Cutoff scores of the Eating Disorder Examination-Questionnaire for the German population.

OBJECTIVE

The Eating Disorder Examination-Questionnaire (EDE-Q) is one of the most widely used self-report measures for the assessment of eating disorder (ED) symptomatology. However, proposed cutoff scores that may indicate the presence of an ED have been heterogeneous. Therefore, the current study derived cutoff scores from two large samples: one representative for the German population and one composed of persons with EDs at admission to inpatient treatment.

METHOD

Receiver operating characteristic analysis was used with the EDE-Q global score as independent variable and group (controls: n = 2519, patients: n = 2038) as dependent variable. These analyses were also conducted separately with the patient group divided into persons with anorexia nervosa (AN; n = 1456), bulimia nervosa (BN; n = 370), and other EDs (n = 212) and after matching groups for age and sex distribution.

RESULTS

The EDE-Q global score discriminated well between controls and patients (AUC >91%, sensitivity >.84, specificity >.79). A score of 1.6 discriminated best between controls and patients in general and persons with AN in particular. Optimal thresholds for discriminating between controls and persons with BN and other EDs ranged between scores of 1.8 and 2.4.

DISCUSSION

In the German population, cutoff scores between 1.6 and 2.4 may be used to screen for the presence or absence of an ED or evaluate treatment outcome, with slightly higher cutoff scores for persons with BN and other EDs than for persons with AN.

PUBLIC SIGNIFICANCE

Questionnaire scores have little value when it is unclear which scores indicate the likely presence of an ED, as such scores can be used to estimate the prevalence of or screen for EDs in the general population and evaluate outcome at the end of ED treatment. The current study indicates a score around 2 on the EDE-Q as an optimal threshold for this.