which self-reported psychological disorders participants met was not collected. Future research should examine whether diagnostic status moderates the associations between the transdiagnostic maintaining mechanisms and eating disorder symptoms. While the aim of the current study was to extend on the findings of E. Jones et al. (2020), comparisons between findings should be made with caution due to the different measures used for perfectionism, mood intolerance, and interpersonal difficulties.

Although we aimed to examine factors in the transdiagnostic model across different genders, only approximately 24% of the sample self-identified as males and the remaining participants self-identified as female. No one in the sample identified as non-binary or gender diverse. As such, findings cannot be generalised to gender diverse samples. Given transgender and other gender diverse youth are at increased risk for developing an eating disorder (Duffy et al., 2019; Watson et al., 2017), it is imperative that researchers endeavour to conduct eating disorder research in a gender inclusive manner, as well as research in gender diverse populations specifically. Further, no association was observed between gender and eating disorder symptoms. The results suggest that males and females in our community sample were experiencing similar rates of eating disorder symptoms. A hypothesis to consider in future research is whether young males now have an improved understanding of and therefore reporting of eating disorder symptoms compared to previous studies where females have been typically reported to have higher eating disorder symptoms. The results may potentially reflect a more equal distribution currently of eating disorder symptoms between males and females residing in the community than reported previously.

In summary, the present results provide support for the association between Fairburn et al. (2003b) transdiagnostic maintaining mechanisms and eating disorder symptoms in a community sample of adolescents. The findings highlight the potential role of core low self-esteem, mood intolerance, interpersonal difficulties, and in particular clinical perfectionism, in symptoms of eating disorders. The results suggest further research is required examining the factors in the transdiagnostic model in community and clinical samples of adolescents using experimental and longitudinal designs.

Disclosure statement

Sarah Egan receives royalties for the books Cognitive-behavioral treatment of perfectionism and Overcoming perfectionism: A self-help guide using scientifically supported cognitive behavioural techniques.

Data availability statement

The data is available upon request from the authors subject to Institutional approvals.

Additional information

Funding

The author(s) reported there is no funding associated with the work featured in this article.