If you’ve ever been assessed for an eating disorder in a clinical setting, there is a good chance you’ve completed the Eating Disorder Examination Questionnaire (EDE-Q). The EDE–Q is a self-report questionnaire widely used in ED assessment and research. Clinicians and researchers calculate several different scores from patient or participant responses to the questionnaire:
- A score on the global scale, which provides a measure of the severity of ED psychopathology
- 4 sub-scales: eating restraint, eating concern, weight concern and shape concern
There are a number of cut-off scores that can be used to distinguish between clinically significant and non-significant cases. In this post, I will look at a few papers critiquing the use of the EDE-Q in clinical and research settings.
The EDE-Q was originally developed as an assessment tool for bulimia nervosa and binge eating and contains few, if any, questions that specifically assess anorexia nervosa symptomology. However, it is used for assessment of all EDs and responses to the questionnaire inform much research into AN.
But is the EDE-Q actually able to detect an ED? In 2013, Thomas, Roberto and Berg looked at the EDE (the longer, interview-based version of the EDE-Q) in some depth, and concluded that it has some serious limitations and should be retired. They raised a number of concerns, including that
- It doesn’t fit well with the new DSM-5 diagnoses
- It is unsuitable for use with men
What does this look like in practice? Let’s look at questions 11 and 12 that, on the surface, appear to address AN symptomology.
On how many of the past 28 days …
11. Have you felt fat?
12. Have you had a strong desire to lose weight?
It’s obvious that responses would carry a completely different meaning if the respondent were overweight or underweight. In an emaciated respondent, even a low score might indicate a serious disturbance in perception, whereas in an overweight person a high score could be clinically unremarkable. But the EDE-Q is unable to make this distinction: it simply assesses the second (person in a larger body) as having a higher level of ED psychopathology.
In fact, it is not uncommon for people with AN to obtain sub-clinical scores. To quote from Thomas, Roberto & Berg:
Indeed, since scoring within one standard deviation of EDE community norms is now being used as a criterion for remission in transdiagnostic treatment studies (e.g. Fairburn et al., 2009), it seems somewhat problematic that, in two recent trials of CBT for AN, 42% of adolescents (Dalle Grave, Calugi, Doll, & Fairburn,2013) and 33% of adults (Fairburn et al., 2013) already scored in the normative range before treatment even took place.
In another study, Ro, Reas and Rosenvinge (2012) looked at how EDE-Q scores vary based on age and BMI. The authors identified that a higher BMI is in itself a predictor of high EDE-Q scores, even in the absence of an ED. Their conclusion was that it may actually be normative (i.e., common, frequent, and even socially sanctioned) for those who are living in larger bodies to have higher scores on the EDE-Q; accordingly, those using the EDE-Q to diagnose EDs should take BMI into account.
A surprising thing about the EDE-Q is how much it focuses on “normative” body image and dieting behaviours while asking nothing about behaviours that might point to a clinical ED–for example, feeling compelled to restrict, or eating for emotional reasons. Therefore, it is perhaps not surprising that it often misses cases of EDs: it doesn’t appear to search for them.
It may be, then, that the EDE-Q is not actually identifying EDs as much as it is identifying symptoms associated with EDs in those who may or may not have clinical EDs.
One of the most serious issues with the EDE-Q is that it does not present the option of any motivation for ED behaviour other than shape and weight concerns. Immediately, this rules out all other commonly cited reasons for restricting, binging and purging–for instance (and among other things) trauma, anxiety, auto-protection and control.
Even sections of the EDE-Q that are unrelated to weight concern or shape concern are built on the assumption that weight or shape concern drive the ED. For example, questions 1-5 (which score in the sub-scale eating restraint) all include the qualifying phrase “to influence your shape or weight.” In other words, dietary restriction such as fasting, eliminating food groups and keeping within low calorie limits is only accepted as indicative of an ED if it is carried out because of concerns about shape and weight.
The reason for this focus is that the EDE-Q follows the cognitive behavioural (CB) model of EDs. This is a trans-diagnostic model that theorises that all ED symptomology (even seemingly opposite behaviours such as binge eating and restricting) arise from the same core psychopathology: an overvaluation of shape and weight. Further, the model implies that binge eating (the original condition for which the questionnaire was compiled) is caused by failed attempts at restricting one’s diet. The questions in the EDE-Q all stem from the assumption, then, that overvaluation of shape and weight drive symptoms.
However, the CB model for EDs is only one of many. Alternative models recognise that a preoccupation with reducing weight can be a symptom–rather than the cause of–an ED. This is one of the most puzzling things about the CB model: it does not clearly present the difference between cause and effect. One might argue that following the CB model, ED symptoms are (mis)understood as indications of pre-existing core beliefs.
Others have raised this issue–notably, Gowers and Shore (2001), writing:
Although weight and shape concern seems commonly to underlie the development of eating disorders, an alternative pathway appears to exist through impulsivity and fear of loss of control. Prevention strategies may usefully focus on the attitudes and concerns that lead to dieting behaviour.
I’m interested to know if anyone is aware of any other instance in healthcare where a questionnaire has decreed in advance that there is only one acceptable motivation for a specific mental health issue.
USE IN ASSESSMENT AND RESEARCH
At present, the EDE-Q is widely used as an assessment tool for individuals seeking help for an ED. It can be quite a distressing experience to fill in this questionnaire, which seems to be calculating how shallow, vain and self-obsessed you are. But, more importantly, the questionnaire requires respondents to tacitly agree with the weight/shape hypothesis even if it doesn’t apply to them. For example, question 1 (the only question about general food restriction):
On how many of the past 28 days ……
1. Have you been deliberately trying to limit the amount of food you eat to influence your shape or weight (whether or not you have succeeded)?
If a respondent with AN reports that they “never” restrict–on the grounds that they are not interested in shape and weight–they are all too aware that this will result in a low score and may affect their treatment prospects. Therefore, many people probably ignore the clause ascribing motivation and simply report the frequency of the behaviour in question. (I have done this myself). Yet, these responses are used in research studies to conclude that people diagnosed with EDs are motivated by high levels of weight/shape concern.
It’s true that there has been a lot of research examining EDE-Q psychometrics (i.e., how well the scale measures what it is supposed to, how consistent the scores are across time, etc.) and discriminative ability (i.e., accuracy in diagnosing). To do this, researchers often compare confirmed/diagnosed cases of EDs with scores on sub-scales like weight concern and shape concern to see whether having a diagnosis of an ED means a person scores higher on these subscales (and, thus, whether scoring high on the subscales would indicate an ED of clinical significance).
However, the whole system is a conceptually closed circle–that is, a question of finding what you’re looking for–and it is concerning that researchers then go on to use these concepts in their research as if they are actually meaningful. For example, Byrne et al. (2015) argue that there is no need to change the course of treatment based on adolescents’ varied weight and shape concerns at intake. Presently, research doesn’t–and can’t–take into account that these concerns were never actually mentioned by the respondents. Those who compiled the questionnaires created the questions (all about shape and weight) and the “answers” (pre-assigning a motivation of shape and weight); respondents then have to indicate frequency. There is no room, here, for respondents to articulate what they think caused their issues.
Over-relying on this one tool in ED research presents a number of dangers. Perhaps most importantly, researchers often use EDE-Q scores as a proxy for an ED diagnosis: this is particularly the case where studies recruit for participants over the Internet. So, data obtained from people who don’t actually have an ED (but who do have high weight/shape and body image concerns) may be used to draw conclusions about ED psychopathology or to develop treatment protocols.
An example of this was a study recently featured on this blog. The researchers selected participants using EDE-Q cut-off scores, even though “most participants had never been diagnosed with an eating disorder”. They then discussed the findings as if they were certainly transferable to a clinical population. Research such as this appears to suggest that there is no meaningful difference between disordered eating and EDs.
And, conversely, when online research studies use EDE-Q cut-off scores to select participants, people with clinically diagnosed EDs can be judged ineligible to take part. If a would-be participant scores below the cut-off, they are deemed not to have an ED and are barred from contributing to the study.
I second Thomas, Roberto and Berg (2013)’s conclusion that the EDE (and EDE-Q) should be retired and that, ideally, questionnaires should be developed that can assess for actual ED behaviours and cognitions–not merely weight and shape concerns. Furthermore, I contend that the EDE-Q actively hinders efforts to understand the true etiology of EDs and improve treatment options. By decreeing in advance that EDs can only arise out of weight/shape concerns and attempts at restraint it keeps much of the discourse around EDs firmly in the domain of dieting and body image–a place where many would argue it does not belongs.
Bryne, C. E., Kass, A. E., Accurso, E. C., Fischer, S., O’Brien, S., Goodyear, A., Lock, J. & LeGrange, D. (2015). Overvaluation of shape and weight in adolescents with anorexia nervosa: Does shape concern or weight concern matter more for treatment outcome? Journal of Eating Disorders, 3, PMID: 26677412.
Gowers, S. G., & Shore, A. (2001). Development of weight and shape concerns in the aetiology of eating disorders. The British Journal of Psychiatry: The Journal of Mental Science, 179, 236-242. PMID: 11532801.
Ro, R., Reas, D., & Rosenvinge, J. (2012). The impact of age and BMI on Eating Disorder Examination Questionnaire (EDE-Q) scores in a community sample. Eating Behaviours, 13(2), 158-161. DOI: 10.1016/j.eatbeh.2011.12.001
Thomas, J., Roberto, C., & Berg, K. (2013). The Eating Disorder Examination: a semi-structured interview for the assessment of the specific psychopathology of eating disorders. Advances in Eating Disorders, 2 (2), 190-203 DOI: 10.1080/21662630.2013.840119