Reevaluating the Binge Eating Scale cut-off using DSM-5 criteria: analysis and replication in preoperative metabolic and bariatric surgery samples.

BACKGROUND

Metabolic and bariatric surgery is effective for sustained weight loss, but binge eating disorder (BED) can be associated with poorer outcomes and lead to weight regain. A common measure used to screen for BED is the Binge Eating Scale (BES). A BES cut-off score of ≥17 previously identified patients who have a high likelihood of meeting criteria for BED. The Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5), lowered the threshold for meeting criteria for BED, and the psychometrics of the BES need to be reevaluated.

OBJECTIVE

The objective of the current investigation is to evaluate whether alternative cut-scores on the BES result in better classification of BED based on the DSM-5 revision of the the BED diagnostic criteria.

SETTING

Academic medical hospital in the Midwestern United States.

METHODS

Patients (n =1133) seeking metabolic and bariatric surgery were randomly split into 2 samples for validation and replication. The validation sample consisted of 561 patients (30.1% men, 35% non-White). The replication sample consisted of 572 patients who were demographically similar to the first random split sample (e.g., 25.3% men, 34.4% non-White).

RESULTS

Of these patients, 13.5% met DSM-5 criteria for BED in the validation sample and 13.8% met criteria for BED in the replication sample. Lowering the interpretative cut-off to ≥15 on the BES yielded sensitivity values of >.72, specificity values of >.67, and an accurate classification of BED in >.70 of patients across both samples. These classification values were as good as or better than the standard cut-off score of ≥17 in both samples.

CONCLUSIONS

The BES is appropriate to screen for BED in patients who are seeking bariatric surgery. A 2-point decrease in the BES score for clinical interpretation is appropriate-lowering it from 17 to 15 given DSM-5 updates to diagnostic criteria.