DSM-IV Diagnostic Criteria for Eating Disorders May Be Too Stringent

Marlene Busko

May 30, 2007

May 30, 2007 — In a large study of psychiatric outpatients, their eating disorders were mainly classed as "not otherwise specified" (NOS) — rather than as bulimia, anorexia, or binge eating — and most patients failed to meet the full diagnostic criteria set forth in the Diagnostic and Statistical Manual of Mental Disorders, 4th Edition ( DSM-IV).

This suggests that DSM-IV diagnostic thresholds for eating disorders are too restrictive, the researchers, led by Mark Zimmerman, MD, from Brown University School of Medicine, in Providence, Rhode Island, write.

The study was presented in a poster at the American Psychiatric Association 2007 Annual Meeting .

Dr. Zimmerman told Medscape that their findings "strongly suggest that there are problems with the diagnostic criteria in the DSM category for eating disorders, because most individuals with an eating disorder don't meet the formal diagnostic criteria." He added that this contrasts with other categories of diagnosis — such as anxiety, personality, and depressive disorders — where "the overwhelming majority" of individuals meet the formal diagnostic criteria. "The conclusion is that there are significant problems with the eating-disorder diagnostic category in the DSM," he said.

Presence of Unspecified Eating Disorders Investigated

The authors explain that in the DSM-IV, bulimia nervosa and anorexia nervosa are officially recognized and formally defined with inclusion and exclusion criteria. Binge eating is listed in the appendix as an eating disorder that requires further study to see whether it should be listed as an official eating disorder in the next edition of the DSM. Any eating disorder that does not fit the full diagnostic criteria for anorexia or bulimia is classed as NOS. The group adds that these criteria evolved from studies of patients seen at treatment centers for eating disorders — patients who likely had the most severe forms of these disorders.

Earlier studies showed that a great number of patients presenting to eating-disorder services had a diagnosis of eating disorder NOS. The group hypothesized that among psychiatric outpatients, failing to meet the diagnostic criteria for eating disorders would be even more common.

They used data from the Rhode Island Hospital Methods to Improve Diagnostic Assessment and Services (MIDAS) project, a large, clinical epidemiologic study. As part of the MIDAS project, 2500 psychiatric outpatients (60.5% women; mean age, 38.2 years) were given a Structural Clinical Interview for DSM-IV.

Based on the interview results, 14.2% (n=354) of the patients were diagnosed as having a current or previous eating disorder:

  • 6.6% (n=165) had a current eating disorder.

  • 2.3% (n=59) were in remission.

  • 5.2% (n=130) had a past diagnosis of eating disorder.

Disorders Perceived to be Clinically Significant

The most frequent eating disorder diagnoses were eating disorder NOS and binge eating.

Rates of DSM-IV Eating Disorders in 2500 Psychiatric Outpatients
Eating Disorder
Current, n (%)
Remission,n (%)
Past, n (%)
Anorexia nervosa
0 (0.0)
12 (0.5)
25 (1.0)
Bulimia nervosa
18 (0.7)
23 (0.9)
29 (1.2)
Binge-eating disorder
63 (2.5)
18 (0.7)
15 (0.6)
Eating disorder NOS
84 (3.4)
6 (0.2)
61 (2.4)

When the categories of binge-eating disorders and eating disorder NOS were combined — as they are in the DSM-IV — then among the psychiatric outpatients with current eating disorders, 89.1% (147/165) were diagnosed with eating disorder NOS.

Most of patients with eating disorder NOS had subthreshold levels of anorexia, bulimia, or binge eating; they failed to meet the sustained amenorrhea criteria for anorexia or the frequency/duration criteria for bulimia or binge eating. Most of these patients still wanted treatment for their disorders, which suggests that they perceived them to be clinically significant.

The authors write that the NOS category was intended to be a residual category, providing clinicians with a diagnostic option for rare cases. "Finding that eating disorder NOS cases predominated suggests a problem with the DSM-IV nomenclature for this class of disorders," they note.

Dr. Zimmerman observed: "All of these data, what they are suggesting is that the thresholds that have been used to diagnose bulimia, anorexia, and binge-eating disorders, if they were lowered, would then incorporate more of these individuals, so that they wouldn't be given an NOS diagnosis."

American Psychiatric Association 2007 Annual Meeting: Abstract NR683. May 19–24, 2007.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.