Screen for Disordered Eating: Improving the accuracy of eating disorder screening in primary care

Abstract Objective To develop a primary care eating disorder screen with greater accuracy and greater potential for generalizability, compared to existing screens. Design Cross-sectional survey to assess discriminative accuracy of a new screen, Screen for Disordered Eating (SDE), compared to Eating Disorders Screen for Primary Care (EDS-PC) and SCOFF screener, using prevalence rates of Binge Eating Disorder (BED), Bulimia Nervosa (BN), Anorexia Nervosa (AN), and Any Eating Disorder (AED), as measured by the Eating Disorder Examination Questionnaire (EDE-Q). Results The SDE correctly classified 87.2% (CI: 74.3%–95.2%) of BED cases, all cases of BN and AN, and 90.5% (CI: 80.4%–96.4%) of AED cases. Sensitivity estimates were higher than the SCOFF, which correctly identified 69.6% (CI: 54.2%–82.3%) of BED, 77.8% (CI: 40.0%–97.2%) of BN, 37.5% (CI: 8.52%–75.5%) of AN, and 66.1% (CI: 53%–77.7%) of AED. While the EDS-PC had slightly higher sensitivity than the SDE, the SDE had better specificity. The SDE outperformed the SCOFF in classifying true cases, the EDS-PC in classifying true non-cases, and the EDS-PC in distinguishing cases from non-cases. Conclusions The SDE is the first screen, inclusive of BED, valid for detecting eating disorders in primary care. Findings have broad implications to address eating disorder screening in primary care settings.