Screening for Eating Disorders in Adolescents and Adults

Abstract

ImportanceEating disorders are associated with adverse health and social outcomes.ObjectiveTo review the evidence on screening for eating disorders in adolescents and adults to inform the US Preventive Services Task Force.Data SourcesMEDLINE, Cochrane Library, PsycINFO, and trial registries through December 19, 2020; surveillance through January 1, 2022.Study SelectionEnglish-language studies of screening test accuracy, randomized clinical trials (RCTs) of screening or interventions for eating disorders in populations with screen-detected or previously untreated eating disorders (trials limited to populations who are underweight were ineligible).Data Extraction and SynthesisDual review of abstracts, full-text articles, and study quality. Meta-analysis of test accuracy studies and intervention trials.Main Outcomes and MeasuresTest accuracy, eating disorder symptom severity, quality of life, depression, and harms.ResultsFifty-seven studies were included (N = 10 773); 3 (n = 1073) limited to adolescents (mean or median age, 14-15 years). No study directly evaluated the benefits and harms of screening. Seventeen studies (n = 6804) evaluated screening test accuracy. The SCOFF questionnaire (cut point ≥2) had a pooled sensitivity of 84% (95% CI, 74% to 90%) and pooled specificity of 80% (95% CI, 65% to 89%) in adults (10 studies, n = 3684). Forty RCTs (n = 3969) evaluated interventions for eating disorders; none enrolled a screen-detected population. Lisdexamfetamine for binge eating disorder (4 RCTs; n = 900) was associated with larger reductions in eating disorder symptom severity on the Yale-Brown Obsessive Compulsive Scale modified for binge eating (YBOCS-BE) than placebo (pooled mean difference, −5.75 [95% CI, −8.32 to −3.17]). Two RCTs (n = 465) of topiramate for binge eating disorder found larger reductions in YBOCS-BE scores associated with topiramate than placebo, from −6.40 (95% CI, −8.16 to −4.64) to −2.55 (95% CI, −4.22 to −0.88). Nine pharmacotherapy trials (n = 2006) reported on harms. Compared with placebo, lisdexamfetamine was associated with higher rates of dry mouth, headache, and insomnia, and topiramate was associated with higher rates of paresthesia, taste perversion, confusion, and concentration difficulty. Twenty-four trials (n = 1644) assessed psychological interventions. Guided self-help for binge eating disorder improved eating disorder symptom severity more than control (pooled standardized mean difference, −0.96 [95% CI, −1.26 to −0.67]) (5 studies, n = 391). Evidence on other interventions was limited.Conclusions and RelevanceNo studies directly assessed the benefits and harms of screening. The SCOFF questionnaire had adequate accuracy for detecting eating disorders among adults. No treatment trials enrolled screen-detected populations; guided self-help, lisdexamfetamine, and topiramate were effective for reducing eating disorder symptom severity among referred populations with binge eating disorder, but pharmacotherapies were also associated with harms.