Low satisfaction with normative life domains in adolescents with anorexia nervosa

Abstract Objective Low satisfaction with normative life domains might be an important factor in the persistence of anorexia nervosa (AN). Initial evidence in non‐clinical samples showed that lower satisfaction with normative life domains was related to more intense eating disorder symptoms. As a critical next step, the current study examined satisfaction with normative life domains in a clinical sample. Specifically, the present study tested whether adolescents with AN reported lower satisfaction with normative life domains than adolescents without an eating disorder. Method Adolescents with AN (n = 69) and adolescents without an eating disorder matched on age, gender and educational level (n = 69) completed the Brief Multidimensional Students' Life Satisfaction Scale to assess satisfaction with five life domains (family, friendships, school, self and living location) and life in general. Results Adolescents with AN reported significantly lower satisfaction with normative life domains than the comparison group. Subsequent analyses showed that this overall group difference was primarily driven by adolescents with AN reporting lower satisfaction with the self, school experience and life in general. Conclusions Findings supported the hypothesis that adolescents with AN show relatively low satisfaction with meaningful, non‐AN‐related life domains. This points to the potential relevance of enhancing satisfaction with specific life domains to optimize treatment effectiveness.

According to an influential cognitive behavioural theory of eating disorders, most of the clinical features seen in AN can be understood as stemming directly from the excessive influence of body weight and shape on self-evaluation (Fairburn et al., 2003). In contrast to most people whose lives are guided by values and goals in different life domains (e.g., work, school and relationships), individuals with AN seem to primarily concern themselves in regards to values related to weight, shape and eating behaviour (Fairburn et al., 2003). This overvaluation may create a self-reinforcing loop, in which the goal to lose weight enhances the motivational salience of goal-related cues (e.g., 'fat' body parts), which causes these cues to grab attention and to be processed more elaborately, which in turn reactivates the goal to lose weight again (Fairburn et al., 2003;Williamson et al., 2004). It has been hypothesized that, as a consequence, the salience landscape (i.e., to what extent environmental stimuli elicit appetitive or aversive motivation) and goal-directed activity of individuals with AN may increasingly narrow to weight and shape-related concerns.
The pursuit of important goals, together with the feeling that life makes sense and that one matters in the world, is thought to play a central role in whether individuals experience their lives as meaningful (George & Park, 2016;Martela & Steger, 2016). In addition, theory and empirical findings suggest that the experience of life meaning depends on the ability to engage with one's goals in a positive, satisfying way (King et al., 2006;Wolf, 2010). According to Fox and Leung (2009), AN may serve as a maladaptive attempt to deal with a sense of purposelessness or existential anxiety. That is, in the short run, AN may provide individuals with goals and feelings of purpose (Ison & Kent, 2010;Serpell et al., 1999) and thus a way of imposing meaning upon their world (Fox & Leung, 2009). However, the narrowed salience landscape and goal-directed activity of individuals with AN (i.e., striving to achieve AN-related goals) may lead to a tendency to neglect or compromise more overarching personal values and common sources of life meaning (Mulkerrin et al., 2016), including interpersonal relationships, personal development and generativity (Emmons, 2003;Schnell, 2011). Therefore, in the long run, AN may create a myopic perspective that prevents access to a wide range of normative goals and domains that provide meaning across time and situations (Marco et al., 2017).
Although excessive engagement with weight and shape-related goals is seen as a core mechanism of AN (Fairburn et al., 2003), relatively little research has examined the role that a lack of engagement and satisfaction with normative, non-AN-related goals might play in the disorder. The only two clinical studies on meaning in life and eating disorder pathology (Marco et al., 2017;Marco et al., 2019) have used the Purpose in Life Questionnaire (Crumbaugh & Maholick, 1969), which cannot differentiate satisfying engagement with normative goals from satisfying engagement with AN-related goals (with items as 'In life I have no goals nor aims at all'). This could be problematic, as measures that do not differentiate between these types of goals could potentially show high purpose in individuals with AN, due to satisfactory engagement in dysfunctional AN-related goals (Ison & Kent, 2010;Serpell et al., 1999).
Several studies can be found demonstrating that individuals with AN report lower levels of satisfaction with life as a whole than individuals without an eating disorder (Garcia et al., 2017;Kitsantas et al., 2003;Magallares et al., 2014). However, thus far, only one study has examined the relationship between eating disorder symptoms and satisfaction with normative, non-AN-related domains (Matthews et al., 2012). In this study, a non-clinical student sample completed the college version of the Brief Multidimensional Students' Life Satisfaction Scale (BMSLSS; Seligson et al., 2003), which allows the assessment of satisfaction with respect to different life domains (e.g., family life, friendships and school experience) and life in general, which are common goals or sources from which people derive meaning (Delle Fave et al., 2011). Results indicated that multiple eating disorder behaviours and weight perceptions (e.g., binge eating, worrying about weight and self-described weight) were inversely related to satisfaction with each individual domain (Matthews et al., 2012). The strongest correlations were found between students' weight perceptions and satisfaction with the self and physical appearance, followed by satisfaction with life in general.
The study of Matthews et al. (2012) provides important initial evidence that, in a non-clinical sample, less satisfaction with normative life domains is associated with more eating disorder symptoms. A crucial next step is to examine whether this inverse relation is also evident in individuals who are clinically diagnosed with AN. It is important to provide empirical evidence on this matter, as the overall impairment in functioning seen in AN (American Psychiatric Association, 2013) does not necessarily equate less satisfaction with normative life domains. Furthermore, the present study differentiates between satisfaction with various domains, which could provide concrete targets for interventions. Our main aim was to examine satisfac-  ( Agh et al., 2016;Engel et al., 2009;Sy et al., 2013). We therefore hypothesized that adolescents with AN would report less satisfaction with normative life domains compared to a comparison group of adolescents without an eating disorder. Furthermore, because the salience

Key Practitioner Message
• Adolescents with AN were less satisfied with normative life domains, compared to adolescents without an eating disorder.
• In particular, adolescents with AN were less satisfied with themselves, with their school experiences and with life in general.
• Addressing low satisfaction with these life domains may be beneficial in optimizing treatment effectiveness.
van DOORNIK ET AL.
landscape and goal-directed activity in individuals with AN seem to increasingly narrow to weight and shape-related concerns (Fairburn et al., 2003), we also hypothesized that individuals with AN would report normative life domains to be less important, compared to the comparison group. Short oral presentations about the current study were given to several groups of students, which were chosen based on the characteristics of the individuals with AN (i.e., age and educational level). After the presentation participants were informed about when the assessment would take place (i.e., approximately 2 weeks later) and received an information letter and informed consent forms for themselves and their parents to take home.    (Bickman et al., 2010;Büssing et al., 2009;Seligson et al., 2003;Zullig et al., 2009). In the current study, the internal consistency was satisfactory for both individuals with AN and the comparison group (α = .72 and α = .87, respectively). More item descriptive statistics can be found in Table 1.

| Satisfaction with and importance of normative life domains
Furthermore, in a separate set of questions, participants' importance ratings for each life domain and overall life were obtained, using a scale from 1 (very unimportant) to 5 (very important; cf. Seligson et al., 2003). The items were scored in the same manner as the aforementioned items, with higher scores indicating the life domains to be more important. The internal consistency for these items was considered to be sufficient for the individuals with AN and excellent for the comparison group (α = .66 and α = .92, respectively).

| Procedure
The current study was approved by the medical ethical committee of the University Medical Center in Groningen in the Netherlands (NL.51694042.14). All participants, and their parents when participants were younger than 18, provided written informed consent. As part of the regular diagnostic procedure at the Eating Disorder

| Statistical analyses
Group differences on adjusted BMI were assessed with an independent sample t test. Because EDE-Q scores were not normally distributed (see the supporting information), a Mann-Whitney U test was used to assess group differences on the EDE-Q. To examine differences on satisfaction with and importance of normative life domains between adolescents with broadly defined AN and those without an eating disorder, a Multivariate Analysis of Variance (MANOVA) was performed with the average satisfaction score and the average importance score as dependent factor, and group (AN vs. non-AN-comparisons) as fixed factor. Afterwards, univariate ANOVAs were used to examine on which variable(s) differences between groups were found.
Power calculations indicated that the power of the MANOVA was .74 to detect medium-sized effects with α = 0.05 and the power of the between subject test was .83.

| Group characteristics
Group characteristics are presented in

| Satisfaction and importance of normative life domains
When checking the assumptions for the analyses, visual inspection showed that the assumption of multivariate normality of residuals for the mean importance score was violated to a certain degree. Additionally, two multivariate outliers (both from the comparison group, regarding mean importance score) were detected when checking Mahalanobis Distance. However, MANOVAs are relatively robust to violations of multivariate normality, especially when using relatively large sample sizes as in the current study (Finch, 2005). Furthermore, because excluding the outliers did not visually improve multivariate normality and also did not significantly change the results, a MAN-OVA using all participants is reported.

| Post hoc tests
As the between subjects tests showed differences between groups in satisfaction with the meaningful life domains, we subsequently examined for which specific domains these differences between groups were evident. Because assumptions of normality were violated when looking at individual domains, we decided to conduct six separate Mann-Whitney U tests. A Bonferroni-Holm correction was applied to correct for familywise error rate. Therefore, the smallest p value was tested against an alpha of.0083; the p values following against .01, .0125, .0167, and .025 and the largest against .05. The analyses rev-

| DISCUSSION
The current study examined whether adolescents with broadly defined AN differed from healthy weight adolescents without an eating disorder in their satisfaction with and importance of normative life domains. In line with our hypotheses, individuals with AN reported lower satisfaction with normative life domains than individuals without an eating disorder. These results may not necessarily be unexpected, given that eating disorder symptoms were expected to have a negative correlation with satisfaction (Claydon et al., 2020). However, we found no significant difference regarding the importance of these domains between groups. Furthermore, exploratory analyses showed that the overall group difference in the satisfaction measure was primarily driven by individuals with AN reporting lower satisfaction with the self, school experiences and life in general.   These findings are consistent with previous research in a nonclinical, college student sample (Matthews et al., 2012), in which disordered eating behaviours and weight perceptions were found to be inversely related to satisfaction with normative life domains. To the best of our knowledge, no other study has examined satisfaction with specific normative life domains in relation to eating disorder pathology. Despite the relative dearth of research studying satisfaction with normative domains in individuals with AN, there are a few studies examining satisfaction with life as a whole, generally using the Satisfaction With Life Scale (Pavot & Diener, 1993). Results show that individuals with AN report lower levels of satisfaction with life as a whole than individuals without eating disorders (Garcia et al., 2017;Kitsantas et al., 2003;Magallares et al., 2014). Because individuals with AN included in these studies were all college students or adults with a mean age of 21 or higher, our results extend previous findings by showing that lowered satisfaction with life in general can also be found in adolescents with AN.

T A B L E 3 Mean satisfaction and importance scores per group
In addition, lower satisfaction with the self was found to differentiate between individuals with AN and individuals without eating disorders in the current study (cf. Matthews et al., 2012). However, when asking individuals to rate how satisfied they are with themselves, it is debatable whether the measure taps into satisfaction with a normative life domain or whether it actually assesses self-esteem.
The difference between the groups on this item is in line with the observation that eating disorder pathology is strongly related to low self-esteem (Brockmeyer et al., 2013;Fairburn et al., 2003). For example, in a longitudinal study, Halvorsen and Heyerdahl (2006)  The present study also showed some unexpected findings. First, it is surprising that individuals with AN did not report lower satisfaction with their friends and family than the comparison group, given earlier findings that individuals with AN report elevated stress related to these domains. For example, previous studies have shown that difficulties with social interactions and friendships may play a role in the development and maintenance of AN (Doris et al., 2014;Treasure & Schmidt, 2013;Westwood et al., 2016). With respect to family functioning, several unhealthy family styles like rigidity, poor communication, low social support and criticism from mothers have also been shown to be related to the development of AN (Ghaderi, 2003;Holtom-Viesel & Allan, 2014;Krug et al., 2015).
Moreover, in two previous studies, both individuals with AN and their parents reported lower satisfaction with their families compared to controls (Fisher & Bushlow, 2015;Laghi et al., 2015). However, two differences between the current study and previous studies might explain the discrepancy between previous and current findings. First, the sample sizes of Fisher and Bushlow (2015) and Laghi et al. (2015) are relatively small (N = 44 and N = 36, respectively), and in the study of Fisher and Bushlow (2015), participants with other eating disorders than AN were included. Second, in the study of Fisher and Bushlow (2015), individuals with AN were assessed at a mean of 175 days since their first visit to the treatment centre. In the current study, assessment took place at the start of treatment or up to 4 weeks after intake. Thus, it is likely that the participants in the study of Fisher and Bushlow (2015) had a longer duration of illness than the participants in the current study, possibly explaining the differences in satisfaction rates. Unfortunately, no information on the moment of assessment or the duration of illness was provided by Laghi et al. (2015). Fourth, the inclusion of individuals with atypical AN potentially complicates findings of the current study, in light of possible differences in individuals with typical and atypical AN (Garber et al., 2019;Sawyer et al., 2016). Unfortunately, our current sample size does not allow us to reliably examine differences between individuals with typical and atypical AN. Yet, if the analyses were restricted to the individuals with typical AN, we do find similar patterns in which individuals with AN report lower satisfaction with normative life domains than individuals without an eating disorder. Fifth, the broad age range in the current study (12 to 22 years old) might also complicate our results. However, it is important to mention that 89.2% of all participants were aged 14 to 19 years old and that post hoc excluding participants outside of this age range did not affect the pattern of results.
Finally, it is unclear whether there are gender effects in the relation between AN and satisfaction with life domains, as participants in the current study were mostly female (97%). Previous research suggests that disordered eating may be differentially related to life satisfaction across men and women (Zullig et al., 2007). It may therefore be relevant for future research to examine whether satisfaction with normative life domains is also different in men and women with AN.
In conclusion, the present study showed that adolescents with broadly defined AN report lower satisfaction with normative life domains, compared to non-eating disordered individuals. Specifically, individuals with AN mainly seem to be less satisfied with themselves,