Network-based methods for psychometric data of eating disorders: A systematic review

Structure and factorial invariance of a brief version of the Eating Attitudes Test in Peruvian university students

Background

University students often experience significant changes in their eating habits, which can increase the risk of developing eating disorders (ED). This situation calls for the creation of brief assessment tools to identify college students who may be most at risk. The aim of the study was to determine the psychometric properties of the Eating Attitudes Test-8 (EAT-8) in a Peruvian university population; additionally, the possible differences in the scores of the instrument according to sociodemographic variables, such as gender and age, were examined.

Methods

A psychometric study was conducted on 610 participants (M = 24.3, SD = 2.16, and 61.5% female), aged 19 to 31 years, belonging to four universities of different professional careers.

Results

The unidimensional eight-item model was found to have fit indices that confirm acceptable factorial validity (X2/df = 3.23, CFI = 0.984, TLI = 0.977, RMSEA = 0.061, SRMR = 0.049) and an internal consistency of 0.833 for the Cronbach's alpha coefficient and 0.838 for the McDonald's omega coefficient. In addition, the EAT-8 was reported to be invariant according to gender and age; likewise, there were no significant differences in the age and gender categories.

Conclusion

The EAT-8 has solid psychometric properties, including validity, reliability, and invariance, in the Peruvian university population, which supports its ability to assess the risk of developing ED in this specific group.

Central psychological symptoms from a network analysis of patients with anxiety, somatoform or personality disorders before psychotherapy

BACKGROUND

Cross-sectional network analysis examines the relationships between symptoms to explain how they constitute disorders. Up to now, research focuses mostly on depression, posttraumatic stress disorder, and rarely assesses larger networks of various symptoms measured with instruments independent of classifications. Studies on large groups of psychotherapy patients are also rare.

METHODS

Analyzing triangulated maximally filtered graph (TMFG) networks of 62 psychological symptoms reported by 4616 consecutive nonpsychotic adults in 1980-2015.

RESULTS

Case-dropping and nonparametric bootstrap proved the accuracy, stability and reliability of networks in patients' sex-, age-, and time of visit divided subgroups. Feeling that others are prejudiced against the patient was the most central symptom, followed by catastrophic fears, feeling inferior and underestimated. Sadness, panic, and sex-related complaints were less central than we expected. All analysed symptoms were connected, and we found only small sex-related differences between subsamples' networks. No differences were observed for time of visit and age of patients.

LIMITATION

Analyses were cross-sectional and retrospective, not allowing examination of directionality or causality. Further, data are at the between-person level; thus, it is unknown whether the network remains constant for any person over time. One self-report checklist and building binary network method may bias results. Our results indicate how symptoms co-occured before psychotherapy, not longitudinally. Our sample included public university hospital patients, all White-Europeans, predominantly females and university students.

CONCLUSIONS

Hostile projection, catastrophic fears, feeling inferior and underestimated were the most important psychological phenomena reported before psychotherapy. Exploring these symptoms would possibly lead to enhancement of treatments.

Risk factors of disordered eating in adolescent girls from a community sample: a multidimensional approach

BACKGROUND

Eating disorders (ED) are severe, chronic, and complex in nature mental illnesses that are difficult to treat. One of the ways to stave off EDs is by screening among adolescents to preempt the development of clinical forms of ED in risk groups.

AIM

1) to investigate the prevalence of ED risk among adolescent girls and compare subgroups at high and low risk of ED; 2) to investigate using a multidimensional approach those variables that can interact with temperament and character traits to predict ED symptomatology.

METHODS

The cross-sectional observational self-report study of a community sample of adolescent girls 1217 years old (n=298; M=14.771.13) was carried out in the city of Ryazan, Russia. The Russian versions of Eating Attitudes Test and Cloningers Temperament and Character Inventory-Revised were used. In addition, an original questionnaire (Risk Factors of Eating Disorders) was developed. Regression models (to test for significant moderation) and path analysis (to test for significant mediations) were used.

RESULTS

Girls at risk of developing EDs are characterized by a heightened level of concern about weight and dissatisfaction with their body, tend to suffer from low self-directedness, higher novelty seeking and tendency to higher harm avoidance, display high alexithymia, experience self-distrust, negative emotionality and are dissatisfied with family relationships. They also suffer from low self-esteem and tend to be perfectionism and engage in risk behavior. Significant moderating effects were uncovered between the following ED risk factors: (1) self-distrust/risk behavior and BMI; (2) alexithymia/negative emotionality/self-esteem and cooperativeness; and (3) negative emotionality/risk behavior and self-transcendence. Family relationship dissatisfaction mediates the association between self-directedness/cooperativeness/self-transcendence and disordered eating.

CONCLUSION

There are various mutual influences between the numerous ED risk and prevention factors, which all together determine the paths between the predictors and final outcome.

Network-based validation of the psychometric questionnaire EDI-3 for the assessment of eating disorders

Assessing the validity of a psychometric test is fundamental to ensure a reliable interpretation of its outcomes. Few attempts have been made recently to complement classical approaches (e.g., factor models) with a novel technique based on network analysis. The objective of the current study is to carry out a network-based validation of the Eating Disorder Inventory 3 (EDI-3), a questionnaire designed for the assessment of eating disorders. Exploiting a reliable, open source sample of 1206 patients diagnosed with an eating disorder, we set up a robust validation process encompassing detection and handling of redundant EDI-3 items, estimation of the cross-sample psychometric network, resampling bootstrap procedure and computation of the median network of the replica samples. We then employed a community detection algorithm to identify the topological clusters, evaluated their coherence with the EDI-3 subscales and replicated the full validation analysis on the subpopulations corresponding to patients diagnosed with either anorexia nervosa or bulimia nervosa. Results of the network-based analysis, and particularly the topological community structures, provided support for almost all the composite scores of the EDI-3 and for 2 single subscales: Bulimia and Maturity Fear. A moderate instability of some dimensions led to the identification of a few multidimensional items that should be better located in the intersection of multiple psychological scales. We also found that, besides symptoms typically attributed to eating disorders, such as drive for thinness, also non-specific symptoms like low self-esteem and interoceptive deficits play a central role in both the cross-sample and the diagnosis-specific networks. Our work adds insights into the complex and multidimensional structure of EDI-3 by providing support to its network-based validity on both mixed and diagnosis-specific samples. Moreover, we replicated previous results that reinforce the transdiagnostic theory of eating disorders.