Commentary on “Next Steps in Use of the Eating Disorder Examination and Related Eating Disorder Assessments: A Call for Consensus” by Reilly et al.

ABSTRACT

The purpose of this Commentary is to expand upon Reilly et al. 2025's critique of the Eating Disorder Examination (EDE) around a dialectic of its major strengths and its limitations. Although notable strengths of the EDE are its diagnostic case identification and detailed assessment of phenomenology, its resource intensity (training and administration) may explain why researchers and clinicians often prefer to use the EDE‐Q—the self‐report version of the EDE. An example of this is in the development of a national eating disorder assessment package for residential care in Australia, where the EDE‐Q (not the EDE) was recommended by a committee of eating disorder experts advising the Federal Government, and subsequently the EDE‐Q was used by researchers in the clinical evaluation of the first Australian residential program. The present and future need for multidimensional assessments beyond symptoms is presented. A proposal for a global consensus on harmonization of constructs in eating disorder assessment, matching a repertoire of relevant instruments suited to diverse times and places, is suggested.

Abbreviations

CREDO

Centre for Research on Eating Disorders at Oxford

DSM

The Diagnostic and Statistical Manual of Mental Illnesses

ICD

International Classification of Diseases

The Eating Disorder Examination (EDE) is arguably the best instrument for the assessment of the complex phenomenology of eating disorders but also the least accessible and with the lowest clinical utility because of its time and expense of training and administration. This dialectic for the EDE is reflected in my personal experience. I was first introduced to an early version of the EDE interview when I was a very naive trainee in psychiatry in 1985 in New Zealand. My supervisor was sent what we judged an excellent and comprehensive interview schedule by post, the latter a reflection of Fairburn's early and sustained commitment to the dissemination and translation of his work. However, as Reilly et al. (2025) comment, the EDE is “resource intensive” and after our initial read we reluctantly decided it was not practical to use in a public hospital outpatient clinic where the consultation time was insufficient to administer an EDE alongside the requisite comprehensive medical and psychiatric assessment for clinical care. About a decade later, however, I found myself in Oxford being trained by Fairburn in the EDE for its use in large clinical trials and general population epidemiological surveys of prevalence and risk factors for disorders of recurrent binge eating. It was during this period that I learned the true value of this well‐constructed semi‐structured investigator‐based interview for exploring the phenomenology of people with eating disorders that in addition to diagnostic criteria provides a reliable and valid assessment of clinical symptoms that are important targets for treatment.

The first versions of the EDE focused on the phenomenology of bulimia nervosa, and later versions incorporated specific features related to restrictive eating disorders, in particular anorexia nervosa, and the diagnostic criteria for binge eating disorder. In this regard, the exploration of control as a driver of both disordered eating behaviors and cognitions was added for anorexia nervosa. Rather oddly, and I'm not sure why, the DSM diagnostic criteria specifier's B&C for binge eating disorder were added, but not in investigator‐based mode, but rather in respondent‐based mode. What is the difference, you may ask? An investigator‐based interview allows for probing and clarification of responses, and the investigator makes the ratings, not the respondent; whereas a respondent‐based interview is similar to a self‐report question where the respondent determines the score on the item. Thus, the investigator can ensure that responses are scored as closely as possible to the constructs being evaluated, as they are trained in an understanding of these that goes beyond that expected of a respondent. This is a key feature of the EDE and emphasized in Fairburn's Guidelines for Interviewers on the Centre for Research on Eating Disorders at Oxford (CREDO) site (reference available on request). Other extant interviews such as the Structured Clinical Interview for DSM Diagnoses (SCID) resemble the DSM criteria turned into direct questions for which the answers are largely respondent dependent albeit it there is a small capacity for interviewer clarification. The SCID and similar interviews also have a “skip question” format such that a decision to continue is made on the basis of a lead question suggesting the person will likely (or not) meet criteria. This approach relies on a high level of understanding of the phenomena behind the question and has been argued to result in low levels of case identification in a large national survey in the US (Brodrick and McAdams 2019). The EDE also has the capacity for the interviewer to return to earlier questions and amend scoring as the understanding of the person's problems develops during the assessment.

The EDE in my view stands out with regards to its careful clinical assessment of phenomenology and ability to flexibly apply criteria from DSM revisions. Following extensive training and experience with the EDE during my doctoral years at Oxford, I found myself extremely well equipped to use the EDE approach in clinical practice and research, and I use the EDE phrases and prompts in day‐to‐day consultations. In general population epidemiological studies from 1995, I also have used questions adapted from the EDE. Thus, both objective and subjective forms and frequencies of weekly as well as twice weekly binge eating were asked by interviewers in this research program. This ensured that when the DSM and ICD diagnostic criteria for bulimia nervosa changed (e.g., from requiring twice weekly binge eating in both schemes, and the current ICD criteria allowing subjective binge eating) estimates over time of prevalence could yet be done using contemporary definitions.

However, as highlighted by Reilly et al. (2025) in reaching a high standard, the EDE has possibly placed itself out of reach for routine clinical assessments and research where there are time and personnel limitations, as well as considerations of participant burden, which preclude very lengthy assessments. Reilly et al. (2025) detail other problems with the EDE, particularly the limitations of eating disorder symptom assessment of people without behaviors driven by weight/shape overvaluation, those of diverse gender, and those from particular populations such as elite athletes and people with high weight. Some of this is fair, although the EDE has been very widely used, including in studies of people with high weight (e.g., Palavras et al. 2018). It also does assess body image concerns beyond the pursuit of thinness, as the overvaluation of weight and shape questions do not specify what these are for, and thus a person with high levels of muscle dysphoria may well report high levels of overvaluation of shape. It is possible that the predominance of restrictive eating questions and thinness‐oriented body image questions in the EDE may bias respondents to only think about thinness‐related concerns; however, as it is an investigator‐based instrument, the interviewer can prompt respondents to other forms of shape concern.

Not surprisingly, pragmatism has won over in large national or similar surveys which, if they have included eating disorder assessment, they have not used the EDE. The field has generally accepted this and the consequent paucity of general population data on the most common eating disorders such as Other Specified Feeding and Eating Disorders. As noted above, most alternate interviews lack the EDE's very thorough assessment of eating disorder phenomena. For example, the EDE assesses other forms of overeating beyond the DSM binge, such as objective overeating, subjective binge episodes, and frequency of picking or nibbling. The EDE also explores in depth the phenomenology and experience of body shape and weight concerns and eating concerns, and distinguishes between restraint and restriction of eating. To do this, however, also requires extensive and often expensive training and supervision, along with ongoing team consensus agreements on scoring, which is often overlooked or unachievable. Reilly et al. (2025) correctly state that there needs to be much greater recognition of this and reporting on the interview protocols followed by research groups as to how they applied and operationalized terms such as “some” versus “definite” discomfort.

An unforeseen consequence of the challenges of using the EDE is that its questionnaire derivative the EDE‐Questionnaire (EDE‐Q) by default has become more widely used in research and clinical practise. (A brief Pub Med search by the author on 13th April this year resulted in 343 results for “EDE‐Q” and only 16 for “EDE interview” in the past 5 years). This is despite its greater limitations. For example, the EDE‐Q does not assess diagnostic criteria for binge eating disorder and, while generally concordant with the EDE, its assessment of binge eating differentiates poorly between subjective and objective binge episodes and may overestimate the frequency of binge eating episodes. While the subscale scores can be calculated, these also are problematic, as noted by Reilly et al. (2025) as the original factor analysis has not been replicated. Reilly et al. (2025) suggest that shorter versions may be preferable. Indeed, there are several brief forms of the EDE‐Q that are sensitive to change and can provide repeated assessments of symptom severity and some that also assess eating disorder behaviors as well as cognitions. However, none of these can be used as a diagnostic instrument.

What then is the future of the EDE? Will it be relegated like the study of Latin and Greek to a higher order level of scholarship that is perceived to be impractical for the majority. Notably there has been no revision since the 17th April 2014 edition, which is the most complex to train in and to administer. I am not aware of any new editions in progress but if there were I would anticipate a thorough review of item redundancy and utility would be undertaken. The extra questions on control as a specifier of reasons for behaviors and weight/shape overvaluation for example are challenging to administer and do not aid in diagnostic assessment. There also needs to be greater access to training, which is expensive and not easy to find. (As an aside I did a Google search for “Eating Disorder Examination training” while preparing this paper and could not find any web links leading directly to a training.) I agree with the sentiment that the EDE should not be abandoned but it does need continual renewal and revision to remain useful to the field. I would add that the field also needs to move beyond the concept of one instrument fits all and I support the call for global “field‐wide discussions” (Reilly et al. 2025, 17).

In an era of precision care and broader understandings of recovery beyond symptoms, international efforts should reach beyond the EDE interview and EDE‐Q and consider a repertoire of validated instruments to be used in the assessment of people with eating disorders. This is cognizant of the capacity for meta‐analytic data pooling where instruments reflect the same or closely similar features. It would also acknowledge that in consideration of biological, psychological, and social determinants of illness and outcomes, no one instrument will ever be sufficient. In addition to measuring eating disorder symptoms and illness‐specific impairment, there are instruments that add value in measuring complex phenomenology and personality structure such as the Eating Disorder Inventory recommended in manualised psychodynamic therapy assessment (Friederich et al. 2019, 13). In this regard, it is interesting to consider an attempt to establish national consensus on outcome assessment instruments, namely those recommended by national experts in eating disorders on the Federal Technical Advisory Group for the first residential program in Australia (available from the author). This set, like that of the first clinical data registry in Australia (Mitchison, n.d.) includes the EDE‐Q, measures of health‐related quality of life (HRQoL) and mood (anxiety and depression) symptoms. But the sets are neither the same nor include mandated mental health assessment instruments used more broadly across private and public care in Australia, such as a generic measure of HRQoL. Salutary to note, they do not include the EDE interview. This move to reliance on self‐report instruments may be appropriate given the context (i.e., outcomes tracking in clinical practice) but there are other contexts where a diagnostic instrument is required (e.g., estimates of disorder prevalence).

In conclusion, despite its limitations, the EDE interview has made a major and arguably unparalleled contribution to eating disorder research and clinical practice. However, its future is unclear, and its questionnaire derivatives may eventually supersede it. Further, the EDE is not sufficient for a comprehensive biopsychosocial formulation that requires an assessment of risk factors and developmental history. Future endeavors should aim for consensus on the constructs of a multidimensional eating disorder evaluation and appropriate instruments to guide researchers and clinicians to enable the harmonization of research endeavors and the personalization of care.

Author Contributions

Phillipa J. Hay: conceptualization, writing – original draft, writing – review and editing, validation, resources.

Conflicts of Interest

Phillipa J. Hay has received sessional fees from the Therapeutic Guidelines publication and the Health Education and Training Institute (HETI, NSW), and royalties/honoraria from Hogrefe and Huber, McGraw Hill Education, Blackwell Scientific Publications, BioMed Central, and PLOS Medicine. She is Chair of the National Eating Disorders Collaboration Steering Committee in Australia. She has been a consultant to Takeda Pharmaceuticals and is a consultant to Tryptamine Therapeutics. All views in this paper are her own.

Hay, P. J. 2025. “Commentary on “Next Steps in Use of the Eating Disorder Examination and Related Eating Disorder Assessments: A Call for Consensus” by Reilly et al..” International Journal of Eating Disorders 58, no. 9: 1683–1685. 10.1002/eat.24471.

Data Availability Statement

Data sharing is not applicable to this article as no new data were created or analyzed in this study.