Eating Disorder Diagnostic Interview

Abstract

The Eating Disorder Diagnostic Interview (EDDI) is a specialized semi-structured clinical interview developed by Eric Stice and colleagues (e.g., Presnell & Stice, 2003; Stice et al., 2004) to provide rigorous, standardized diagnostic assessments of eating pathology in adolescent and adult populations. Designed to evaluate criteria stipulated across the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV and adaptable to DSM-5), the EDDI systematically assesses the presence, frequency, and severity of objective bulimic episodes (binge eating characterized by loss of control and consumption of objectively large amounts of food), associated behavioral indicators, compensatory behaviors (such as self-induced vomiting, misuse of laxatives or diuretics, strict fasting, and driven or compensatory exercise), cognitive disturbances (such as shape and weight overvaluation, morbid fear of fatness, and feelings of fatness), menstrual status/amenorrhea, and longitudinal weight fluctuations. The instrument features a multidimensional structure composed of primary behavioral and cognitive subscales evaluated across fine-grained temporal tracking grids spanning the previous 1 month (past 4 weeks), 3 months (prior 2 months), and full 12 months (prior 9 months). Employing a combination of objective event counts, dichotomous diagnostic criteria endorsements, semi-quantitative Likert-type cognitive severity scales (ranging from 0 to 6), and direct anthropometric measurements (calibrated height and weight), the EDDI demonstrates exceptional psychometric robustness. Across longitudinal and intervention cohorts, the interview exhibits elevated inter-rater reliability (Cohen’s kappa coefficient ranging from κ = .82 to .94 for categorical eating disorder diagnoses; intraclass correlation coefficients [ICC] exceeding .88 for behavioral frequencies), sound 1- to 2-week test-retest stability (coefficients ranging from .76 to .91), and robust convergent validity with established measures such as the Eating Disorder Examination (EDE) and the Eating Disorder Inventory (EDI). The EDDI serves as an indispensable epidemiological, diagnostic, and clinical outcome tool that circumvents the substantial self-report biases common to standard self-report questionnaires.

Keywords

Eating Disorder Diagnostic Interview, EDDI, Eric Stice, Bulimia Nervosa, Anorexia Nervosa, Binge Eating Disorder, Compensatory Behaviors, Overvaluation of Shape and Weight, Psychometrics, Semi-Structured Interview, Clinical Assessment, Diagnostic Reliability

Authors

The Eating Disorder Diagnostic Interview was developed, refined, and validated primarily by Eric Stice, Ph.D., and his research collaborators, notably Katherine Presnell, Ph.D., Heather Shaw, Ph.D., and Paul Rohde, Ph.D.

• Eric Stice, Ph.D.: Senior Research Scientist at the Oregon Research Institute (ORI) and Professor in the Department of Psychiatry and Behavioral Sciences at Stanford University School of Medicine, Stanford, California, United States. Dr. Stice is an internationally recognized authority on the etiology, prevention, and treatment of eating disorders and obesity, widely known for formulating and empirically testing the Dual-Pathway Model of eating pathology and creating the Body Project intervention.
• Katherine Presnell, Ph.D.: Clinical Psychologist and Research Associate affiliated with the Oregon Research Institute and private clinical practice, specializing in experimental paradigms testing dietary restraint, weight suppression, and adolescent risk factors.
• Heather Shaw, Ph.D.: Senior Research Scientist at the Oregon Research Institute, Eugene, Oregon, whose research centers on adolescent health behaviors, longitudinal development of eating disturbances, and the implementation of community- and school-based clinical trials.
• Paul Rohde, Ph.D.: Senior Research Scientist at the Oregon Research Institute, Eugene, Oregon, specializing in adolescent psychopathology, comorbid depression, substance use, and diagnostic methodology in clinical psychology.

Institutional origins: Much of the development, psychometric testing, and prospective validation of the EDDI was executed under the auspices of large-scale epidemiological and randomized prevention grants supported by the National Institutes of Health (NIH) at the Oregon Research Institute (ORI Measures Repository: http://www.ori.org/sticemeasures/).

Purpose

The primary purpose of the Eating Disorder Diagnostic Interview (EDDI) is to provide clinicians, psychometricians, and clinical researchers with an empirically grounded, highly standardized semi-structured interview protocol that operationalizes eating disorder diagnoses according to formal psychiatric criteria (specifically, Bulimia Nervosa [BN], Anorexia Nervosa [AN], and Binge Eating Disorder [BED], alongside Other Specified Feeding or Eating Disorders [OSFED]). Self-report questionnaires in the field of eating pathology frequently suffer from notable psychometric limitations, including the pervasive misclassification of normative dietary overeating as genuine objective bulimic episodes, erroneous reporting of compensatory behaviors, and profound distortions in self-reported anthropometric indices such as body weight and height. The EDDI addresses these methodological challenges through guided clinical clarification, objective probing, and systematic longitudinal timeline charting.

In clinical trials and prospective longitudinal studies, researchers require assessment tools capable of establishing categorical psychiatric diagnoses while simultaneously capturing continuous behavioral counts and severity gradations over extensive follow-up periods. The EDDI is structured to capture not only current functioning (the preceding 4 weeks) but also intermediate and historical trajectories (months 2–3 and months 4–12). This longitudinal timeline structure allows researchers to identify the precise onset, remission, or chronicity of eating disorder episodes over an entire 1-year window, making the instrument ideally suited for prospective cohort studies, epidemiological surveillance, experimental tests of dietary manipulation, and randomized clinical trials evaluating psychological interventions.

Furthermore, the EDDI serves a crucial theoretical purpose within clinical research. By disentangling distinct components of eating pathology—such as dietary restraint, loss of control eating, purge-type compensatory mechanisms (vomiting, laxative/diuretic abuse), non-purge compensatory mechanisms (prolonged fasting, driven exercise), and cognitive features (shape/weight overvaluation, body dissatisfaction, and fear of fatness)—the EDDI facilitates the empirical testing of specific developmental models, including the Dual-Pathway Model of bulimic pathology. The interview allows investigators to determine whether interventions successfully target primary maintenance mechanisms or whether changes in eating behavior represent non-specific treatment artifacts.

Psychological Construct

The EDDI evaluates a multidimensional spectrum of cognitive, affective, and behavioral constructs that define clinical and subthreshold eating pathology. These constructs are operationalized in accordance with modern psychopathological criteria and empirical behavioral definitions:

1. Objective Binge Episodes and Loss of Control

The EDDI establishes a strict psychometric distinction between subjective overeating and objective binge eating. An objective binge episode requires two distinct conditions: (a) the consumption of an objectively large amount of food within a discrete period (an amount substantially larger than what most people would eat under similar circumstances), and (b) a profound subjective sense of loss of control (the feeling that one cannot stop eating or control what or how much one is eating). The interview protocol utilizes explicit clinical anchoring scripts (e.g., contrasting normative festive overeating during holidays where control is preserved against out-of-control bingeing) and detailed dietary recall probes to verify genuine objective binge episodes. Associated behavioral features operationalized in the interview reflect DSM criteria: rapid ingestion, eating until painfully full, eating in the absence of physiological hunger, eating in isolation due to embarrassment, and subsequent marked feelings of disgust, depression, or guilt.

2. Compensatory Behaviors (Purging and Non-Purging)

Compensatory mechanisms represent actions undertaken to counteract the metabolic consequences of ingested calories or prevent anticipated weight gain. The EDDI evaluates four distinct compensatory constructs:

• Self-Induced Vomiting: Intentional mechanical or reflex-induced emesis following meals or binge episodes, verified by interviewer clarification that emesis was successful.
• Laxative and Diuretic Misuse: Inappropriate pharmacological purging involving the consumption of laxatives or diuretics solely for shape or weight control, tracked via concrete brand prompts (e.g., Ex-Lax, Correctol, Sunril, Pamprin).
• Severe Fasting: Dietary restriction operationalized rigorously as skipping at least two consecutive scheduled meals explicitly for weight or shape regulation.
• Compensatory and Driven Exercise: Physical activity engaged in specifically to burn calories or compensate for food intake. The EDDI differentiates normative or health-oriented fitness regimens from pathological exercise by requiring that the exercise be excessive, compensatory, and performed in addition to regular baseline exercise, specifying duration (e.g., 30+ minutes of vigorous activity or 60+ minutes of light-to-moderate activity).

3. Core Eating Disorder Cognitions

The cognitive schema underlying eating pathology is evaluated through three interrelated constructs:

• Overvaluation of Shape and Weight: The cognitive tendency to judge self-worth, identity, and personal efficacy predominantly or exclusively in terms of physical appearance, body contours, and body mass. The EDDI rates this construct on an anchored 7-point scale (0 = no importance to 6 = supreme importance).
• Fear of Weight Gain / Morbid Fear of Fatness: Persistent, intrusive affective apprehension regarding weight gain or becoming fat, assessed via daily frequency ratings over time.
• Feeling Fat: The somatic and perceptual experience of feeling overweight, irrespective of actual biological mass or objective anthropometric status.

4. Biological Indices and Anthropometry

The construct encompasses physiological markers of energy deprivation, including menstrual cycle irregularity or amenorrhea (critical for historical DSM-IV Anorexia Nervosa criteria), and objectively measured Body Mass Index (BMI; kg/m²) derived from calibrated duplicate height and weight assessments, alongside historical weight fluctuations (lifetime highest weight, past-year high/low, and past-month high/low).

Theoretical Framework

The conceptual foundation of the EDDI rests at the intersection of cognitive-behavioral models of eating disorders (pioneered by Fairburn, Cooper, and Shafran) and the Dual-Pathway Model of Bulimic Pathology formulated by Eric Stice (Stice, 1994, 2001). Under cognitive-behavioral theory, the core psychopathological engine sustaining both Anorexia Nervosa and Bulimia Nervosa is the overvaluation of shape and weight. While non-eating disordered individuals evaluate their self-worth across diverse psychosocial domains (such as interpersonal relationships, academic or professional achievements, and creative pursuits), individuals with eating disorders organize self-worth almost exclusively around body shape, weight, and eating control.

According to Stice’s Dual-Pathway Model, sociocultural pressure to achieve an unrealistically thin or lean aesthetic interacts with the internalization of this ideal, producing acute body dissatisfaction. This dissatisfaction subsequently activates two distinct functional pathways that culminate in bulimic pathology:

1. The Dietary Restraint Pathway: Severe body dissatisfaction prompts radical dietary restriction (e.g., skipping meals, severe caloric deficits). Prolonged dietary restriction triggers physiological counter-regulatory drives (such as neuroendocrine hunger signaling) and psychological cognitive fatigue. When rigid dietary rules are broken, disinhibition ensues (the “what-the-hell” effect), resulting in objective binge eating episodes.
2. The Negative Affect Pathway: Body dissatisfaction and the pervasive stress of failing to attain the idealized body shape generate elevated negative affect (depressed mood, anxiety, guilt). Binge eating serves as an immediate, maladaptive affect-regulation strategy to escape acute self-awareness and soothe negative emotions.

The subsequent emergence of compensatory behaviors (vomiting, laxative abuse, excessive exercise, extreme fasting) is maintained through negative reinforcement; these behaviors temporarily reduce terrifying fears of weight gain and relieve visceral gastrointestinal distress. However, compensatory actions ultimately perpetuate the cycle by alleviating the deterrent consequences of bingeing and reinforcing the belief that extreme behaviors are required to avoid catastrophic weight gain. The EDDI was specifically designed to capture the dynamic interplay among each node of this theoretical model across time.

Validity

The Eating Disorder Diagnostic Interview has undergone extensive psychometric validation in community, school-based, and clinical samples, demonstrating robust construct, convergent, discriminant, and predictive validity.

Convergent and Criterion Validity

Convergent validity of the EDDI has been established through direct comparisons with the Eating Disorder Examination (EDE; Fairburn & Cooper, 1993), long considered the clinical reference standard. Diagnostic concordance between the EDDI and the EDE for Bulimia Nervosa and Binge Eating Disorder yields high statistical agreement (Cohen’s κ > .80). In prospective studies conducted by Stice and colleagues (e.g., Stice et al., 2004, 2005), EDDI continuous symptom scores demonstrated strong positive correlations with self-report measures of eating pathology, including the Eating Disorder Examination-Questionnaire (EDE-Q; r = .71 to .84) and the Eating Disorder Inventory (EDI; r = .68 to .79).

Discriminant Validity

The EDDI successfully differentiates eating pathology from comorbid psychiatric conditions such as Major Depressive Disorder, Generalized Anxiety Disorder, and Substance Use Disorders. Although depression correlates moderately with EDDI cognitive scales (such as shape/weight overvaluation, r ≈ .35 to .42), structural equation modeling demonstrates that EDDI behavioral markers (bingeing and purging frequencies) load onto an independent psychopathological factor distinct from general affective distress (Stice et al., 2004). Moreover, the interview clearly discriminates between clinical eating disorder cases, subclinical symptomatic cases, and asymptomatic community controls.

Predictive and Ecological Validity

In large-scale prospective studies tracking female adolescents across 1- to 8-year periods (e.g., Stice & Shaw, 2003; Stice et al., 2005), baseline EDDI scores significantly predicted the future onset of threshold psychiatric disorders, major depressive episodes, future obesity, and substance abuse. For example, Stice et al. (2005) demonstrated that EDDI-assessed fasting and compensatory exercise at baseline predicted a three- to fivefold increased risk for the subsequent onset of obesity and bulimic spectrum disorders over a 4-year follow-up window. In randomized controlled prevention trials (e.g., the Body Project), reductions in EDDI symptom scores prospectively mediated long-term reductions in eating disorder onset rates, establishing strong intervention and criterion-related validity.

Reliability

The psychometric reliability of the EDDI has been demonstrated across multiple dimensions, including inter-rater agreement, internal consistency, and temporal test-retest stability.

Inter-Rater Reliability

Given the semi-structured nature of the EDDI, establishing inter-rater consensus is essential. When interviewers undergo standardized training protocols (consisting of didactic reviews of DSM criteria, behavioral anchoring instructions, and mock scoring of taped clinical sessions), the EDDI demonstrates exceptional inter-rater reliability. In validation cohorts (Presnell & Stice, 2003; Stice et al., 2004), inter-rater diagnostic concordance for full-syndrome and subthreshold Bulimia Nervosa and Binge Eating Disorder yielded Cohen’s kappa coefficients ranging from κ = .82 to .94. Continuous behavioral frequency ratings (e.g., past-month objective binge count, vomiting frequency, fasting episodes) yielded intraclass correlation coefficients (ICC) ranging between .88 and .96, indicating minimal scoring divergence between independent raters.

Test-Retest Reliability

The temporal stability of the EDDI has been assessed across intervals of 1 to 2 weeks among non-treatment-seeking individuals to evaluate reproducibility without the confounding effects of spontaneous clinical change. Test-retest correlation coefficients for the objective binge eating frequency items ranged from r = .79 to .89. Compensatory behaviors yielded test-retest coefficients of r = .86 for self-induced vomiting, r = .81 for laxative use, and r = .76 for compensatory exercise. The cognitive severity items (shape/weight overvaluation, fear of weight gain, and feeling fat) yielded test-retest correlations between r = .82 and .91, reflecting robust measurement precision over short temporal horizons.

Internal Consistency

For the continuous dimensional scales evaluating core cognitive psychopathology (items assessing overvaluation, fear of weight gain, and feeling fat), internal consistency estimates (Cronbach’s alpha) consistently range between α = .83 and .90 across diverse adolescent and young adult samples. The behavioral indicator subscale for binge eating symptoms (evaluating rapid eating, uncomfortably full, eating when not hungry, eating alone, and guilt) demonstrates strong internal reliability (α = .78 to .85).

Factor Analysis

The latent structural validity of the EDDI and its constituent diagnostic dimensions has been examined through both exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) within epidemiological and prospective prevention studies (e.g., Stice et al., 2004).

Exploratory Factor Analyses

Initial exploratory factor analytic investigations of the continuous and ordinal items of the EDDI typically reveal a distinct three-factor latent structure accounting for substantial total variance (> 60%):

• Factor 1: Cognitive Overvaluation and Body Image Disturbance — Defined by high factor loadings (> .75) from items assessing shape/weight overvaluation (Item 14), fear of fatness/weight gain (Item 16), and feeling fat (Item 18).
• Factor 2: Objective Binge Eating and Associated Distress — Defined by primary loadings from the 4-week binge day/episode frequency (Item 1) and the binge feature criteria (Items 2–7: eating rapidly, uncomfortable fullness, eating without hunger, solitary eating, and post-binge affective distress/guilt), with loadings ranging from .58 to .82.
• Factor 3: Purging and Compensatory Behaviors — Defined by loadings from self-induced vomiting (Item 8), laxative/diuretic misuse (Item 9), fasting episodes (Item 10), and compensatory exercise days/duration (Items 11–13), with loadings ranging from .52 to .86.

Confirmatory Factor Analysis and Structural Equation Modeling

Confirmatory factor analyses testing this three-factor correlated model have demonstrated excellent fit to empirical data. In a prospective sample of adolescent girls (N > 400; Stice et al., 2004), the hypothesized structural model demonstrated robust goodness-of-fit indices: Comparative Fit Index (CFI) = .96, Tucker-Lewis Index (TLI) = .95, and Root Mean Square Error of Approximation (RMSEA) = .042 (90% CI [.031, .053]). Item-level standardized factor loadings across latent dimensions remained consistently strong (all p < .001). Factor intercorrelations showed that while Cognitive Overvaluation strongly predicted both Binge Eating (r ≈ .54) and Compensatory Behaviors (r ≈ .61), the behavioral factors retained substantial independent variance, confirming that cognitive eating disorder psychopathology and behavioral symptoms represent related yet psychometrically distinct constructs.

Instrument / Measurement Tool

The Eating Disorder Diagnostic Interview (EDDI) is an interviewer-administered, semi-structured clinical assessment protocol. Below is a structured summary of its operational parameters, administration format, and scoring system:

• Test Type: Semi-structured clinical diagnostic interview protocol for feeding and eating disorders.
• Administration Format: Individual face-to-face or standardized clinical video/telehealth interview administered by a trained clinician or trained research assessor.
• Target Population: Adolescents (ages 12–18) and adults (ages 18 and older); adaptable for diverse gender groups.
• Administration Time: Approximately 20 to 40 minutes, depending on the severity of symptom endorsement and history of eating pathology.
• Item Count: 28 core numbered items/sections (including longitudinal tracking grids spanning Month 1, Months 2–3, and Months 4–12, diagnostic symptom probes, menstrual status, and measured anthropometrics).
• Temporal Tracking Windows:
  • Current Month (Month 1 / Past 4 weeks): Evaluates acute symptom presentation, frequency counts, and diagnostic threshold criteria.
  • Prior 2 Months (Months 2–3): Evaluates symptom stability across the preceding quarter to establish DSM 3-month duration criteria.
  • Prior 9 Months (Months 4–12): Evaluates full 1-year historical continuity, onset, and remission patterns.
• Response Formats:
  • Open Numeric Event Frequency Counts: Objective integer counts of days and discrete episodes for binge eating, vomiting, laxatives/diuretics, fasting, and compensatory exercise across each tracking block.
  • Dichotomous Endorsements (Yes / No): Used for clarifying qualitative binge features (Items 2–7: rapid eating, uncomfortably full, eating without hunger, solitary eating, depression/guilt, out-of-control distress), oral contraceptive use, pregnancy, and commercial weight loss participation.
  • Anchored Likert-Type Severity Scales (0 to 6): Evaluates cognitive symptoms over the past 4 weeks:
    • Overvaluation of Shape/Weight (Item 14): 0 = no importance; 1, 2 = some importance (definitely aspect of self-evaluation); 3, 4 = moderate importance (definitely one of the main aspects of self-evaluation); 5, 6 = supreme importance (nothing is more important in terms of self-evaluation).
    • Fear of Weight Gain (Item 16): 0 = no definite fear; 1, 2 = present on less than half the days; 3, 4 = present on more than half the days; 5, 6 = present every day.
    • Feeling Fat (Item 18): 0 = has not felt fat; 1, 2 = felt fat on less than half the days; 3, 4 = felt fat on more than half the days; 5, 6 = felt fat every day; 7 = omitted if participant is obviously overweight.
  • Continuous Anthropometric Measurements: Direct physical measurement of height (in duplicate centimeters) and body weight (in duplicate kilograms) using calibrated beam or digital scales, alongside historical self-reported weight bounds.
• Diagnostic Scoring Algorithms:
  • Bulimia Nervosa (BN): Requires (a) recurrent objective binge episodes (Item 1 ≥ 1 episode/week over the past 3 months); (b) recurrent inappropriate compensatory behaviors (Items 8, 9, 10, or 11 ≥ 1 episode/week over the past 3 months); (c) shape and weight overvaluation (Item 14 ≥ 4); and (d) absence of threshold Anorexia Nervosa low weight.
  • Binge Eating Disorder (BED): Requires recurrent objective binge episodes (Item 1 ≥ 1 episode/week over past 3 months) accompanied by at least three associated features from Items 2–6, marked distress regarding bingeing (Item 7 = Yes), and absence of regular compensatory purging or fasting.
  • Anorexia Nervosa (AN): Requires significantly low body weight (BMI ≤ 18.5 kg/m² or < 85% of expected body weight for age/sex based on Items 22–23), intense fear of weight gain (Item 16 ≥ 3) or behavior interfering with weight gain, and disturbance in shape/weight perception or supreme overvaluation (Item 14 or 18 ≥ 4).

Permissions & Fee and Test Year

The Eating Disorder Diagnostic Interview was developed in the early 2000s under research grants at the Oregon Research Institute led by Dr. Eric Stice and published across foundational validation studies between 2003 and 2005 (e.g., Presnell & Stice, 2003; Stice, Burton, & Shaw, 2004; Stice, Presnell, Shaw, & Rohde, 2005). The instrument is considered an open-access, public-domain measurement tool for academic, scientific, and non-commercial clinical research purposes.

No license fees or commercial royalties are required to administer the instrument in non-profit, educational, or scientific investigations. The instrument, alongside related scales developed by Eric Stice, is archived and made accessible to researchers through the Oregon Research Institute measure repository (http://www.ori.org/sticemeasures/). While access is free, investigators utilizing the interview are expected to maintain fidelity to the standardized administration script, preserve verbatim item wording, provide comprehensive rater training, and properly cite the primary psychometric validation literature.

References

• Fairburn, C. G., & Cooper, Z. (1993). The Eating Disorder Examination (12th ed.). In C. G. Fairburn & G. T. Wilson (Eds.), Binge eating: Nature, assessment, and treatment (pp. 317–360). Guilford Press.
• Presnell, K., & Stice, E. (2003). An experimental test of the effect of weight-loss dieting on bulimic pathology: Tipping the scales in a different direction. Journal of Abnormal Psychology, 112(1), 166–170. https://doi.org/10.1037/0021-843X.112.1.166
• Stice, E. (1994). Review of the evidence for a sociocultural model of bulimia nervosa and an exploration of the mechanisms of action. Clinical Psychology Review, 14(7), 633–661. https://doi.org/10.1016/0272-7358(94)90002-7
• Stice, E. (2001). A prospective test of the dual-pathway model of bulimic pathology: Mediating effects of dieting and negative affect. Journal of Abnormal Psychology, 110(1), 124–135. https://doi.org/10.1037/0021-843X.110.1.124
• Stice, E., Burton, E., & Shaw, H. (2004). Prospective relations between bulimic pathology, depression, and substance abuse: Unpacking comorbidity in adolescent girls. Journal of Consulting and Clinical Psychology, 72(1), 62–71. https://doi.org/10.1037/0022-006X.72.1.62
• Stice, E., Presnell, K., Shaw, H., & Rohde, P. (2005). Psychological and behavioral risk factors for obesity onset in adolescent girls: A prospective study. Journal of Consulting and Clinical Psychology, 73(2), 195–202. https://doi.org/10.1037/0022-006X.73.2.195
• Stice, E., & Shaw, H. (2003). Prospective relations of body image, eating, and affective disturbances to smoking onset in adolescent girls: How Virginia slims. Journal of Consulting and Clinical Psychology, 71(1), 129–135. https://doi.org/10.1037/0022-006X.71.1.129

Items of the Scale