EDE-Q Structure Differs in Borderline Personality Disorder

Researchers have identified that the standard four-factor structure of the Eating Disorder Examination Questionnaire (EDE-Q) does not fully generalize to individuals with borderline personality disorder (BPD). In an exploratory graph analysis published in Borderline Personality Disorder and Emotion Dysregulation, a team led by Nora M. Laskowski and colleagues found that a three-dimensional model better captures the attitudinal eating disorder psychopathology in this specific clinical population. The study highlights the importance of validating assessment instruments within specific clinical groups, suggesting that standard subscale scores may require cautious interpretation for patients with BPD.

The investigation addressed a gap in the literature regarding the dimensional structure of commonly used assessment instruments in BPD populations. Although BPD frequently co-occurs with eating disorders, the stability of the EDE-Q’s original four-factor structure has not been sufficiently examined in this group. Previous research has questioned the validity of this structure across diverse populations, including men and gender minority individuals, and recent systematic reviews have indicated that empirical support for the original model is limited. The present study aimed to explore how eating-related symptoms are organized within individuals with BPD using a network-based approach.

The study utilized a clinical sample of 73 adult patients with a confirmed diagnosis of BPD. These participants were recruited between May 2021 and December 2024 as part of a larger study and were undergoing Dialectical Behaviour Therapy (DBT). The sample consisted predominantly of women, with 84.9% identifying as female, 13.7% as male, and 1.4% as gender diverse. The mean age of the participants was 27.57 years, and the mean body mass index was 26.96 kg/m². The analytic sample did not differ significantly from the full recruitment sample in terms of age, BMI, gender distribution, or educational attainment.

To examine the structure of the EDE-Q, the researchers applied Exploratory Graph Analysis (EGA), a network-based method for dimensionality assessment. Unlike traditional latent variable approaches, EGA identifies dimensions as communities of densely connected items within a Gaussian graphical model. This approach is particularly suited for BPD populations, which are characterized by heterogeneous and dynamically interacting symptom profiles. The analysis was restricted to the 22 attitudinal items of the EDE-Q, which assess cognitive and emotional aspects of eating disorder psychopathology. Behavioral items measuring the frequency of binge eating and purging were excluded because they use a different response format.

The baseline EGA on all 22 attitudinal items yielded a four-dimensional structure. However, a nonparametric bootstrap procedure with 1,000 resamples revealed substantial variability in the dimensional solution, with a median dimensionality of four but a 95% confidence interval ranging from two to six dimensions. Item stability estimates indicated that nine items did not meet the predefined stability threshold of 0.70. These items, which included those related to guilt, secrecy, and loss of control, were iteratively removed to identify a stable subset of items contributing to the dimensional structure.

After the iterative item reduction, a three-dimensional solution emerged comprising 13 items. This final structure demonstrated high robustness across multiple sensitivity analyses, including median imputation, variation of the regularization parameter, and comparison of community detection algorithms. The three identified dimensions were labeled Restraint, Cognitive–Interpersonal Eating Concerns, and Body Dissatisfaction. The Restraint dimension, consisting of items 1 through 5, reflects restrictive eating and dietary control. The Cognitive–Interpersonal Eating Concerns dimension, comprising items 7, 8, and 21, is characterized by persistent thoughts about eating and concerns about eating in the presence of others.

The third dimension, Body Dissatisfaction, included items 11 and 25 through 28, reflecting negative body-related affect and the importance of body-related criteria for self-worth. This dimension primarily captured affective and identity-related aspects, whereas items related to avoidance behaviors and appearance ideals were not retained in the final stable model. The convergence of items from the original Shape and Weight Concern subscales into a single Body Dissatisfaction dimension suggests that in individuals with BPD, the distinction between shape and weight concerns may be less clearly differentiated than in classical eating disorder samples.

The findings suggest that body image disturbances in BPD may extend beyond classical weight- or shape-focused eating disorder psychopathology. Previous research indicates that individuals with BPD often report markedly negative attitudes toward their bodies and heightened discomfort related to body image, which are only partially explained by co-occurring eating disorders. The Body Dissatisfaction dimension identified in this study may reflect broader difficulties in self-concept and self-evaluation rather than isolated concerns about body shape or weight. This interpretation is consistent with theoretical models emphasizing identity disturbance and unstable self-image as core features of BPD.

The Restraint dimension was fully replicated and highly stable, mirroring one of the most robust findings across eating disorder research. In contrast, the original Eating Concern subscale was reduced to a narrower cognitive–interpersonal core. Items reflecting preoccupation with eating and social-evaluative concerns were retained, while those capturing guilt and loss of control were unstable and excluded. This fragmentation suggests that the Eating Concern dimension is particularly sensitive to sample characteristics and may not represent a stable or unified construct in the context of BPD.

The study has several limitations that must be considered when interpreting the results. The sample size was relatively small, which limits the generalizability of the findings. Additionally, the sample consisted of individuals with BPD undergoing DBT, predominantly from inpatient settings and largely female. Participants completed the questionnaires at varying stages of treatment, meaning the observed dimensional structure may partly reflect variation in treatment stage. Information on formal eating disorder diagnoses and psychiatric comorbidities beyond BPD was not extracted for this secondary analysis, so it remains unclear to what extent the identified structure may have been influenced by co-occurring psychopathology.

Despite these limitations, the convergence of findings across multiple analytic approaches and sensitivity analyses strengthens confidence in the robustness of the identified dimensional structure. The cross-sectional design precludes conclusions regarding temporal stability or predictive validity, and the exclusion of behavioral items limits conclusions regarding the full spectrum of eating disorder pathology. However, the consistent emergence of a Body Dissatisfaction dimension highlights its potential as a transdiagnostic target for future research and intervention in individuals with BPD.

The results underscore the necessity of validating assessment instruments within specific clinical populations. The routine use of standard EDE-Q subscale scores in BPD samples may be problematic, as these scores may not reflect valid underlying constructs in this group. Network-based approaches such as EGA allow for a more flexible and empirically grounded identification of dimensional structures, providing insight into the organization of psychopathology. These findings support a cautious interpretation of standard EDE-Q subscale scores in individuals with BPD and highlight the complex interplay between personality pathology and eating disorder symptoms.

Subject of Research: Dimensional structure of the EDE-Q in borderline personality disorder

Article Title: Which EDE-Q items are structurally relevant in borderline personality disorder? An exploratory graph analysis

Article References: Laskowski, N. M., Hambruch, H. R., Häkkinen, I., Pollmann, I., Deutschmann, C. M., Brandt, G., Ballero Reque, C., & Paslakis, G. (2026). Which EDE-Q items are structurally relevant in borderline personality disorder? An exploratory graph analysis. Borderline Personality Disorder and Emotion Dysregulation, 13(1), Article 17. https://doi.org/10.1186/s40479-026-00355-1

Image Credits: AI Generated

DOI: 10.1186/s40479-026-00355-1

Keywords: borderline personality disorder, EDE-Q, eating disorders, exploratory graph analysis, psychometrics, clinical psychology, body dissatisfaction, network analysis, items, structurally, relevant, borderline