More on the Problematic EDNOS Category (and Diagnostic Crossover)

I’ve been thinking a lot lately about the eating disorder not otherwise specified (EDNOS) category. ED-NOS is a diagnostic category for all individuals with subthreshold anorexia or bulimia nervosa or those with a mix of symptoms that don’t fit neatly into AN or BN. ED-NOS is essentially everything else. A mixed bag, if you will. It doesn’t tell the clinician nor the researcher anything useful, outside of what it isn’t. So, is there any use for it? If it doesn’t tell the clinician about patient symptoms or guide choice of treatment, why even bother? Does it help researchers understand EDs or do they just want to avoid this messy and heterogenous group (that by the way makes up most of those with eating disorders)? In this entry (and many more to come), I want to further explore these questions.

There’s been a push by researchers to minimize the amount of people who fall into this category. This has namely been done by loosening the criteria for anorexia and bulimia diagnoses for the next edition of the Diagnostic and Statistical Manual (scheduled to be out in 2013). (As a side note: be weary of the “anorexia/bulimia on the rise” when all that’s happened is the diagnostic criteria became less stringent when it comes to symptom frequency and weight criteria.) Some of the proposed changed include getting rid of the amenorrhea requirement for AN and reducing the weight loss required for diagnosis, for BN the plan is to reduce the frequency of bingeing/purging episodes necessary for the diagnosis (from 2x/week to 1x/week).

In general, I think this is a good move because it gets away from (what seem to me) arbitrary numbers: why 2x/week? Why not 2x/month or 2x/day? What about 85% of normal? “Normal” is not even defined! Some use BMI of 17.5, but what if that’s you healthy BMI? Or what if your normal is BMI 20 or BMI 30? Should the same arbitrary BMI 17.5 cut-off apply?

At the moment, the diagnostic criteria for anorexia and bulimia nervosa, at least it terms of symptom frequency and weight criteria, are pretty arbitrary. It is no wonder that most people (whether they seek treatment and get a formal diagnosis or not) fall into ED-NOS.

However, there’s something more important that I think is missed in the quest to minimize the EDNOS category: diagnostic crossover. I talked about this in my previous posts (Eddy et al., 2008 and Eddy et al., 2010). The Eddy et al. studies showed that a significant portion of people who are initially diagnosed with AN or BN, fluctuate between diagnoses (both within subtypes and between disorders), at times quite frequency, over the course of their eating disorders.

The current proposals for the DSM-5 change the proportion of people fitting the AN or BN diagnostic criteria at any given point, but, in my opinion, don’t deal with the fundamental problem that, in my view, plagues the literature, which is diagnostic crossover.

Why is it a problem that a substantial number of patients do not fit into the AN or BN categories (a majority of outpatients, as indicated by some studies (Fairburn et al., 2007))? I outlined some problems in a previous post and many of these are echoes and elaborated on in the Wilfley et al. 2007 paper.

  • It is not informative with regard to patient symptoms, treatment or prognosis.
  • It is often not covered by insurance companies.
  • It is not perceived as a “legitimate” or “real” disorder by sufferers, and sometimes clinicians.
  • Researchers use DSM criteria to guide their inclusion and exclusion criteria for various groups in the study (AN, BN)

One of my biggest problems is with the last point above: how valid or “real” are the findings that compare different ED groups, divided into restricting-type, binge-purging type and bulimic groups?

Denise Wilfley echoes this point from a different perspective. In the 2007 paper, she writes that the DSM is used by both clinicians (to guide and provide information for diagnosis, treatment and prognosis) and researchers use it to guide their studies. This, as she points out, can have unintended consequences:

Rigid adoption of DSM definition can (and likely does) hinder investigation of the etiology of eating disorders, because researchers tend to study what is defined. Thus valuable information is not collected, and the current diagnostic criteria are not challenged. Currently defined eating disorder categories are not well validated; as such, they should not be considered definitive answers but rather suggestions subject to change and refinement as evidence-based findings dictate.

I am of the opinion that there is virtually no evidence to show that ED-NOS is a different disorder. Rather, to take again from Wilfley et al.: it is a “mislabelling of the same disorder at different stages.” And perhaps most importantly: “It fails to inform research”, because researchers tend to study AN or BN groups who have shown no or limited history of diagnostic crossover, when this group is actually in the minority of those with eating disorders. While it may be interesting from a purely research and basic science perspective (and I fully support that), I think it paints a wrong picture of what most of us, with EDs, actually experience in the course of the ED.

One problem I’ve always had is that AN and BN patients, if you read a lot of the literature (I will find papers on this a bit later) or even listen to a video, appear to have different personality features. The underlying assumption can be, and often is that these are either predisposing factors or more or less static factors that may be related to, but are disconnected from the eating disorder.

But, in my own experience I’ve found that during the course of my eating disorder, my way of dressing, how structured my day is, my obsessionality about cleanliness and order of things, how impulsive I am, how social and outgoing I am, and all sorts of other things change DRASTICALLY as I go through phases of restrictive eating (usually associated with happier moods, actually, as I feel this is my “natural” and basal state that I fall back into when I’m feeling good) and bulimic episodes (characterized by being more unhappy, stressed and anxious).

My boyfriend once pointed out to me that he knew I was feeling better when he woke up and found our apartment clean and me doing laundry. The state of my house is a reflection of my inner state. Messy = bulimic episodes, clean = restrictive episodes (but lately not accompanies by other features of anorexia, as in, I rarely feel fat at low weights anymore and don’t have body image issues as much as I did in high school). Things change, and evolve. But regardless of where I am on the AN to BN spectrum, I still have an eating disorder (unfortunately). Would I fit into a neat study group? Someone with stable symptoms for a year? A stable diagnosis for even 6 month? No. But I would make an educated guess that I’m probably in the majority, or a sizeable minority.

Anyway, I’ll end here. I apologize this entry is wordy and not much on the science. I’ve had a mediocre last two weeks and not much time to write. I promise more on this later.


Fairburn, C., Cooper, Z., Bohn, K., O’Connor, M., Doll, H., & Palmer, R. (2007). The severity and status of eating disorder NOS: Implications for DSM-V Behaviour Research and Therapy, 45 (8), 1705-1715 DOI: 10.1016/j.brat.2007.01.010

Wilfley, D., Bishop, M., Wilson, G., & Agras, W. (2007). Classification of eating disorders: Toward DSM-V International Journal of Eating Disorders, 40 (S3) DOI: 10.1002/eat.20436

Liked it? Take a second to support Tetyana on Patreon!


Tetyana is the creator and manager of the blog. She has an Honours BSc in Neuroscience and an MSc in Medical Science. She can be reached at tetyana[at]scienceofeds[dot]org.


  1. We did some advocacy around this issue a couple years ago, and got quite a strong response from people with EDs.
    You can still find the signatures and comments from a petition we started before TJP ever existed:

    The Joy Project also conducted an informal online survey regarding ED’d individuals’ perceptions of how they may have been impacted by the current diagnostic criteria structure. A non-insignificant number reported delaying treatment-seeking (thinking they needed to be ‘thinner’ in order to be taken seriously), receiving inadequate treatment, having their illness dismissed by care providers, and having issues with insurance not covering EDNOS.

Leave a Reply

Your email address will not be published.

5 × three =