This paper has addressed the neglected DSM-IV diagnosis eating disorder NOS. Two misconceptions appear to keep eating disorder NOS on the margins of eating disorders. The first is the assumption that cases of eating disorder NOS are mild and therefore unimportant. The findings reviewed above indicate that this view is mistaken. The second misconception is that eating disorder NOS is uncommon. Data from eating disorder clinics give the lie to this view (see ), but it is perhaps perpetuated by the “residual” status of NOS diagnoses in general.
We have suggested that two challenges have to be met for the problems of people with eating disorder NOS to get the attention that they deserve. One is that positive diagnostic criteria are needed and we have described a research strategy whereby they could be developed. It has not escaped our attention that doing this would be of value beyond simply defining eating disorder NOS and, in the process, what is an “eating disorder”. For example, it would provide a definition of caseness for epidemiological and clinical purposes and it would provide a new and clinically meaningful way of defining outcome for studies of treatment and natural course. At present most such studies ignore eating disorder NOS as a potential outcome thereby possibly inflating recovery rates. Having what constitutes an eating disorder delineated, with a good outcome being defined as being “over the edge” (i.e., no longer having an eating disorder), would provide a unified and consistent index of remission and recovery that would be the same whatever the eating disorder being studied. It might therefore make redundant the varied and somewhat inconsistent ways of representing outcome that are in use today. We are also aware that the proposed research strategy has broader implications too for it could be used to define the outer boundaries of other classes of psychiatric disorder.
The second challenge involves re-conceptualising the clinical problems that are currently categorised as eating disorder NOS. This is essential if the nosological anomaly of eating disorder NOS is to be resolved. Three solutions have been proposed. In the short term we favour the second solution because the first ignores the fact that many of the cases within eating disorder NOS are of the mixed variety. It involves relaxing the diagnostic criteria for anorexia nervosa and bulimia nervosa to extract the subthreshold cases from eating disorder NOS, the remaining cases being re-classified as cases of mixed eating disorder or BED. We are aware that the introduction of a new eating disorder diagnosis is inconsistent with the conservative spirit of DSM-IV but, as Nielsen and Palmer point out, “There is room for a measure of conservatism but we cannot be satisfied until the EDNOS issue is more adequately addressed” (Nielsen & Palmer, 2003, p. 162
The second solution would have the effect of eliminating the concept of eating disorder NOS, at least for the meantime. The diagnosis would re-appear, however, once specific criteria for the “edges” were formulated (i.e., criteria for what constitutes an eating disorder) since in practice some “cases” of clinical severity would inevitably be encountered that would fall outside the new boundary, however well it was defined. These cases should be modest in number, rendering eating disorder NOS a small residual category, as NOS categories are intended to be.
We acknowledge that this re-classification of the cases within eating disorder NOS is something of a sleight of hand, but it is a sleight of hand with a purpose since it is intended to place these cases in specific and appropriate diagnostic categories. This might enhance the credibility and usefulness of the scheme for classifying eating disorders and, hopefully, it might also facilitate research on these problems including research on their treatment. We believe that this proposal would fulfil the First et al. (2004)
criteria for “clinical utility”.
As regards the “transdiagnostic” solution, we believe that in the longer term it has the most to recommend it. The existing scheme for classifying eating disorders is a historical accident that is a poor reflection of clinical reality. The transdiagnostic solution would encourage and permit the classification of eating disorders to be examined afresh. The collection of good transdiagnostic data, particularly cross-diagnostic information on course and response to treatment, is needed if new clinically informative subdivisions are to be identified.