Skip to main content
NCBI home page
Wiley Open Access Collection logoLink to Wiley Open Access Collection
. 2025 Jan 31;58(4):789–801. doi: 10.1002/eat.24389

Exploring the Multifaceted Burdens and Experiences of Parents With a Child Diagnosed With Anorexia Nervosa: A Psychological Network Analysis

Michael Zeiler 1, Julia Philipp 1, Stefanie Truttmann 1, Konstantin Kopp 1, Gabriele Schöfbeck 1, Dunja Mairhofer 1, Hartmut Imgart 2, Annika Zanko 2, Ellen Auer‐Welsbach 3, Andreas Karwautz 1, Gudrun Wagner 1,✉
PMCID: PMC11969034  PMID: 39891395

ABSTRACT

Objective

Parents of children diagnosed with anorexia nervosa (AN) are facing multiple burdens when caring for their child. This study uses a psychological network approach to identify central factors among parental caregiving burdens and experiences which then can constitute promising targets for caregiver interventions.

Method

A total of 348 parents (247 mothers, 101 fathers) of children diagnosed with AN (96% female, 90% restrictive type) provided data including parental psychopathology, eating‐disorder related burden, high‐expressed emotion and perceived caregiver skills at baseline and after having participated in a parental skills training, multi‐family therapy or systemic family therapy. Regularized partial correlation networks including 14 variables were estimated for baseline and follow‐up data were estimated.

Results

High‐expressed emotion, particularly parental emotional overinvolvement emerged as the most central variable in the network showing substantial correlations with depression and low self‐care behavior. Emotional overinvolvement also functioned as a bridge variable between parental psychopathology, perceived caregiver skills, and eating disorder‐related burden. Moreover, parental criticism was strongly associated with burden due to the child's confrontational behavior and low levels of parental frustration tolerance. Network comparison tests neither revealed statistically significant differences in network structure and global network strength between baseline and follow‐up, nor between mothers and fathers.

Discussion

This study highlights the importance of parental high‐expressed emotion as a promising treatment target. Adding or intensifying intervention components promoting parental emotion regulation strategies as well as intensified training in Motivational Interviewing may be useful therapeutic approaches to decrease overall parental burden.

Keywords: anorexia nervosa, caregivers, caregiver skills, depression, eating disorders, eating disorder‐related burden, fathers, high‐expressed emotion, mothers, network analysis

Summary.

  • Parents caring for a child suffering from anorexia nervosa face numerous burdens and difficult situations.

  • This study used a statistical method called “psychological network analysis” to identify key factors among the many experiences and stressors they encounter in their caregiving role.

  • We found that being overly emotionally involved (e.g., intense worry about the affected child, ruminating about causes of the illness, self‐blame, thinking they may be becoming ill themselves, taking over full control) was a central factor for parents.

  • We concluded that parents could profit from emotional regulation trainings to alleviate their burden.

1. Introduction

There is consistent evidence that caring for a child diagnosed with anorexia nervosa (AN) puts complex burden on parents. The cognitive interpersonal maintenance model of AN (Schmidt and Treasure 2006; Treasure et al. 2020) proposes that the child's illness elicits strong emotional reactions in their parents which not only lead to maladaptive behavior toward the child (e.g., high‐expressed emotion), but also to psychological distress in themselves. Confidence in being able to manage the challenges arising from the child's illness, respectively high levels of caregiver skills, may mitigate parental distress.

Parental burden may arise from the awareness that the child's AN is potentially life‐threatening, the symptomatology of the child's eating disorder itself (e.g., binge‐purging behavior, comorbid psychopathology), the child's confrontational behavior (e.g., rigid behavior, mealtime arguments), subjective beliefs about the etiology (e.g., rumination about one's own role in the illness development associated with self‐blame) and social impacts of the disorder (e.g., social isolation) (Pehlivan et al. 2024; Sepulveda et al. 2008; van Hoeken and Hoek 2020). Parental eating‐disorder related burden have also been associated with high general parental distress, depressive symptoms and anxiety (Kyriacou, Treasure, and Schmidt 2008; Zeiler et al. 2023c; Zhang et al. 2021) and even post‐traumatic stress symptoms following the child's inpatient admission (Timko et al. 2023). While the majority of previous research was based on mothers, some studies have also demonstrated high levels of stress and burden in fathers of patients with AN, although to a lesser extent compared to mothers (Martín et al. 2013; Rhind et al. 2016; Zeiler et al. 2023a). Parental burden has also been linked to the lack of information about the course and treatment of the child's eating disorder (Kocsis‐Bogar et al. 2023) as well as to lower levels of perceived caregiver skills which include, for example, communicating about the eating disorders in the family, frustration tolerance, parental self‐care and bigger‐picture thinking (Hibbs et al. 2015b; Zeiler et al. 2021). Moreover, high‐expressed emotion has also been reported in parents of children with eating disorders. According to Rienecke (2018), high‐expressed emotion is not a measure of emotional expressiveness, but is defined as a set of a relative's attitudes and behaviors toward an ill family member including critical comments, hostility (often subsumed under critical comments) and emotional overinvolvement (Rienecke 2018). Critical comments or criticism comprise being too directive, negative and hostile toward the patient or blaming the patient (e.g., being angry with him/her, constantly insisting to change behavior, not being able to hold back with criticism with him/her). Emotional overinvolvement involves an overprotective and overinvolved caregiving style (e.g., suppressing own needs or not taking care of your own health due to the child's illness, constantly ruminating about the child's illness and causes of it, giving up important things in your own life). Thus, the concept of high‐expressed emotion describes a high level of (maladaptive) reactions toward the patient and the parents' own handling of the illness. High‐expressed emotion in parents has been linked to the level of eating disorder symptoms and worse treatment outcome in patients (Philipp et al. 2020; Rienecke 2018). Parental confidence in being able to effectively support their child with an eating disorder and to manage the challenges of the illness themselves (hereafter referred to as the “perceived level of caregiver skills”) appears particularly important given the pivotal role parents play in AN treatment. Often regarded as ‘co‐therapists’ (Altdorf et al. 2022) parents require a high level of caregiver skills to feel empowered in this role. Existing AN treatment approaches involving parents include parental specific skills trainings, for example based on the Maudsley approach (Hibbs et al. 2015a), Family‐based Treatment (Lock et al. 2024) and Multi‐Family Therapy (Baudinet et al. 2021; Dennhag, Henje, and Nilsson 2019). These approaches do not only aim at improving eating disorder outcomes of those affected by AN but also to reduce burden for parents and the entire family. However, in order to be able to optimally support parents in their caregiving role through targeted interventions, identifying key variables from the multiple burdens and experiences parents are faced with, seems to be highly important.

Psychological network analyses may be an innovative statistical method to model complex interactions between a large set of variables and to identify central characteristics in a network of associated variables (Hevey 2018). Thus, this method may be a useful way to identify central characteristics among factors known to impact parents caring for a child with AN. Once identified, such central characteristics may present promising targets for parental interventions as it is assumed that interventions improving these variables will likely have an impact on other associated variables in the network. Psychological network analyses have been increasingly conducted in the field of eating disorders while the focus was on identifying central symptoms among networks of psychopathological symptoms in patients with eating disorders (Tomei, Pieroni, and Tomba 2022). So far, only two studies applying network analyses including caregiver variables have been performed. A network analysis including a range of caregiver and patient variables revealed that the caregivers' level of depression and emotional overinvolvement were the most central variables in the network and that the caregivers' accommodation to the child's illness as well as the patients' level of depression functioned as bridge variables between caregiver and patient outcomes (Monteleone et al. 2023). Another network analysis focused on family functioning and patients' psychopathology and found that family communication was the node with the highest overall centrality and that problem‐solving as well as maturity fear, interpersonal insecurity an interpersonal alienation were bridge variables between family functioning and the child's eating disorder pathology (Monteleone et al. 2024).

The present study aims to expand this knowledge by performing a psychological network analysis of variables obtained in parents having a child with AN including variables assessing the eating disorder‐related caregiver burden, parental psychopathological symptoms, perceived caregiver skills and high‐expressed emotion. These groups of variables were selected as they are relevant for whether or not parents can optimally support their child diagnosed with AN and act as “co‐therapists.” Parents facing high eating disorder related burden, for example burden arising from self‐blame, difficulties during mealtimes and social isolation can hinder interaction with the child and adherence to treatment plans. In a similar manner, parents showing psychopathology such as depression or high stress levels themselves may face challenges in supporting their child effectively. Parental depression and distress levels may also affect high‐expressed emotion toward the child. Parental emotional overinvolvement and parental criticism may interfere with treatment success as they can exacerbate the child's symptoms. Furthermore, caregiver skills including frustration tolerance, emotional intelligence, and big‐picture thinking and self‐care are vital for managing the challenges of the illness. These abilities enable parents to remain resilient in stressful situations and maintain a long‐term perspective. These dimensions selected for the network analysis are in line with the cognitive interpersonal maintenance model of anorexia nervosa which highlights the interplay between them (Schmidt and Treasure 2006; Treasure et al. 2020).

In contrast to the research performed by Monteleone et al. (2023, 2024), this study focuses solely on caregiver variables allowing a more detailed view on which aspects of parental experiences when caring for a child with AN may be most central. In this regard, the primary objective of this study is to identify central variables in the network which can then shape the focus of future caregiver interventions. Secondary objectives are to explore whether the network and its central indices change from baseline to after having participated in a parental intervention (3‐month follow‐up) and to explore differences in the network structure and central indices between mothers and fathers which also expands the knowledge gained by previous studies. We did not have a specific expectation which variables could be more or less central at baseline or after a parental intervention. However, based on previous research showing that previous research showing that mothers of a child with an eating disorder had higher stress, anxiety and depression levels compared to fathers (Le Grange et al. 2011; Martín et al. 2013; Zeiler et al. 2023a), one may expect that affective components would be more central aspects in mothers. In contrast, the network model in fathers who often react less emotionally to the child's illness and tend to be more avoidant (Treasure 2018) may show higher centrality in other aspects, for example criticism and big picture thinking. If differences in the network structure and central indices are found, this may indicate that different factors are more or less relevant in various stages of caregiver support, respectively that different factors may be more relevant for either mothers or fathers. These findings may help to develop targeted and individualized interventions for parents.

2. Methods

2.1. Sample

This study is a secondary analysis of a quasi‐randomized trial evaluating the efficacy of a skills training for parents caring for a child with anorexia nervosa and studies evaluating the implementation of this intervention (Philipp et al. 2020; Truttmann et al. 2020; Zeiler et al. 2023a; Zeiler et al. 2023b). In the present study, we included data from 348 parents (247 mothers, 101 fathers) who provided data at baseline and after the participation in a parental intervention (3‐month follow‐up) while the great majority of participants (n = 316) received an 8‐week parental skills training (“SUCCEAT”; Franta et al. 2018) based on the Maudsley approach at the Department for Child and Adolescent Psychiatry, Medical University of Vienna and the Department for Neurology and Psychiatry of Children and Adolescents, Klagenfurt (Austria), 23 received multi‐family therapy for anorexia nervosa at the Parkland Clinic Bad Wildungen (Germany) and 9 received systemic family therapy (in Klagenfurt, Austria). Data collection took place between 2015 and 2022. The parents were 47.9 (SD: 5.1) years old on average, 57.6% had a university degree and 81.7% were married or lived in a partnership. The majority of parents were citizens of Austria (83.3%) or another country of the European Union (13.8%). 8.1% of parents reported that they were diagnosed with an eating disorder at any time point in their life and 13.4% had another lifetime psychiatric disorder. All patients whom the parents cared for were diagnosed with AN (90% restrictive type), 96.4% were female, 47.9% received inpatient and 52.1% outpatient treatment, the mean age was 14.9 (SD: 1.9) years (range: 10–23), the mean eating disorder duration was 13 months (SD: 11.0) and the average BMI was 15.8 (SD: 2.2).

2.2. Item Selection for the Network Analysis

The items considered for the psychological network analysis (subsequently named as “nodes”) were obtained using standardized self‐report questionnaires. The nodes represent the scale scores of these questionnaires.

Eating disorder related burden was assessed with the Eating Disorder Symptom Impact Scale (EDSIS; Sepulveda et al. 2008). A total of 24 items assessed on a 5‐point Likert scale are summed up to six scale scores representing burden in different areas (“guilt,” “social isolation,” “confrontational behavior,” “binge‐purge impacts,” “mealtime difficulties,” “illness awareness”) within the last month. Of note, the six‐factor model proposed by Coomber and King (2013) which was recently confirmed in the German version (Zeiler et al. 2023c) was used.

Parental psychopathology was assessed using three questionnaires: General psychological distress was obtained with the General Health Questionnaire (GHQ; Goldberg et al. 1997). In the 12‐item version, the items are assessed on a 4‐point scale and the total score represents the level of psychological distress during the previous weeks. Parental depression was obtained with the Beck Depression Inventory (BDI‐II, Hautzinger, Keller, and Kühner 2006) in which 21 items assessing depressive symptoms are rated on a 4‐point scale and the total score reflecting the level of depression during the preceding two weeks. The level of anxiety was measured with the State/Trait Anxiety Inventory (STAI; Laux et al. 1981) in which 20 item rated on a 4‐point scale are aggregated to a state anxiety score representing the level of inner tension and concerns during the last month. Only state anxiety was considered in the present network analysis.

The Caregiver Skills Scale (CASK; Hibbs et al. 2015b) was used to assess the perceived skills in caring for a person with an eating disorder in the last week (27 items rated on a scale from 0 to 100). The 6‐factor model which was also confirmed in the German version (Zeiler et al. 2021) comprises skills related to ‘bigger picture’ thinking, ‘self‐care’, avoiding repetitive nagging conversations (“biting‐your‐tongue”), accepting and managing negative emotions (“insight and acceptance”), discussing and managing feelings (emotional intelligence) and “frustration tolerance.” Higher scores represent higher perceived skills.

High‐expressed emotion was obtained with the Family Questionnaire (FQ; Wiedemann et al. 2002). Twenty items assessed on a 4‐point scale are aggregated to a “criticism”‐score representing the level of critical comments toward the patient and an “emotional overinvolvement” score representing the level of overprotective and emotionally over‐sensitive behavior recently.

More details about these items considered for the network analysis are shown in Table S1.

2.3. Missing Data

Considering data from all participants, selected nodes for the network analysis and time points (baseline, 3‐month follow‐up), only a minimal percentage of data (0.9%) were missing. The Little's MCAR test was statistically significant (χ2(582) = 968.193, p < 0.001) indicating that missing that were not completely at random. Missing data were handled using the expectation–maximization (EM) imputation method including all available questionnaire scores as predictors (assuming normal distribution and using 25 iterations). Due to the very low percentage of missing data, we regarded a single imputation method as sufficient.

2.4. Check for Topological Overlap Between Selected Nodes

We used the goldbricker algorithm implemented in the R function networktools (Jones 2017) to identify redundant nodes in the network which would limit the generalizability of the results. This function computes the proportions of correlations that significantly differ for each pair of nodes. In the present study, we used 0.25 as the threshold for inclusion and a p‐value of 0.05 for determining statistical significance. This means that the function returns a list of node pairs that differ in ≤ 25% of their correlations to other nodes in the network which are considered as redundant nodes. By applying the goldbricker algorithm, three pairs of nodes were identified to be redundant (1. BDI and STAI state scores, 2. CASK “Self Care” and CASK “Insight and Acceptance” scores, 3. CASK ‘Frustration Tolerance and CASK “Biting‐your‐tongue” scores). We thus decided to exclude one item of each pair from the network analysis, namely the STAI state score as well as the CASK ‘Insight and Acceptance’ and ‘Biting‐your‐tongue’ scores leaving a total of 14 nodes included in the final network analysis.

2.5. Network Estimation

We used the R packages qgraph (Epskamp et al. 2012; version 1.6.5) and bootnet (Epskamp, Borsboom, and Fried 2018; version 1.4.3) to estimate and visualize regularized partial correlation networks. In a first step, this was performed including the selected variables obtained at baseline. In a network, the included variables are presented as “nodes” which are connected by “edges” which represent the partial correlations between two nodes controlling for all other correlations in the network. In the network plot, positive associations are shown in green and negative association in red while the thickness of the edges represents the strength of the association. The regularized partial correlation networks are calculated using the graphical least absolute shrinking and selection operator (gLASSO; Friedman, Hastie, and Tibshirani 2008) in combination with the Extended Bayesian Information Criterion (EBIC; Chen and Chen 2008) which sets small or unstable correlations to zero and thus, produces more parsimonious and better interpretable networks (λ defining the level of network sparsity was set to 0.5). We first estimated the network for data obtained at baseline. Subsequently, we estimated the network for the 3‐months follow‐up assessment while in the network plot, coordinates of the baseline‐network were used; thus, differences in edge strengths are shown.

2.6. Measures of Node Centrality

The qgraph package (Epskamp et al. 2012) was used to calculate for common measures of node centrality including “strength,” “betweenness,” “closeness” and “one‐step expected influence.” Strength represents the weighted number and strength of all connections of a specific node, betweenness refers to the number of shortest paths that pass through a specific node, closeness quantifies the number of direct and indirect associations between a specific node to all other nodes. One‐step expected influence represents a similar measure of node strength, however, also takes the sign of edges into account and is preferable for networks containing positive and negative edges (Robinaugh, Millner, and McNally 2016). As we expect positive and negative associations between the nodes, interpretations of the results will mainly focus on this measure. In order to evaluate baseline to 3‐months FU changes in the network structure, we also estimated node centrality using the 3‐months FU data. The Network Comparison Test (NCT; van Borkulo et al. 2023) implemented in the NetworkComparisonTest package in R was conducted to analyze differences regarding the baseline versus 3‐month FU network structure and global network strength as well as differences in centrality indices of specific nodes. Although not an integral part of a network analysis, we additionally analyzed whether the scores of the variables included in the network model changed from baseline to follow‐up using paired t‐tests.

2.7. Network Stability and Edge Accuracy

We used the case‐dropping subset bootstrap approach implemented in the R package bootnet (Epskamp, Borsboom, and Fried 2018) to estimate the stability of central indices. We performed 1.000 bootstraps of each centrality index with progressively smaller subsamples of the data. The correlation stability (CS) coefficients that are calculated for each central index should be above 0.5; values ≥ 0.7 are regarded as excellent (Epskamp, Borsboom, and Fried 2018). This value represents the maximum proportion of cases that can be dropped from the full dataset so that the correlation between the originally calculated central indices and those retrieved from bootstrap subsets has a 95% probability of being r = 0.7 (respectively r = 0.5) or higher. Moreover, non‐parametric bootstrapping (1.000 boots) was performed to evaluate edge weights accuracy by calculating 95% confidence intervals.

2.8. Bridge Symptoms

Moreover, we evaluated whether there exist bridge symptoms/variables between different clusters (here called “communities”) of variables. We pre‐defined four communities: Variables measuring parental mental health, caregiver skills, eating‐disorder related burden and high‐expressed emotion. We used the bridge‐function of the networktools package (Jones 2017) to calculate the one‐step bridge expected influence. This centrality measure quantifies the summed edge weights of a node of interest and all other nodes that are not in the same community. Moreover, we used the NCT again to analyze differences in one‐step bridge expected influence of specific nodes.

2.9. Subgroup Analyses

We repeated the above‐described analyses (baseline data only) separately for parental sex to explore potential differences in the network structure and node centrality between mothers and fathers. A NCT was performed to analyze whether the network structure and global network strength differ between the mothers' and fathers' subsamples. Noteworthy, due to the relatively low number of fathers (n = 101), this analysis should be regarded as preliminary.

The analysis script is provided in the (Supporting Information).

3. Results

3.1. Network Structure at Baseline and 3‐Months Follow‐Up

Figure 1 shows the gLASSO network plots for baseline and 3‐months follow‐up data. High‐expressed emotion variables, namely emotional overinvolvement and criticism were placed in the central part of the plot indicating that these variables have many and strong connections to other nodes in the network. Emotional overinvolvement regarding the child's eating disorder was positively associated with parental depressive symptoms and social isolation as well as negatively with self‐care behavior. Parental criticism showed connections to the child's confrontational behavior as well as with high illness awareness and low frustration tolerance. Feelings of guilt with regard to the child's eating disorder were related to depressive symptoms. Low levels of parental self‐care behavior were associated with higher levels of depression. Variables assessing caregiver skills were clustered close to each other in the plot. The node ‘binge‐purge impacts’ was a peripheral node, likely due to the low number of patients with a binge‐purging subtype of AN in the sample.

FIGURE 1.

FIGURE 1

Open in a new tab

gLASSO networks of the entire sample at (a) baseline and (b) 3‐months follow‐up.

3.2. Node Centrality

Node centrality parameters are shown in Figure 2. The standardized centrality indices are also provided in Table S2 of the Supporting Information. Regarding node strength, emotional overinvolvement, criticism and depression emerged as the most central nodes in the baseline and follow‐up networks. Regarding one‐step expected influence taking into account the sign of the association, also emotional overinvolvement showed the highest centrality in both networks, followed by big picture thinking as well as confrontational behavior (at baseline) and illness awareness (at follow‐up). Additionally, emotional overinvolvement and criticism exhibited highest betweenness and closeness indicating that these variables act as a bridge between other nodes in the network and that many direct and indirect pathways go through these nodes.

FIGURE 2.

FIGURE 2

Open in a new tab

Centrality plot showing node centrality measures at baseline and 3‐months follow‐up (for explanations of node abbreviations, see legend in Figure 1).

3.3. Network Stability and Edge Accuracy

Case‐dropping subset bootstrapping (see Figure 3 for the baseline data and Figure S1 in the Supporting Information for the follow‐up data) demonstrated relatively high stability of central indices. The CS coefficients were as follows (baseline network: strength: 0.75, one‐step expected influence: 0.75, betweenness: 0.44, closeness: 0.67; follow‐up network: strength: 0.75, one‐step expected influence: 0.75, betweenness: 0.36, closeness: 0.60). This indicates most central indices can be meaningfully interpreted; however, betweenness should be interpreted with caution. Figure 4 (respectively Figure S2 in the Supporting Information) shows 95% confidence intervals for edges weights derived from non‐parametric bootstrapping. As the confidence intervals were not excessively large, they seem to be sufficiently accurate to be reasonably interpreted.

FIGURE 3.

FIGURE 3

Open in a new tab

Plot showing the stability of central indices derived from the case‐dropping subset bootstrap approach (based on baseline data).

FIGURE 4.

FIGURE 4

Open in a new tab

Edge accuracy plot showing 95% confidence intervals obtained from 1.000 bootstrap samples (based on baseline data).

3.4. Bridge Expected Influence

As displayed in Figure 5, emotional overinvolvement showed far the highest bridge expected influence in the baseline and follow‐up networks indicating that this variable acts at bridge between different communities of variables including parental general mental health, eating‐disorder related burden and perceived caregiver skills.

FIGURE 5.

FIGURE 5

Open in a new tab

Bridge expected influence calculated for (a) baseline (T0) and (b) 3‐months follow‐up (T1) data (for explanations of node abbreviations, see legend in Figure 1).

3.5. Differences Between Baseline (T0) and Follow‐Up (T1) Networks and Questionnaire Scores

The NCT revealed no statistically significant differences between the baseline and follow‐up networks regarding general network structure (p = 0.426) and global network strength (p = 0.857). Regarding differences in centrality indices, we observed lower strength of the GHQ at follow‐up (p = 0.039) and a similar trend for expected influence (p = 0.062) as well as lower closeness of the EDSIS guilt variable at follow‐up (p = 0.006). Moreover, lower bridge expected influence was shown for the GHQ at follow up (p = 0.021). No statistically significant differences in centrality indices were observed for any other variables in the network.

Regarding questionnaire scores, parental psychopathology (GHQ and BDI), eating disorder related burden (EDSIS subscales guilt, confrontational behavior, mealtime difficulties and illness awareness) and high‐expressed emotion (FFB Emotional overinvolvement, Criticism) significantly decreased and perceived caregiver skills (CASK all subscales) significantly increased from baseline to follow up (all p‐values > 0.001). No statistically significant changes were observed for the EDSIS social isolation (p = 0.341) and binge‐purge impacts (p = 0.087) subscales.

3.6. Comparison Between Mothers and Fathers

As sensitivity analysis, we compared the network estimated between mothers and fathers using the data obtained at baseline. Overall, we found a similar network solution and similar central indices for mothers and fathers resembling the overall network estimation presented above. The NCT showed no statically significant differences between the subsamples regarding overall network structure (p = 0.439) and global network strength (p = 0.193). For centrality indices, we just observed higher strength (p = 0.037) and expected influence (p = 0.029) of the EDSIS illness awareness in mothers compared to fathers. Visual exploration of the network plots in mothers and fathers (see Figure S3 in the Supporting Information) indicates that there may be a tendency of stronger associations in mothers compared to fathers between illness awareness and mealtime difficulties, between guilt and depression/general distress/ social isolation as well as between criticism and confrontational behavior/frustration tolerance. However, the network solution for fathers needs to be interpreted with caution as the CS‐coefficients for the fathers' subsample were weak (strength: 0.29, one‐step expected influence: 0.21, betweenness: 0.29, closeness: 0.13) compared to the subsample of mothers (strength: 0.67, one‐step expected influence: 0.60, betweenness: 0.44, closeness: 0.52). For details on the central indices, network stability and edge accuracy for both subsamples, see Table S3 and Figures S4–S6.

4. Discussion

This study was designed to shed light on central factors among the multifaced burden and experiences of parents caring for a child diagnosed with AN. Parental emotional overinvolvement emerged as the most central variable having strong connections to other variables in the network and serving as a bridge between symptoms of parental psychopathology, parental burden and perceived caregiver skills. Moreover, although to a lower extent, also parental criticism and depression, burden through the child's confrontational behavior and awareness about the illness severity showed substantial associations to other nodes in the network.

Parental emotional overinvolvement, along with caregiver depression, has also been identified as a central variable in the network analysis conducted by Monteleone et al. (2023). The results from the present study confirm the key role that emotional overinvolvement plays among caregivers' experiences and coping with the eating disorder of the child. As outlined above, emotional overinvolvement comprises parental overprotective and emotionally over‐sensitive behavior (Wiedemann et al. 2002). As observed in our study, parental emotion overinvolvement has been strongly associated with parental psychopathology, including symptoms of depression and anxiety (Kyriacou, Treasure, and Schmidt 2008; Schwarte et al. 2017). Furthermore, parental criticism when communicating to the affected child has been higher when parents felt burdened by the child's confrontational behavior (e.g., comprising verbal aggression, out of control temper, rigid behavior). This is consistent with other studies having found profound associations between criticism and child's difficult behavior (Blondin et al. 2019; Kyriacou, Treasure, and Schmidt 2008). In the present study high levels of criticism were also linked to low frustration tolerance when caring for the child. In this regard, high‐expressed emotion can be regarded as an important target in caregiver interventions. This seems all the more relevant as familiar high‐expressed emotion predicted worse treatment outcomes such as higher eating disorder pathology and lower chance of remission in patients (Allan et al. 2018; Rienecke et al. 2017). Moreover, it has also been listed as an important factor contributing to the maintenance of the eating disorder in the cognitive interpersonal maintenance model of anorexia nervosa (Schmidt et al. 2016; Treasure et al. 2020).

Notably, emotional overinvolvement was also the most central variable in the network after parents have participated in an 8‐week skills training or other type of family therapy. In overall, the network structure and centrality indices of most nodes in the network model remained rather stable between baseline and 3‐months follow‐up. This result must be interpreted considering that the level of parental psychopathology, eating disordered related burden and high‐expressed emotion significantly decreased and perceived caregiver skills significantly increased. The effectiveness of parental trainings in reducing parental burden and high‐expressed emotion have also been demonstrated in previous studies (Dimitropoulos et al. 2019; Hibbs et al. 2015a; Moskovich et al. 2017; Philipp et al. 2020; Zeiler et al. 2023a). At first glance, these findings (almost no change in centrality indices in the network model including emotional overinvolvement as the most central node, but significant improvements in questionnaire scores) may appear contradictory. However, the most plausible interpretation is that, while the level of emotional overinvolvement decreased over the course of parental intervention, it remained a central characteristic relative to other assessed variables. Thus, although the parental skills training “SUCCEAT” (Franta et al. 2018) covers content and exercises to reduce parental high‐expressed emotion in context with the child's eating disorder, this was one of many topics addressed in the intervention. The present study indicates that probably too little focus was placed on this aspect and that addressing parental high‐expressed emotion would further improve the effectiveness of parental interventions. For example, parents with extreme emotional overinvolvement may benefit from an individual emotion regulation training in addition to group‐based intervention to tackle their own emotional difficulties in coping with the child's illness. This was, for example, offered for caregivers of adults with psychiatric and neurocognitive disorders (Behrouian et al. 2020, 2021; Moskowitz et al. 2019). Furthermore, parents with high criticism may benefit from an extended individual training in Motivational Interviewing (Miller and Rollnick 2002) additional to the Motivational Interviewing exercises performed in the SUCCEAT group sessions. The latter may be particularly supportive when the affected child shows high confrontational behavior toward the parents.

In addition to high‐expressed emotion, parental depression and big picture thinking emerged as rather central in some of the central indices. Parental symptoms of depression have been also emphasized as an important treatment target by Monteleone et al. (2023) and the high level of depression among parents of children with eating disorders is evident (Anastasiadou et al. 2014). Particularly as parental depression was associated with low self‐care behavior and feelings of guilt related to the child's eating disorder, psychotherapeutic treatment for parents experiencing high psychopathology themselves should be considered. Based on network modeling in our study, it is suggested that reducing parental psychopathology may also affect the parent's reaction to the child's illness, such as emotional overinvolvement. Although we did not have the information about how many parents received individual psychotherapy during the study period, we expect this percentage to be very low. However, individual psychotherapy should be offered for parents facing high psychopathology themselves such as depressive symptoms or anxiety. Indeed, anecdotal evidence from parents shows that some of them started individual psychotherapy after the group‐based SUCCEAT training to cope with own psychopathology and burden in a deeper way. Targeted support for mothers and fathers to meet their own needs and decrease their own psychopathology seems crucial to enhance their capacity to be effective co‐therapists in the child's treatment.

Moreover, in a similar way in that neurocognitive trainings are addressing central coherence, respectively bigger picture thinking, in patients with AN, e.g. in Cognitive Remediation Therapy (Giombini et al. 2022; Tchanturia et al. 2017), also parents may profit from exercises improving bigger picture thinking. For example, such exercises may promote the ability to separate the illness from the child, appreciate also small steps toward recovery and strengthen hope that the child will recover. As bigger picture thinking was closely related to frustration tolerance in the present study, promoting this skill may also help the parents coping with disappointing or discouraging situation when caring for their child. This skill may be also addressed in joint sessions between parents and the child, such as done in Family‐based Therapy (Lock and Nicholls 2020).

Moreover, we found that the network structure and central indices were similar in mothers and fathers. However, this finding must be interpreted with caution due to the low sample size and low stability of central indices in the subsample of fathers and needs validation in future studies. However, this is the first study indicating that fathers show a similar pattern of associations between variables assessing caregiver burden and experiences compared to mothers. Recently, a study showed that also fathers, although to a lower extent than mothers, show high burden due to the child's eating disorder and that both parents can profit from a caregiver skills training (Zeiler et al. 2023a). Taken together, the present study provides further evidence for the usefulness and importance of providing caregiver interventions to both parents.

4.1. Strengths and Limitations

The inclusion of mothers and fathers instead of focusing on primary caregivers only as well as the inclusion of two timepoints (prior and after parents having received a skills training or other type of family therapy) can be considered as strengths of this study. However, as mentioned beforehand, the relatively low sample size of fathers is a limitation as the network estimation for this subgroup may not be valid. Moreover, we used a quite short follow‐up period of 3 months. Thus, it cannot be ruled out that we would have seen more profound longitudinal differences in the network structure when a longer follow‐up period had been used. Furthermore, as in each network analysis, the network estimation depends on the variables that are considered as nodes and finally included in the model. The inclusion of further or different caregiver variables could have resulted in a different network solution. For example, parental accommodating behavior to the child's illness which was not obtained in the present study has been identified as a central node in a previous network analysis (Monteleone et al. 2023). Moreover, it should be noted that the assessed variables do not refer to the exact same timeframe. While the CASK questionnaire refers to perceived caregiver skills in the last week, the BDI assesses depressive symptoms during the previous two weeks, the EDSIS refers to burden during the last months, the FBB assesses high‐expressed emotion occurring “recently,” and the GHQ obtains psychological distress during the “previous weeks.” However, despite this inconsistency, all measures included caregiver burden and experiences made in a period relatively close to when the assessment was carried out. Finally, the sample included relatively highly educated parents with a high proportion of parents holding a university degree (possibly since the majority of the sample was recruited in large urban regions) and parents living in partnerships. Thus, generalizability of the results for parents with a lower level of education and parents in non‐intact families should be evaluated in future studies.

4.2. Conclusions

This psychological network analysis provides evidence that parental high‐expressed emotion, particularly emotional overinvolvement, constitutes a central factor among the multifaceted burdens of parents caring for a child with AN and thus, is suggested a promising target in caregiver interventions. Adding emotion regulation trainings to parents facing high emotional overinvolvement as well as extended trainings in Motivational Interviewing for those showing high criticism toward the affected child support parents coping with the child's illness. High‐expressed emotion was closely connected to parental psychopathology, burden arising from child's confrontational behavior and awareness of high illness severity as well as frustration tolerance and bigger picture thinking. Thus, we suggest that interventions succeeding to reduce high‐expressed emotion would likely also affect these associated variables in a positive way. Moreover, strengthening appropriate support and skills to meet the parents' own needs is important to enhance their capacity to be effective co‐therapists in the treatment of the child's illness. Validation studies involving larger samples sizes, however, are needed to reveal potential gender differences between network structures of mothers and fathers.

Author Contributions

Michael Zeiler: conceptualization, data curation, formal analysis, methodology, visualization, writing – original draft. Julia Philipp: conceptualization, investigation, project administration. Stefanie Truttmann: conceptualization, investigation, project administration. Konstantin Kopp: investigation. Gabriele Schöfbeck: investigation, resources. Dunja Mairhofer: investigation. Hartmut Imgart: investigation, resources. Annika Zanko: investigation, project administration. Ellen Auer‐Welsbach: investigation, project administration, resources. Andreas Karwautz: conceptualization, funding acquisition, resources, supervision, writing – review and editing. Gudrun Wagner: conceptualization, funding acquisition, resources, supervision, writing – review and editing.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Data S1. R_Syntax.

EAT-58-789-s002.docx (18.8KB, docx)

Data S2.

Table S1. Detailed description of the items considered for the network analysis.

Table S2. Standardized centrality indices of the EBIC gLASSO network for the entire sample at baseline and 3‐months follow‐up.

Table S3. Centrality indices by parental sex (based on baseline data).

Figure S1. Stability of central indices derived from the case‐dropping subset bootstrap approach (entire dataset, 3‐month follow‐up).

Figure S2. Edge accuracy plot showing 95% confidence intervals obtained from 1.000 bootstrap samples (entire dataset, 3‐months follow‐up data).

Figure S3. Network plots in the subsample of (a) mothers and (b) fathers (based on baseline data).

Figure S4. Centrality plot showing standardized centrality measures in the subsample of mothers and fathers (based on baseline data).

Figure S5. Stability of central indices derived from the case‐dropping subset bootstrap approach for the subgroups of (a) mothers and (b) fathers.

Figure S6. Edge accuracy plot showing 95% confidence intervals obtained from 1.000 bootstrap samples for the subgroups of (a) mothers and (b) fathers.

EAT-58-789-s001.docx (178.9KB, docx)

Acknowledgments

Parts of the data collection for this study were funded by Gemeinsame Gesundheitsziele – Pharma Master Agreement (a cooperation between the Austrian pharmaceutical industry and the Austrian social insurance): reference: #99901002500.

Action Editor: Jake Linardon

Funding: This work was supported by “Gemeinsame Gesundheitsziele – Pharma Master Agreement” (a cooperation between the Austrian pharmaceutical industry and the Austrian social insurance), 99901002500.

Andreas Karwautz and Gudrun Wagner are senior authors.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author (G.W.) upon reasonable request.

References

  1. Allan, E. , Le Grange D., Sawyer S. M., McLean L. A., and Hughes E. K.. 2018. “Parental Expressed Emotion During Two Forms of Family‐Based Treatment for Adolescent Anorexia Nervosa.” European Eating Disorders Review 26, no. 1: 46–52. 10.1002/erv.2564. [DOI] [PubMed] [Google Scholar]
  2. Altdorf, S. , Dempfle A., Heider K., Seitz J., Herpertz‐Dahlmann B., and Dahmen B.. 2022. “Parents as co‐Therapists in Home Treatment for Adolescents With Anorexia Nervosa—Factors and Mechanisms.” Praxis der Kinderpsychologie und Kinderpsychiatrie 71, no. 5: 467–486. 10.13109/prkk.2022.71.5.467. [DOI] [PubMed] [Google Scholar]
  3. Anastasiadou, D. , Medina‐Pradas C., Sepulveda A. R., and Treasure J.. 2014. “A Systematic Review of Family Caregiving in Eating Disorders.” Eating Behaviors 15, no. 3: 464–477. 10.1016/j.eatbeh.2014.06.001. [DOI] [PubMed] [Google Scholar]
  4. Baudinet, J. , Eisler I., Dawson L., Simic M., and Schmidt U.. 2021. “Multi‐Family Therapy for Eating Disorders: A Systematic Scoping Review of the Quantitative and Qualitative Findings.” International Journal of Eating Disorders 54, no. 12: 2095–2120. 10.1002/eat.23616. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Behrouian, M. , Ramezani T., Dehghan M., Sabahi A., and Ebrahimnejad Zarandi B.. 2020. “The Effect of Emotion Regulation Training on Stress, Anxiety, and Depression in Family Caregivers of Patients With Schizophrenia: A Randomized Controlled Trial.” Community Mental Health Journal 56, no. 6: 1095–1102. 10.1007/s10597-020-00574-y. [DOI] [PubMed] [Google Scholar]
  6. Behrouian, M. , Ramezani T., Dehghan M., Sabahi A., and Ebrahimnejad Zarandi B.. 2021. “The Effect of the Emotion Regulation Training on the Resilience of Caregivers of Patients With Schizophrenia: A Parallel Randomized Controlled Trial.” BMC Psychology 9, no. 1: 39. 10.1186/s40359-021-00542-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Blondin, S. , Meilleur D., Taddeo D., and Frappier J.‐Y.. 2019. “Caregiving Experience and Expressed Emotion Among Parents of Adolescents Suffering From Anorexia Nervosa Following Illness Onset.” Eating Disorders 27, no. 5: 453–470. 10.1080/10640266.2018.1553431. [DOI] [PubMed] [Google Scholar]
  8. Chen, J. , and Chen Z.. 2008. “Extended Bayesian Information Criteria for Model Selection With Large Model Spaces.” Biometrika 95, no. 3: 759–771. 10.1093/biomet/asn034. [DOI] [Google Scholar]
  9. Coomber, K. , and King R. M.. 2013. “An Investigation of the Psychometric Properties of the Eating Disorder Symptom Impact Scale Within an Australian Sample.” Australian Journal of Psychology 65, no. 2: 71–78. 10.1111/j.1742-9536.2012.00057.x. [DOI] [Google Scholar]
  10. Dennhag, I. , Henje E., and Nilsson K.. 2019. “Parental Caregiver Burden and Recovery of Adolescent Anorexia Nervosa After Multi‐Family Therapy.” Eating Disorders 29, no. 5: 463–479. 10.1080/10640266.2019.1678980. [DOI] [PubMed] [Google Scholar]
  11. Dimitropoulos, G. , Landers A., Freeman V., Novick J., Schmidt U., and Olmsted M.. 2019. “A Feasibility Study Comparing a Web‐Based Intervention to a Workshop Intervention for Caregivers of Adults With Eating Disorders.” European Eating Disorders Review 27, no. 6: 641–654. 10.1002/erv.2678. [DOI] [PubMed] [Google Scholar]
  12. Epskamp, S. , Borsboom D., and Fried E. I.. 2018. “Estimating Psychological Networks and Their Accuracy: A Tutorial Paper.” Behavior Research Methods 50, no. 1: 195–212. 10.3758/s13428-017-0862-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Epskamp, S. , Cramer A. O. J., Waldorp L. J., Schmittmann V. D., and Borsboom D.. 2012. “Qgraph: Network Visualizations of Relationships in Psychometric Data.” Journal of Statistical Software 48, no. 4: 1–18. 10.18637/jss.v048.i04. [DOI] [Google Scholar]
  14. Franta, C. , Philipp J., Waldherr K., et al. 2018. “Supporting Carers of Children and Adolescents With Eating Disorders in Austria (SUCCEAT): Study Protocol for a Randomised Controlled Trial.” European Eating Disorders Review 26, no. 5: 447–461. 10.1002/erv.2600. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Friedman, J. , Hastie T., and Tibshirani R.. 2008. “Sparse Inverse Covariance Estimation With the Graphical Lasso.” Biostatistics 9, no. 3: 432–441. 10.1093/biostatistics/kxm045. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Giombini, L. , Nesbitt S., Kusosa R., et al. 2022. “Neuropsychological and Clinical Findings of Cognitive Remediation Therapy Feasibility Randomised Controlled Trial in Young People With Anorexia Nervosa.” European Eating Disorders Review 30, no. 1: 50–60. 10.1002/erv.2874. [DOI] [PubMed] [Google Scholar]
  17. Goldberg, D. P. , Gater R., Sartorius N., et al. 1997. “The Validity of Two Versions of the GHQ in the WHO Study of Mental Illness in General Health Care.” Psychological Medicine 27, no. 1: 191–197. 10.1017/s0033291796004242. [DOI] [PubMed] [Google Scholar]
  18. Hautzinger, M. , Keller F., and Kühner C.. 2006. Beck Depressions Inventar. 2. Auflag. (BDI‐II). Hallbergmoos: Pearson. [Google Scholar]
  19. Hevey, D. 2018. “Network Analysis: A Brief Overview and Tutorial.” Health Psychology and Behavioral Medicine 6, no. 1: 301–328. 10.1080/21642850.2018.1521283. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Hibbs, R. , Rhind C., Leppanen J., and Treasure J.. 2015a. “Interventions for Caregivers of Someone With an Eating Disorder: A Meta‐Analysis.” International Journal of Eating Disorders 48, no. 4: 349–361. 10.1002/eat.22298. [DOI] [PubMed] [Google Scholar]
  21. Hibbs, R. , Rhind C., Salerno L., et al. 2015b. “Development and Validation of a Scale to Measure Caregiver Skills in Eating Disorders.” International Journal of Eating Disorders 48, no. 3: 290–297. 10.1002/eat.22362. [DOI] [PubMed] [Google Scholar]
  22. Jones, P. J. 2017. “Networktools: Assorted Tools for Identifying Important Nodes in Networks. R Package Version 1.5.0. [Computer Software].” https://CRAN.R‐project.org/package=networktools.
  23. Kocsis‐Bogar, K. , Ossege M., Aigner M., Wancata J., and Friedrich F.. 2023. “Involvement, Depressive Symptoms, and Their Associations With Problems and Unmet Needs in Caregivers of Adult Eating Disorder Patients.” Eating and Weight Disorders 28, no. 1: 45. 10.1007/s40519-023-01572-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Kyriacou, O. , Treasure J., and Schmidt U.. 2008. “Understanding How Parents Cope With Living With Someone With Anorexia Nervosa: Modelling the Factors That Are Associated With Carer Distress.” International Journal of Eating Disorders 41, no. 3: 233–242. 10.1002/eat.20488. [DOI] [PubMed] [Google Scholar]
  25. Laux, L. , Glanzmann P., Schaffner P., and Spielsberger C. D.. 1981. STAI ‐ Das State‐Trait‐Angstinventar. Weinheim: Beltz Test. [Google Scholar]
  26. Le Grange, D. , Hoste R. R., Lock J., and Bryson S. W.. 2011. “Parental Expressed Emotion of Adolescents With Anorexia Nervosa: Outcome in Family‐Based Treatment.” International Journal of Eating Disorders 44, no. 8: 731–734. 10.1002/eat.20877. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Lock, J. , and Nicholls D.. 2020. “Toward A Greater Understanding of the Ways Family‐Based Treatment Addresses the Full Range of Psychopathology of Adolescent Anorexia Nervosa.” Frontiers in Psychiatry 10: 968. 10.3389/fpsyt.2019.00968. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Lock, J. D. , Le Grange D., Bohon C., Matheson B., and Jo B.. 2024. “Who Responds to an Adaptive Intervention for Adolescents With Anorexia Nervosa Being Treated With Family‐Based Treatment? Outcomes From a Randomized Clinical Trial.” Journal of the American Academy of Child and Adolescent Psychiatry 63, no. 6: 605–614. 10.1016/j.jaac.2023.10.012. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Martín, J. , Padierna A., Aguirre U., González N., Muñoz P., and Quintana J. M.. 2013. “Predictors of Quality of Life and Caregiver Burden Among Maternal and Paternal Caregivers of Patients With Eating Disorders.” Psychiatry Research 210, no. 3: 1107–1115. 10.1016/j.psychres.2013.07.039. [DOI] [PubMed] [Google Scholar]
  30. Miller, W. R. , and Rollnick S.. 2002. Motivational Interviewing: Preparing for Change. 2nd ed. New York, NY: Guilford. [Google Scholar]
  31. Monteleone, A. M. , Cascino G., Salerno L., et al. 2023. “A Network Analysis in Adolescent Anorexia Nervosa Exploring the Connection Between Both Patient and Carer Reactions and Outcome.” European Eating Disorders Review 31, no. 1: 65–75. 10.1002/erv.2933. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Monteleone, A. M. , Marchetto C., Cascino G., et al. 2024. “The Bidirectional Connection Between Family Functioning and Psychopathology: A Network Analysis in a Large Sample of Adolescents With Anorexia Nervosa and Their Parents.” Family Process 63: 2229–2242. 10.1111/famp.12983. [DOI] [PubMed] [Google Scholar]
  33. Moskovich, A. A. , Timko C. A., Honeycutt L. K., Zucker N. L., and Merwin R. M.. 2017. “Change in Expressed Emotion and Treatment Outcome in Adolescent Anorexia Nervosa.” Eating Disorders 25, no. 1: 80–91. 10.1080/10640266.2016.1255111. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Moskowitz, J. T. , Cheung E. O., Snowberg K. E., et al. 2019. “Randomized Controlled Trial of a Facilitated Online Positive Emotion Regulation Intervention for Dementia Caregivers.” Health Psychology 38, no. 5: 391–402. 10.1037/hea0000680. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Pehlivan, M. J. , Rodgers B., Schlage J., Maguire S., and Miskovic‐Wheatley J.. 2024. “Characteristics, Correlates of Burden and Support Service Use of a Help‐Seeking Carers of Loved Ones With an Eating Disorder.” European Eating Disorders Review 32: 458–475. 10.1002/erv.3059. [DOI] [PubMed] [Google Scholar]
  36. Philipp, J. , Truttmann S., Zeiler M., et al. 2020. “Reduction of High Expressed Emotion and Treatment Outcomes in Anorexia Nervosa—Caregivers' and Adolescents' Perspective.” Journal of Clinical Medicine 9, no. 7: 2021. 10.3390/jcm9072021. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Rhind, C. , Salerno L., Hibbs R., et al. 2016. “The Objective and Subjective Caregiving Burden and Caregiving Behaviours of Parents of Adolescents With Anorexia Nervosa.” European Eating Disorders Review 24, no. 4: 310–319. 10.1002/erv.2442. [DOI] [PubMed] [Google Scholar]
  38. Rienecke, R. D. 2018. “Expressed Emotion and Eating Disorders: An Updated Review.” Current Psychiatry Reviews 14, no. 2: 84–98. 10.2174/1573400514666180808115637. [DOI] [Google Scholar]
  39. Rienecke, R. D. , Lebow J., Lock J., and Le Grange D.. 2017. “Family Profiles of Expressed Emotion in Adolescent Patients With Anorexia Nervosa and Their Parents.” Journal of Clinical Child and Adolescent Psychology: The Official Journal for the Society of Clinical Child and Adolescent Psychology, American Psychological Association, Division 53 46, no. 3: 428–436. 10.1080/15374416.2015.1030755. [DOI] [PubMed] [Google Scholar]
  40. Robinaugh, D. J. , Millner A. J., and McNally R. J.. 2016. “Identifying Highly Influential Nodes in the Complicated Grief Network.” Journal of Abnormal Psychology 125, no. 6: 747–757. 10.1037/abn0000181. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Schmidt, U. , Ryan E. G., Bartholdy S., et al. 2016. “Two‐Year Follow‐Up of the MOSAIC Trial: A Multicenter Randomized Controlled Trial Comparing Two Psychological Treatments in Adult Outpatients With Broadly Defined Anorexia Nervosa.” International Journal of Eating Disorders 49, no. 8: 793–800. 10.1002/eat.22523. [DOI] [PubMed] [Google Scholar]
  42. Schmidt, U. , and Treasure J.. 2006. “Anorexia Nervosa: Valued and Visible. A Cognitive‐Interpersonal Maintenance Model and Its Implications for Research and Practice.” British Journal of Clinical Psychology 45, no. 3: 343–366. 10.1348/014466505X53902. [DOI] [PubMed] [Google Scholar]
  43. Schwarte, R. , Timmesfeld N., Dempfle A., et al. 2017. “Expressed Emotions and Depressive Symptoms in Caregivers of Adolescents With First‐Onset Anorexia Nervosa—A Long‐Term Investigation Over 2.5 Years.” European Eating Disorders Review 25, no. 1: 44–51. 10.1002/erv.2490. [DOI] [PubMed] [Google Scholar]
  44. Sepulveda, A. R. , Whitney J., Hankins M., and Treasure J.. 2008. “Development and Validation of an Eating Disorders Symptom Impact Scale (EDSIS) for Carers of People With Eating Disorders.” Health and Quality of Life Outcomes 6, no. 1: 28. 10.1186/1477-7525-6-28. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Tchanturia, K. , Giombini L., Leppanen J., and Kinnaird E.. 2017. “Evidence for Cognitive Remediation Therapy in Young People With Anorexia Nervosa: Systematic Review and Meta‐Analysis of the Literature.” European Eating Disorders Review 25, no. 4: 227–236. 10.1002/erv.2522. [DOI] [PubMed] [Google Scholar]
  46. Timko, C. A. , Dennis N. J., Mears C., Rodriguez D., Fitzpatrick K. K., and Peebles R.. 2023. “Post‐Traumatic Stress Symptoms in Parents of Adolescents Hospitalized With Anorexia Nervosa.” Eating Disorders 31, no. 3: 212–224. 10.1080/10640266.2022.2099604. [DOI] [PubMed] [Google Scholar]
  47. Tomei, G. , Pieroni M. F., and Tomba E.. 2022. “Network Analysis Studies in Patients With Eating Disorders: A Systematic Review and Methodological Quality Assessment.” International Journal of Eating Disorders 55, no. 12: 1641–1669. 10.1002/eat.23828. [DOI] [PubMed] [Google Scholar]
  48. Treasure, J. 2018. “Inclusion of Fathers in the Treatment of Eating Disorders.” Lancet Child & Adolescent Health 2, no. 6: 385–387. 10.1016/S2352-4642(18)30136-6. [DOI] [PubMed] [Google Scholar]
  49. Treasure, J. , Willmott D., Ambwani S., et al. 2020. “Cognitive Interpersonal Model for Anorexia Nervosa Revisited: The Perpetuating Factors That Contribute to the Development of the Severe and Enduring Illness.” Journal of Clinical Medicine 9, no. 3: 630. 10.3390/jcm9030630. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Truttmann, S. , Philipp J., Zeiler M., et al. 2020. “Long‐Term Efficacy of the Workshop vs. Online SUCCEAT (Supporting Carers of Children and Adolescents With Eating Disorders) Intervention for Parents: A Quasi‐Randomised Feasibility Trial.” Journal of Clinical Medicine 9, no. 6: 1912. 10.3390/jcm9061912. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. van Borkulo, C. , van Bork R., Boschloo L., et al. 2023. “Comparing Network Structures on Three Aspects: A Permutation Test.” Psychological Methods 28, no. 6: 1273–1285. 10.1037/met0000476. [DOI] [PubMed] [Google Scholar]
  52. van Hoeken, D. , and Hoek H. W.. 2020. “Review of the Burden of Eating Disorders: Mortality, Disability, Costs, Quality of Life, and Family Burden.” Current Opinion in Psychiatry 33, no. 6: 521–527. 10.1097/YCO.0000000000000641. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Wiedemann, G. , Rayki O., Feinstein E., and Hahlweg K.. 2002. “The Family Questionnaire: Development and Validation of a New Self‐Report Scale for Assessing Expressed Emotion.” Psychiatry Research 109, no. 3: 265–279. 10.1016/s0165-1781(02)00023-9. [DOI] [PubMed] [Google Scholar]
  54. Zeiler, M. , Philipp J., Truttmann S., et al. 2021. “A German Version of the Caregiver Skills Scale for Caregivers of Patients With Anorexia Nervosa.” European Eating Disorders Review 29, no. 2: 257–268. 10.1002/erv.2817. [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Zeiler, M. , Philipp J., Truttmann S., et al. 2023a. “Fathers in the Spotlight: Parental Burden and the Effectiveness of a Parental Skills Training for Anorexia Nervosa in Mother–Father Dyads.” Eating and Weight Disorders ‐ Studies on Anorexia, Bulimia and Obesity 28, no. 1: 65. 10.1007/s40519-023-01597-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Zeiler, M. , Schneider A., Philipp J., et al. 2023b. “Psychological Distress and Caregiving Experience During the First Two Years of the COVID‐19 Pandemic Among Parents of an Offspring With Anorexia Nervosa.” European Eating Disorders Review 31, no. 4: 529–538. 10.1002/erv.2976. [DOI] [PubMed] [Google Scholar]
  57. Zeiler, M. , Truttmann S., Philipp J., et al. 2023c. “An Investigation of the Factor Structure of a German Version of the Eating Disorder Symptom Impact Scale (EDSIS) Among Parents of Adolescents and Young Adults With Anorexia Nervosa.” Eating Behaviors 48: 101695. 10.1016/j.eatbeh.2022.101695. [DOI] [PubMed] [Google Scholar]
  58. Zhang, L. , Wu M. T., Guo L., et al. 2021. “Psychological Distress and Associated Factors of the Primary Caregivers of Offspring With Eating Disorder During the Coronavirus Disease 2019 Pandemic.” Journal of Eating Disorders 9, no. 1: 58. 10.1186/s40337-021-00405-9. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Data S1. R_Syntax.

EAT-58-789-s002.docx (18.8KB, docx)

Data S2.

Table S1. Detailed description of the items considered for the network analysis.

Table S2. Standardized centrality indices of the EBIC gLASSO network for the entire sample at baseline and 3‐months follow‐up.

Table S3. Centrality indices by parental sex (based on baseline data).

Figure S1. Stability of central indices derived from the case‐dropping subset bootstrap approach (entire dataset, 3‐month follow‐up).

Figure S2. Edge accuracy plot showing 95% confidence intervals obtained from 1.000 bootstrap samples (entire dataset, 3‐months follow‐up data).

Figure S3. Network plots in the subsample of (a) mothers and (b) fathers (based on baseline data).

Figure S4. Centrality plot showing standardized centrality measures in the subsample of mothers and fathers (based on baseline data).

Figure S5. Stability of central indices derived from the case‐dropping subset bootstrap approach for the subgroups of (a) mothers and (b) fathers.

Figure S6. Edge accuracy plot showing 95% confidence intervals obtained from 1.000 bootstrap samples for the subgroups of (a) mothers and (b) fathers.

EAT-58-789-s001.docx (178.9KB, docx)

Data Availability Statement

The data that support the findings of this study are available from the corresponding author (G.W.) upon reasonable request.

Articles from The International Journal of Eating Disorders are provided here courtesy of Wiley

RESOURCES