Background: DSM-5 revisions have been criticized in the popular press for overpathologizing normative eating patterns-particularly among individuals with obesity. To evaluate the evidence for this and other DSM-5 critiques, we compared the point prevalence and interrater reliability of DSM-IV versus DSM-5 eating disorders (EDs) among adults seeking weight-loss treatment.
Method: Clinicians (n = 2) assigned DSM-IV and DSM-5 ED diagnoses to 100 participants via routine clinical interview. Research assessors (n = 3) independently conferred ED diagnoses via Structured Clinical Interview for DSM-IV and a DSM-5 checklist.
Results: Research assessors diagnosed a similar proportion of participants with EDs under DSM-IV (29%) versus DSM-5 (32%). DSM-5 research diagnoses included binge eating disorder (9%), bulimia nervosa (2%), subthreshold binge eating disorder (5%), subthreshold bulimia nervosa (2%), purging disorder (1%), night eating syndrome (6%), and other (7%). Interrater reliability between clinicians and research assessors was "substantial" for both DSM-IV (κ = 0.64, 84% agreement) and DSM-5 (κ = 0.63, 83% agreement).
Conclusion: DSM-5 ED criteria can be reliably applied in an obesity treatment setting and appear to yield an overall ED point prevalence comparable to DSM-IV.