Background: Binge eating disorder (BED) was first included as a specific diagnosis in the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5). Twelve-month prevalence rates of BED in US adult females and males are estimated at 1.6% and 0.8%, respectively. BED may be under-diagnosed due to a lack of awareness and familiarity in the general medical community. The objective of this project was to develop a patient-reported screener to identify individuals with probable BED for further evaluation by their physicians and/or referral to specialists.
Method: Based on the DSM-5 diagnostic criteria and existing instruments, draft BED Screener (BEDS) items were developed. Items were reviewed with 3 academic and clinical BED experts, revised accordingly, and debriefed in two iterative sets of cognitive debriefing interviews with adults reporting BED-consistent characteristics. In a multisite, cross-sectional, prospective, non interventional study, the BEDS was administered to 97 adults, each required to have a body mass index (BMI) of at least 19, and with approximately half of the sample self-reporting DSM-5 BED characteristics. A structured clinical interview (updated to reflect the DSM-5) was administered to participants, who also completed the BEDS and two additional self-report instruments (the Binge-Eating Scale [BES] and the Rand-36, measuring general health-status). Data were used to evaluate the sensitivity and specificity of candidate classification algorithms, based on participants' BED diagnosis (presence or absence of BED). It was the BEDS developers' intent to optimize sensitivity–to ensure that those who truly have BED would be detected–as well as, to be conservative in identifying a broader range of individuals with probable BED–to ensure that no one would be missed for further evaluation and/or referral.
Results: Following feedback from the experts, a 19-item version (including 6 alternative items) was tested in cognitive debriefing interviews (N = 13), resulting in a 13-item BEDS pilot version. Based on the clinical interview, 16 participants (16.5%) were diagnosed with BED. An algorithm composed of 7 items (BEDS-7) produced optimal sensitivity (100%, 16/16) with reasonable specificity (38.7%, 29/75). These 7 items address the presence of overeating episodes in the last 3 months; lack of control; eating when not hungry; feelings of distress, embarrassment, and disgust; and self-induced vomiting. Based on this algorithm, participants who were correctly identified–true positives–tended to have worse average BES and RAND-36 scores, indicating poorer health status, than participants identified as true negatives.
Conclusions: The BEDS-7 is a sensitive and valid patient-reported screening tool that can be used to identify individuals with probable BED, identify the need for further evaluation, and facilitate referral. Future research should be conducted to test the real-world feasibility and value of the BEDS.
