Screening for Eating Disorders in Adolescents and Adults (PDF)

Screening for Eating Disorders in Adolescents and Adults Evidence Report and Systematic Review for the US Preventive Services Task Force Cynthia Feltner, MD, MPH; Christine Peat, PhD; Shivani Reddy, MD, MS; Sean Riley, MA, MSc; Nancy Berkman, PhD; Jennifer Cook Middleton, PhD; Casey Balio, PhD; Manny Coker-Schwimmer, MPH; Daniel E. Jonas, MD, MPH IMPORTANCE Eating disorders are associated with adverse health and social outcomes. OBJECTIVE To review the evidence on screening for eating disorders in adolescents and adults to inform the US Preventive Services Task Force. DATA SOURCES MEDLINE, Cochrane Library, PsycINFO, and trial registries through December 19, 2020; surveillance through January 1, 2022. STUDY SELECTION English-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 SYNTHESIS Dual review of abstracts, full-text articles, and study quality. Meta-analysis of test accuracy studies and intervention trials. MAIN OUTCOMES AND MEASURES Test accuracy, eating disorder symptom severity, quality of life, depression, and harms. RESULTS Fifty-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 RELEVANCE No 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. JAMA. 2022;327(11):1068-1082. doi:10.1001/jama.2022.1807 Editorial page 1029 Multimedia Related article page 1061 and JAMA Patient Page page 1100 Supplemental content Related article at jamainternalmedicine.com Author Affiliations: Author affiliations are listed at the end of this article. Corresponding Author: Cynthia Feltner, MD, MPH, Cecil G. Sheps Center for Health Services Research, University of North Carolina at Chapel Hill, 725 Martin Luther King Jr Blvd, CB#7295, Chapel Hill, NC 27599 (cindy_f[email protected]). Clinical Review & Education JAMA | US Preventive Services Task Force | EVIDENCE REPORT 1068 (Reprinted) jama.com © 2022 American Medical Association. All rights reserved.  E atingdisordersareconditionsmarkedbyadisturbanceineat- ingoreating-relatedbehaviorsthatimpairfunctioning.1This review focused on common eating disorders that could be asymptomatic or undetected in routine primary care: anorexia nervosa, avoidant/restrictive food intake disorder, bulimia nervosa, bingeeatingdisorder,andotherspecifiedeatingorfeedingdisorder. Estimated lifetime prevalences for anorexia nervosa, bulimia ner- vosa,andbingeeatingdisorderinadultwomenare1.42%,0.46%,and 1.25%, respectively, and are lower in adult men (anorexia nervosa, 0.12%; bulimia nervosa, 0.08%; binge eating disorder, 0.42%).2 In adolescentsaged12to17years,estimatedlifetimeprevalenceforan- orexianervosa,bulimianervosa,andbingeeatingdisorderare0.3%, 1.3%, and 2.3%, respectively, for females and 0.3%, 0.5%, and 1.3% formales.3Estimatedprevalenceforsomedisordersvarybyraceand ethnicity and age category (eTable 1 in the Supplement). Eating disorders are associated with adverse health outcomes which vary by diagnosis, duration, and frequency of certain behav- iors. For bulimia nervosa, purging behaviors (eg, self-induced vomit- ing) can lead to electrolyte disturbances and dental erosion.4 Binge eatingdisordercancontributetoobesity,5andanorexianervosaisas- sociated with morbidity attributed to weight loss and malnutrition.6 Eating disorders are also commonly comorbid with mood and sub- stance abuse disorders.7 Measurementofweight,height,andbodymassindexisroutine in primary care practice and may detect some eating disorders, par- ticularlyanorexianervosa.Disorderswithoutphysicalsymptomsmay go unrecognized, and some individuals experiencing symptoms may not seek care. Routine screening could detect eating disorders early, lead to earlier treatment, and reduce future morbidity. The US Preventive Services Task Force (USPSTF) has not pre- viously made a recommendation on screening for eating disorders. This review evaluated the evidence on screening adolescents and adults for eating disorders for populations and settings relevant to primarycareintheUStoinformarecommendationbytheUSPSTF. Methods Scope of the Review Detailedmethodsareavailableinthefullevidencereport.8Figure1 showstheanalyticframeworkandkeyquestions(KQs)thatguided the review. Data Sources and Searches PubMed/MEDLINE, the Cochrane Library, PsycINFO, and ClinicalTrials.gov were searched for English-language articles pub- lished through June 23, 2020 (eMethods in the Supplement). The searchesweresupplementedwithreferencelistsofpertinentarticles and studies suggested by peer reviewers or public comment respon- dents.SinceJune2020,ongoingsurveillancewasconductedthrough articlealertsandtargetedsearchesofjournalstoidentifymajorstud- ies published in the interim that may affect the conclusions or under- standing of the evidence and the related USPSTF recommendation through January 1, 2022. No relevant studies were identified. Study Selection Two investigators independently reviewed titles, abstracts, and full-text articles using prespecified eligibility criteria (eMethods in the Supplement). Disagreements were resolved by consensus. For all KQs, English-language studies of adolescents and adults 10 years or older conducted in settings generalizable to primary care, includ- ing school-based health centers, and in countries categorized as “very high” on the United Nations Human Development Index were included.9 The scope of this review was focused on populations with eating disorders unlikely to be detected in the context of rou- tine primary care. Studies limited to populations with physical signs of eating disorders (eg, populations who are underweight) were ineligible because eating disorders would be part of the diagnostic assessment for individuals presenting with an abnormally low body weight. For KQ1 and KQ3 (direct evidence of benefits and harms of screening), randomized clinical trials (RCTs) comparing screening with no screening in asymptomatic populations were eligible. For KQ2 (screening test accuracy), studies comparing a screening test with a diagnostic reference standard for eating disorders (struc- tured or semistructured diagnostic interview or diagnostic ques- tionnaire) were eligible. Eligible screening tests included those fea- sible for use in primary care settings (brief, easy to interpret) and designed to detect any eating disorder or specific disorders (eg, binge eating disorder); longer questionnaires (eg, the 26-item Eating Attitudes Test) were excluded. For KQs on benefits and harms of treatment (KQ5 and KQ6), RCTs enrolling populations with screen-detected eating disorders, or populations from specialty settings or via advertisements who had not been previously treated for eating disorders, were included. Eligible treatments included psychological interventions (eg, cognitive behavioral) delivered in a group, individual, or family- based format, including self-help interventions, or pharmaco- therapy with US Food and Drug Administration–approved medica- tions. Eligible RCTs had to compare treatment with an inactive control (ie, no treatment, wait-list, minimal intervention [eg, brief education about eating disorders], or placebo). RCTs evaluating combined psychological and pharmacotherapy interventions were eligible if they included an inactive control group. Eligible outcomes for KQs on the benefits of screening or treat- mentincludedmeasuresofeatingdisordersymptomseverity,health- related quality of life or function, depression, and others. Intermedi- ateoutcomessuchasmeanchangeinfrequencyofspecificbehaviors (eg,changeinfrequencyofbingeeatingepisodes)wereexcluded.Eli- gibleoutcomesforKQ3(harmsofscreening)includedincreasedanxi- ety, labeling, and stigma associated with screening; for KQ5 (harms of interventions), outcomes included any harms attributed to inter- ventions, such as harms associated with medications. Data Extraction and Quality Assessment For each study, 1 investigator extracted information about popula- tions, tests or interventions, comparators, outcomes, settings, and designs,andasecondinvestigatorreviewedtheinformationforcom- pleteness and accuracy. Two investigators independently assessed eachstudy’smethodologicalquality,usingtheCochraneRiskofBias Tool (RoB 2.0)10 and the QUADAS-2 (Quality Assessment of Diag- nostic Accuracy Studies 2) for studies of test accuracy.11 Disagree- ments in quality ratings were resolved through discussion or inde- pendent assessment from a third senior investigator. Risk-of-bias assessments using these instruments were translated into an over- all study quality rating of good, fair, or poor using predefined crite- ria developed by the USPSTF and adapted for this topic (eMethods USPSTF Review: Screening for Eating Disorders in Adolescents and Adults US Preventive Services Task Force Clinical Review & Education jama.com (Reprinted) JAMA March 15, 2022 Volume 327, Number 11 1069 © 2022 American Medical Association. All rights reserved.  in the Supplement). Individual study quality ratings are reported in eTables 4-7 in the Supplement. Data Synthesis and Analysis Findings for each KQ were summarized in tables, figures, and narra- tiveformat.ForKQ2,pooledsensitivitiesandspecificitiesforscreen- ing tests were calculated using a hierarchical summary receiver operatingcharacteristiccurveanalysiswhenatleast4similarstudies were available. For KQ4, random-effects restricted maximum likeli- hoodmodelswereconductedoncontinuousmeasuresofeatingdis- orderanddepressionsymptomseverity(analyzingstandardizedmean difference or unstandardized mean difference in change between groups) when at least 3 similar studies were available. When studies reported more than 1 continuous outcome for eating disorder symptomseverity,theoutcomemostcommonlyreportedbysimilar studiesinpooledestimateswaspreferentiallyselected.Statisticalsig- nificancewasassumedwhen95%CIsofpooledresultsdidnotcross the null. All testing was 2-sided. Comprehensive Meta-Analysis ver- sion3.4(BiostatInc)andStataversion16(StataCorp)12 wereusedto conduct all quantitative analyses. The overall strength of the evidence for each KQ was assessed as high, moderate, low, or insufficient based on the overall quality of thestudies,consistencyofresultsbetweenstudies,precisionoffind- ings, risk of reporting bias, and limitations of the body of evidence, using methods developed for the USPSTF (and the Evidence-based PracticeCenterprogram).13Additionally,theapplicabilityofthefind- ings to US primary care populations and settings was assessed. Dis- crepancies were resolved through consensus discussion. Results A total of 57 studies (59 articles) with 10 773 participants were in- cluded (Figure 2). Benefits of Screening KeyQuestion1.Doesscreeningforeatingdisordersinadolescents and adults improve health outcomes, including for specific sub- groups of interest? No eligible studies were identified. Screening Accuracy Key Question 2. What is the accuracy of primary care–relevant screening tests for eating disorders in adolescents and adults, in- cluding for specific subgroups of interest? Tengood-quality14-25and7fair-quality20,26-31studies(18articles)as- sessed the accuracy of 9 screening questionnaires; 5 were de- signedtodetectanyeatingdisorder,14,15,17,18,20,21,24-28,31and4were designed to detect eating disorders characterized by binge eating (bulimia nervosa or binge eating disorder).19,23,29,30 Detailed char- acteristicsarereportedinTable1.14-32Ofthestudies,11assessedthe 5-item SCOFF questionnaire. (Some experts recommend not con- sidering SCOFF an acronym since it is based on terminology from signaling questions that may not translate well [eg, “Have you re- cently lost more than One stone in a 3-month period?”].) Refer- ence standards used to evaluate screening test accuracy included a diagnostic clinical interview or a longer diagnostic questionnaire. Figure 1. Analytic Framework: Screening for Eating Disorders in Adolescents and Adults Key questions Does screening for eating disorders in adolescents and adults improve health outcomes, including for specific subgroups of interest? 1 What is the accuracy of primary care-relevant screening tests for eating disorders in adolescents and adults, including for specific subgroups of interest? 2 How effective are interventions for improving health outcomes in screen-detected or previously untreated adolescents and adults with eating disorders, including for specific subgroups of interest? 4 What are the harms of screening for eating disorders in adolescents and adults, including for specific subgroups of interest? 3 What are the harms of interventions for eating disorders, including for specific subgroups of interest? 5 Adolescents and adults Morbidity Mortality Health outcomes 2 Harms of screening 3 4 Harms of interventions 5 Screening Intervention Early detection of eating disorders 1 Evidence reviews for the US Preventive Services Task Force (USPSTF) use an analytic framework to visually display the key questions that the review will address to allow the USPSTF to evaluate the effectiveness and safety of a preventive service. The questions are depicted by linkages that relate interventions and outcomes. For additional information see the USPSTF Procedure Manual.13 Clinical Review & Education US Preventive Services Task Force USPSTF Review: Screening for Eating Disorders in Adolescents and Adults 1070 JAMA March 15, 2022 Volume 327, Number 11 (Reprinted) jama.com © 2022 American Medical Association. All rights reserved.  Most studies enrolled participants from university set- tings17,18,20,22,25,31 and outpatient clinics (primary care,15,19,24-26 psychiatry,14 and obesity clinics).23,28,30 Six studies were set in the US15,18-20,26,29; others were set in the UK,24,25 Taiwan,14 Malaysia,17 and various European countries.21-23,27,28,30,31 Most studies en- rolled only females15-18,22,24,28,31 or predominantly females (>60%)14,20,23,25,29,30;2enrolledamajorityofmales.19,21Twostud- ieslimitedtoadolescentswithameanormedianageof14years,and all others enrolled adults (mean age, 20-63 years).21,23 In 4 studies evaluating a screening tool for bulimia nervosa or binge eating dis- order, prevalence (based on the reference standard) ranged from 8% to 22%19,23,29,30; the prevalence of any eating disorder ranged from 2% to 46%. Table 2 summarizes results of screening test accuracy. In stud- ies of adults (10 studies, n = 4348), the SCOFF questionnaire (cutpoint�2)hadapooledsensitivityof84%(95%CI,74%to91%) andpooledspecificityof80%(95%CI,65%to89%)(Table2;eFig- ure1intheSupplement).Sevenstudies(n = 3424)assessedtheac- curacyatahighercutpoint(�3)14,17,22,24,26,28;pooledsensitivitywas lower at 69% (95% CI, 56% to 80%), and specificity was higher at 90% (95% CI, 69% to 98%) (Table 2; eFigure 2 in the Supple- ment).OnestudyevaluatedtheSCOFFquestionnaire(cutpoint�2) amongadolescents(n = 954;meanage,14years)21;sensitivitywas 73% (95% CI, 63% to 83%), and specificity was 78% (95% CI, 75% to 80%). Eightotherscreeningquestionnaireswereassessedacross8in- cluded studies.15,18,19,23,25,29-31 One, the EDS-PC (5 items, devel- opedforuseinprimarycare25)wasevaluatedin2studies(n = 627) enrollingdifferentpopulations(Table2);sensitivityrangedfrom97% to 100%, and specificity ranged from 40% to 71%.15,25 All other screeningquestionnaireswereassessedby1studyeach;resultsare summarized in Table 2.19,23,29,30 Harms of Screening Key Question 3. What are the harms of screening for eating disor- ders in adolescents and adults, including for specific subgroups of interest? No eligible studies were identified. Benefits of Treatment Key Question 4. How effective are interventions for improving health outcomes in screen-detected or previously untreated ado- lescentsandadultswitheatingdisorders,includingforspecificsub- groups of interest? Figure 2. Literature Search Flow Diagram: Screening for Eating Disorders in Adolescents and Adults 16 700 Citations identified through database search 8927 PubMed 5453 PsycINFO 2320 Cochrane 263 Additional citations identified through other sources 262 ClinicalTrials.gov 1 Public comment 13 586 Citations excluded at title and abstract stage 1392 Excluded 279 Wrong or no outcome 261 Wrong or no comparator 231 Wrong population 215 Wrong screening test 174 Wrong study design 76 Wrong intervention 49 Poor quality 30 Wrong setting 21 Abstract only 18 Wrong condition 14 Intermediate outcome only 13 Not original research 7 Wrong country 4 Wrong language/non-English-language 0 Articles included for KQ1 0 Articles included for KQ3 18 Articles (17 studies) included for KQ2 41 Articles (40 studies) included for KQ4 8 Articles (9 studies) included for KQ5 59 Articles (57 studies) included in qualitative synthesis of systematic review 1451 Full-text articles assessed for eligibility 15 037 Citations screened after duplicates removed KQ indicates key question. USPSTF Review: Screening for Eating Disorders in Adolescents and Adults US Preventive Services Task Force Clinical Review & Education jama.com (Reprinted) JAMA March 15, 2022 Volume 327, Number 11 1071 © 2022 American Medical Association. All rights reserved.  Table 1. Characteristics of Included Studies for KQ2 Source, country Quality Screener Reference standard: eating disorder diagnoses assessed Recruitment setting Population Eating disorder prevalence, %a Age, mean (SD), y Female, % Non-White, % BMI, mean (SD) Cohort studies Lui et al,14 2015 Taiwan Good SCOFF SCID (DSM-IV): anorexia nervosa, bulimia nervosa, BED, EDNOS Outpatient psychiatric clinics 1541 Adults (18-45 y) recruited at their first outpatient psychiatric visit Any eating disorder: 16 31 (7.9) 61 NRb 22.2 (5.4) Graham et al,18,32 2019 US Good SWED EDE (DSM-5): anorexia nervosa, bulimia nervosa, BED University campuses 549 College-age women (18-25 y) responding to recruitment ads and flyers for an eating disorder prevention trial Any eating disorder: 19c,d 21 (1.97) 100 44d 24.5 (5.02) Dorflinger et al,19 2017 US Good VA-BES QEWP-R: BED VHA medical center 116 Veterans recruited at primary care–based weight management group BED: 8 62 (8.73) 11 26 37.9 (7.35) Rosenvinge et al,31 2001 Norway Fair EDS-5 SCID (DSM-III-R): any eating disorder University campuses 51 College-age women (20-42 y) recruited at their teaching and nursing colleges CED: 20d 25.2 (5.33) 100 NR NR Mond et al,26 2008 US Fair SCOFF EDE Primary care practices 147 Adult women (18-40 y) recruited at their primary care visit Any eating disorder: 17 28 (6.50) 100 12 28.10 (7.20) Cotton et al,25 2003 UK Good SCOFF, EDS-PC QEDD University campuses and primary care 225 Students (18-65 y) recruited from posters and lecture announcements and adults (18-65 y) recruited at a primary care visit Any eating disorder: 12 29 77 NR 22 Lähteenmäki et al,27 2009 Finland Fair SCOFF SCID (DSM-IV): anorexia nervosa, bulimia nervosa, EDNOS Households 541 Young adults recruited via mail Current anorexia nervosa, bulimia nervosa, EDNOS: 1d Lifetime anorexia nervosa, bulimia nervosa, EDNOS: 4d NR NR NR NR Cross-sectional studies Maugen et al,15 2018 US Good EDS-PC, SCOFF, SDE EDE-Q (DSM-5e): anorexia nervosa, bulimia nervosa, BED VHA medical center 402 Female veterans (18-70 y) responding to mailed questionnaires Any eating disorder: 16d 49 (NR)d 100 52f NR Chamay- Weber et al,23 2017 Switzerland Good ADO-BED SCID (DSM-IV): BED Outpatient pediatric obesity center 94 Adolescents (12-18 y) recruited at their outpatient pediatric visit BED: Sub: 28 Full: 22 Overall: 50 Median (range): 14 (11-18) 60 NR NR Luck et al,24 2002; Hill et al,16 2010 UK Good SCOFF Clinical interview (DSM-IV): anorexia nervosa, bulimia nervosa, EDNOS Primary care practices 341 Women (18-50 y) attending primary care practices Any eating disorder: 4d NR 100 NR NR Siervo et al,28 2005 Italy Fair SCOFF “Clinical diagnosis” (DSM-IV): bulimia nervosa, BED Outpatient diet clinics 162 Women (16-35 y) recruited at an outpatient dietetic clinic Any eating disorder: 46d 24d 100 NR 29.6d (continued) Clinical Review & Education US Preventive Services Task Force USPSTF Review: Screening for Eating Disorders in Adolescents and Adults 1072 JAMA March 15, 2022 Volume 327, Number 11 (Reprinted) jama.com © 2022 American Medical Association. All rights reserved.  Table 1. Characteristics of Included Studies for KQ2 (continued) Source, country Quality Screener Reference standard: eating disorder diagnoses assessed Recruitment setting Population Eating disorder prevalence, %a Age, mean (SD), y Female, % Non-White, % BMI, mean (SD) Parker et al,20 2005 US Good SCOFF EDE-Q (DSM-IV): anorexia nervosa, bulimia nervosa, EDNOS University health center 297 Adults (20-51 y) recruited at their campus health visit Any eating disorder: 20 <23 y: 10 23-26 y: 66 >26 y: 23 72 33 Range: 16-44 Ricca et al,30 2000 Italy Fair BES SCID (DSM-IV): bulimia nervosa, BED Outpatient clinic for metabolic diseases 344 Patients recruited at an outpatient clinic for metabolic diseases including obesity BED: 8 43.5 (13.6) 83 NR 35.8 (6.1) Wan Wahida et al,17 2017 Malaysia Good SCOFF EAT-26 University 292 Undergraduate students (18-22 y) who understood English Any eating disorder: 11 20 (0.5) 65 Malay: 44 Chinese: 42 Indian: 14 NR Garcia et al,22 2010 France Good SCOFF MINI (DSM-IV-TR): any eating disorder, anorexia nervosa, bulimia nervosa University clinic 400 Female undergraduate students (18-35 y) Any eating disorder: 9 21 (2.5) 100 14f 21.98 (3.5) Muro-Sans et al,21 2008 Spain Good SCOFF EDI-2 Primary and secondary schools 954 Adolescents (10-17 y) recruited from schools Any eating disorder: 8d 14 (1.31) 49 NR NR Striegel- Moore et al,29 2010 US Fair PHQ-ED EDE (DSM-IV): bulimia nervosa, BED Health maintenance organization 348 Adults (18-35 y) selected from the EHR of an HMO via letter Bulimia nervosa or BED: 8d,e 28 (5.38) 82 13 NR Abbreviations: ADO-BED, Adolescent Binge-Eating Disorder Questionnaire; BED, binge eating disorder; BES, Binge Eating Scale; BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); CED, clinical eating disorder; DSM-III-R, Diagnostic and Statistical Manual of Mental Disorders (Third Edition Revised); DSM-IV, Diagnostic and Statistical Manual of Mental Disorders (Fourth Edition); DSM-IV-TR, Diagnostic and Statistical Manual of Mental Disorders (Fourth Edition, Text Revision); DSM-5, Diagnostic and Statistical Manual of Mental Disorders (Fifth Edition); EAT-26, Eating Attitudes Test; EDE, Eating Disorder Examination; EDE-Q, Eating Disorder Examination Questionnaire; EDI-2, Eating Disorder Inventory 2; EDNOS, eating disorder not otherwise specified; EDS-5, Eating Disturbance Scale 5; EDS-PC, Eating Disorder Screen for Primary Care; EHR, electronic health record; HMO, health maintenance organization; KQ, key question; MINI, Mini International Neuropsychiatric Interview; NR, not reported; PHQ-ED, eating disorder module of the Patient Health Questionnaire; QEDD, Questionnaire for Eating Disorder Diagnoses; QEWP-R, Questionnaire of Eating and Weight Patterns–Revised; SCID, Structured Clinical Interview for DSM Disorders; SDE, Screen for Disordered Eating; SWED, Stanford-Washington University Eating Disorder screen; VA-BES, Veterans Affairs Binge Eating Screener; VHA, Veterans Health Administration. a Full refers to meeting the full diagnostic criteria for a given eating disorder; sub refers to a subthreshold condition definition. bConducted in Taiwan and required to understand Mandarin. c Also conducted analysis including subthreshold bulimia nervosa, BED, and purging disorder (in addition to threshold anorexia nervosa, bulimia nervosa, and BED). dComputed by data abstractors. e BED defined as an average of 1 or more objective binge episodes per week without compensatory or purging behaviors. f Enrolled all screen-positive participants and a random sample of those who screened negative. USPSTF Review: Screening for Eating Disorders in Adolescents and Adults US Preventive Services Task Force Clinical Review & Education jama.com (Reprinted) JAMA March 15, 2022 Volume 327, Number 11 1073 © 2022 American Medical Association. All rights reserved.  Fortyfair-togood-qualityRCTs(n = 3969)oftreatmentforeat- ing disorders were included—18 (19 publications) assessing pharmacotherapy33-51 and 24 assessing therapy (eTables 9 and 10 intheSupplement);ofthese,2assessedbothpharmacotherapyand therapy interventions compared with a control.50,51 All enrolled populationsreferredorrecruitedtotreatment;noneenrolledpopu- lations detected by screening in primary care. In 17 studies describ- ingraceorethnicity,1waslimitedtoLatinasonly,522enrolledapopu- lation that was 54% to 55% non-White (from the US),53,54 and all others enrolled a majority of White participants. Among18RCTsevaluatingthebenefitofpharmacotherapycom- pared with placebo over 6 to 16 weeks (eTable 9 in the Supple- ment), 14 enrolled populations with binge eating disorder (defined byDiagnosticandStatisticalManualofMentalDisorders[FourthEdi- tion] or Diagnostic and Statistical Manual of Mental Disorders [Fifth Edition] [DSM-5] criteria), and 4 enrolled populations with bulimia nervosa defined by Diagnostic and Statistical Manual of Mental Dis- orders(ThirdEdition)criteria.41,43,48,51Allenrolledadults,and1trial (n = 50) enrolled both adults and adolescents as young as 16 years (mean age, 25 years).55 Detailed characteristics of populations and pharmacotherapy are reported in eTable 9 in the Supplement. Four RCTs (described in 3 publications) compared lisdexamfet- amine with placebo among adults with binge eating disorder.33,35,37 AllmeasuredbingeeatingdisordersymptomseverityusingtheYale- BrownObsessiveCompulsiveScalemodifiedforbingeeating(YBOCS- BE); for doses ranging from 50 to 60 mg/d, the pooled mean differ- ence in change from the baseline score over 11 to 12 weeks (4studies,n = 900)was−5.75(95%CI,−8.32to−3.17)(Figure3).This difference falls within the range considered a minimum clinically im- portant change on the YBOCS-BE (−4 to −17).56 Other eligible out- comes were reported by only 1 or 2 studies each (eTable 9 in the Supplement). Two trials of topiramate (n = 465) measured reduc- tionineatingdisordersymptomseverityusingtheYBOCS-BEover14 to 15 weeks (mean or median dose, 212-300 mg/d).44,48 Both found significantimprovementfavoringtopiramate(Figure3);1foundadif- ferencebetweengroupsinmeanchangefrombaselinescore(−6.50) within the range considered a minimum clinically important change (−4to−17),44andtheotherfoundasmallerdifferenceinmeanscore change (−2.55).48 Five RCTs assessed a selective serotonin reuptake inhibitor (SSRI) for improving binge eating disorder, including fluoxetine (2 studies)38,50 and 1 study each of fluvoxetine,42 sertraline,40 and escitalopram.46 None selected participants based on the presence of comorbid depression; however, in 4 trials prevalence of lifetime depressionrangedfrom37%to77%,38,40,46,50andin3trialspreva- lenceofcurrentmajordepressionrangedfrom18%to25%.38,40,46 Only 2 trials measured eating disorder symptom severity (the Eat- ing Disorder Examination–Questionnaire [EDE-Q] and YBOCS-BE); although both found a reduction in symptom scores favoring SSRIs (eFigure 3 in the Supplement), results were imprecise. All reported on change in depression symptoms (eFigure 3 in the Supplement); SSRIs were associated with a larger reduction in depression symp- tom scores than placebo (pooled standardized mean difference [SMD], −0.6 [95% CI, −0.90 to −0.33]) (5 studies; n = 208).36,39,45 Three trials assessed fluoxetine for populations with bulimia ner- vosaandfoundinconsistentresultsforeatingdisordersymptomse- verity and depression (eTable 12 in the Supplement). One trial each evaluated duloxetine,39 bupropion,36 and imipramine45 for popu- lations with binge eating disorder, and 1 evaluated desipramine for bulimia nervosa41 (eTable 9 in the Supplement); none found a sig- nificant differences between groups on measures of eating disor- der symptom severity or depression. Twenty-four trials (n = 1644) assessed the benefit of a psycho- logical intervention compared with an inactive control (eTable 10 in the Supplement).50-55,57-74 Most enrolled populations with binge eating, either binge eating disorder or bulimia nervosa with recur- rent binge eating behavior; 1 trial enrolled those with bulimia ner- vosa without mention of binge eating,73 and 3 enrolled women with any DSM-5 eating disorder.55,57,62 One trial (n = 25) was lim- ited to adolescents (mean age, 15 years),74 1 (n = 82) enrolled both adults and adolescents (as young as 14 years),55 and all others enrolled adults only. Table 2. Summary of Accuracy for Included Screening Tests (KQ2) Screener (cut point) Eating disorder diagnosis No. of studies (No. of participants) % (95% CI) LR+ (95% CI) LR– (95% CI) Sensitivity Specificity SCOFF ≥3 Any 7 (2749) Pooled: 69 (56-80) Pooled: 90 (69-98) Pooled: 7.3 (2.2-24.0) Pooled: 0.34 (0.25-0.46) ≥2 Any 10 (3684) Pooled: 84 (74-90) Pooled: 80 (65-89) Pooled: 4.1 (2.3-7.3) Pooled: 0.20 (0.12-0.33) SWED (>59)18 Any 1 (549) 80 (NR) 82 (NR) NR NR EDS-PC (≥2)15,25 Any 2 (627) 97 (88-100) 40 (35-46) NR NR 100 (90-100) 71 (64-77) SDE (≥2)15 Any 1 (402) 91 (80-96) 58 (80-96) NR NR EDS-5 (≥16)31 Any 1 (51) 90 (NR) 88 (NR) NR NR PHQ-ED (NA)29 BN, BED 1 (348) 100 (NR) 30 (NR)a ADO-BED (NA)23 BED 1 (94 adolescents) 100 (NR) 27 (NR) NR NR VA-BES (≥1)19 BED 1 (162) 89 (NR) 65 (NR) NR NR BES (≥17)30 BED 1 (344) 85 (NR) 75 (NR) NR NR Abbreviations: ADO-BED, Adolescent Binge-Eating Disorder Questionnaire; BED, binge eating disorder; BES, Binge Eating Scale; BN, bulimia nervosa; EDS-5, Eating Disturbance Scale 5; EDS-PC, Eating Disorder Screen for Primary Care; KQ, key question; LR, likelihood ratio; NR, not reported; PHQ-ED, eating disorder module of the Patient Health Questionnaire; SDE, Screen for Disordered Eating; SWED, Stanford-Washington University Eating Disorder screen; VA-BES, Veterans Affairs Binge Eating Screener. a Value calculated based on individual cell frequencies differs from reported specificity value reported in study (91.7% vs 27.7%, respectively). Clinical Review & Education US Preventive Services Task Force USPSTF Review: Screening for Eating Disorders in Adolescents and Adults 1074 JAMA March 15, 2022 Volume 327, Number 11 (Reprinted) jama.com © 2022 American Medical Association. All rights reserved.  Included trials focused on a variety of psychological inter- ventions (eTable 10 in the Supplement); most evaluated a form of self-help based on cognitive behavioral therapy or other strate- gies, designed to help participants cope with eating disorder symptoms.51-55,59-61,63,66,68,72,75 Seven trials evaluated a type of grouptherapy57,65,67,69-71,73and4evaluatedaformofindividualcog- nitivebehavioraltherapy.50,62,64,74eTable11intheSupplementpro- vides additional detail related to the intervention approach, com- ponents, and intensity. Thirteen trials evaluated a self-help intervention, 7 assessed a form of “guided” self-help,52,58,59,61,63,68,72 and 7 assessed an “unguided” self-help intervention.51,53-55,60,63,66 One trial com- pared both guided and unguided self-help interventions with a control.63Guidedinterventionsincludedongoingsupportandguid- ance, for example, several brief (25- to 30-minute) individually guidedsessions,regularemailcontactforsupport,52,59,61,63,72orin- dividual feedback on assignments.58 Unguided interventions in- volved providing the intervention materials with instructions only. Guided self-help was associated with a larger reduction in eat- ing disorder severity than control (measured by the EDE or EDE-Q) over 12 to 24 weeks (pooled SMD, −0.96 [95% CI, −1.26 to −0.67] [5 studies; n = 391]) (Figure 4). Results from trials of unguided self- help (6 studies; n = 368) were consistent in favoring self-help, but pooledestimateswerenotstatisticallysignificant(SMD,−0.18[95% CI, −0.38 to 0.03]) (Figure 4). For measures of depression, pooled results demonstrated larger reductions in mean scores compared with controls for both guided self-help (SMD, −0.73 [95% CI, −1.04 to −0.43]; 4 studies; n = 324) and unguided self-help (SMD, −0.37 [95%CI,−0.68to−0.05];3studies;n = 156).Fewtrialsofself-help measured other eligible outcomes. Seven RCTs assessed a group-based psychological interven- tion for binge eating disorder and bulimia nervosa with recurrent bingeeatingusingdifferenttherapeuticapproachesandnumberof sessions(eTable11intheSupplement).57,65,67,69-71,73Grouptherapy (7studies;n = 253)wasassociatedwithlargerreductionsindepres- sionscoresfrombaselinethaninactivecontrol(pooledSMD,−0.48 [95% CI, −0.69 to 0.27]). Three trials of group therapy measured eating disorder symptom severity using the EDE-Q and found in- consistentresults(eFigure4intheSupplement).65,67Fourtrialsas- sesseddifferentformsofindividualtherapyamongadultsandfound mixed results (eFigure 4 in the Supplement).50,62,64 One trial of in- dividual cognitive behavioral therapy in adolescents (n = 25; mean age,15)foundnosignificantdifferencesbetweengroupsat12or24 weeks on depression symptoms and psychosocial functioning.74 Harms of Treatment KeyQuestion5.Whataretheharmsofinterventionsforeatingdis- orders, including for specific subgroups of interest? No included studies of psychological interventions reported on harms. Nine studies of pharmacotherapy reported various harms associated with 4 medications, including lisdexamfetamine (4 studies),33,35,37 topiramate (2 studies),44,48 fluoxetine (2 studies),38,43 and escitalopram.46 Characteristics are described in KQ4 and eTable 9 in the Supplement. In 1 trial of lisdexamfetamine (n = 259) over 11 weeks, 1 partici- pant died during the study, and postmortem toxicology analysis found that methamphetamine/amphetamine levels were consis- tent with a methamphetamine overdose (death was not attributed to the study drug).37 Across all 4 trials of lisdexamfetamine, treat- ment-emergent harms were higher for the treatment groups than the placebo groups; commonly reported harms were dry mouth, insomnia, and jitteriness (eTable 13 in the Supplement).33,35,37 Two trials of topiramate (duration, 14-16 weeks) found significantly higherratesofparesthesiaandtasteperversion44,48associatedwith topiramate than placebo. One trial found significantly higher rates ofdifficultyconcentrating44andtheotherfoundsignificantlyhigher rates of confusion.48 In 3 trials of SSRIs, 1 found significantly higher rates of several harms in the fluoxetine group than in the placebo group (eTable 13 in the Supplement), such as insomnia, nausea, and tremor.43 The other2trialsreportednosignificantdifferencesbetweengroupsfor any adverse effects over 6 weeks.38,46 Figure 3. Results of Randomized Clinical Trials of Lisdexamfetamine and Topiramate vs Placebo for Binge Eating Disorder (KQ4) –10 5 0 Mean difference (95% CI) –5 Favors medication Favors placebo Outcome Weeks Mean dose, mg/d No. of participants Medication Placebo Source Lisdexamfetamine (eating disorder symptom severity) Mean difference (95% CI) YBOCS-BE 12 25 60 25 Guerdjikova et al,33 2016 –2.80 (–7.28 to 1.68) YBOCS-BE 12 190 57 184 McElroy et al,35 2016 –7.40 (–8.92 to –5.88) Topiramate (eating disorder symptom severity) YBOCS-BE 16 202 300 202 McElroy et al,44 2007 –6.40 (–8.14 to –4.66) YBOCS-BE 14 30 212 31 McElroy et al,48 2003 –2.55 (–4.29 to –0.81) Topiramate (depression, anxiety) HAM-D 14 30 212 31 McElroy et al,48 2003 –0.55 (–1.57 to 0.46) MADRS 16 202 300 202 McElroy et al,44 2007 –0.50 (–1.79 to 0.79) HAM-A 16 202 300 202 –0.60 (–1.52 to 0.32) YBOCS-BE 12 174 58 176 –7.94 (–9.51 to –6.37) YBOCS-BE 11 64 50 62 McElroy et al,37 2015 –3.25 (–5.61 to –0.89) Heterogeneity: τ2 = 5.31, I2 = 82.75%, H2 = 5.80 –5.75 (–8.32 to –3.17) HAM-A indicates Hamilton Anxiety Rating Scale; HAM-D, Hamilton Rating Scale for Depression; KQ, key question; MADRS, Montgomery-Åsberg Depression Rating Scale; YBOCS-BE, Yale-Brown Obsessive Compulsive Scale modified for binge eating. USPSTF Review: Screening for Eating Disorders in Adolescents and Adults US Preventive Services Task Force Clinical Review & Education jama.com (Reprinted) JAMA March 15, 2022 Volume 327, Number 11 1075 © 2022 American Medical Association. All rights reserved.  Discussion Thissystematicreviewevaluatedevidencerelevanttoscreeningfor eating disorders in adults and adolescents. A summary of findings, including an assessment of the strength of evidence for each KQ, is presentedinTable3.Todate,thereisnodirectevidencefromtrials comparingthebenefitsandharmsofroutinescreeningvsnoscreen- ing.Thus,thisreviewanswered2questions:howwellscreeningde- tects eating disorders and how effective are interventions at treat- ing eating disorders among populations with screen-detected or previously untreated eating disorders. Screening tools are available for clinical practice that may rea- sonably identify adults with eating disorders, primarily the SCOFF questionnaire. Other tools were assessed by only 1 study each, lim- iting the ability to make stronger conclusions about screening test accuracy.TheestimatesofSCOFFscreeningtestaccuracywerede- rived from populations with a current prevalence of eating disor- ders ranging from 4% to 46% based on the reference standard, higher than recent estimates of eating disorders in the US. Poten- tial harms of screening include false-positive screening results that lead to unnecessary referrals or labeling. Based on the pooled esti- matesofSCOFFaccuracyfordetectinganyeatingdisorder(Table2) among adults (10 studies, 4348 participants), the expected rate of false-positive test results would be 20%. Most RCTs evaluating interventions for eating disorders were limited to adult women with binge eating disorder and bulimia ner- vosa, enrolled treatment-seeking populations (either respondents toadvertisementsorreferrals),andmeasuredoutcomesoverarela- tively short duration. Some recruited participants using ads that in- dicatedtrialsoftreatmentforbingeeatingandobesity.Bothlisdex- amfetamine and topiramate were effective in reducing eating disorderseverityamongadultswithbingeeatingdisorderbutwere also associated with various harms. Few trials of SSRIs reported an eligiblehealthoutcomespecifictoeatingdisordersymptoms;how- ever, results of 5 trials enrolling adults with binge eating disorder found consistent improvement in depression symptoms associ- ated with various SSRIs. Although trials did not enroll participants based on depression status, lifetime depression rates ranged from 37% to 77% in 4 trials reporting on mental health comorbidity. Whetherimprovementondepressionscoresindicatesimprovedeat- ing disorder symptom severity is not clear. Among the 24 trials assessing psychological interventions, guided self-help improved eating disorder symptom severity and depressive symptoms among adults with binge eating disorder; results for unguided self-help were consistent in direction of effect, Figure 4. Results of Randomized Clinical Trials of Self-help Interventions for Eating Disorders (KQ4) –2 .5 0 Standardized mean difference (95% CI) –.5 –1 –1.5 Favors self-help Favors placebo Intervention type Measure Weeks Source Guided self-help (eating disorder symptom severity) Standardized mean difference (95% CI) DBT EDE-Q 24 Carter et al,63 2020 –0.40 (–0.98 to 0.18) CBT EDE-Q 16 Wagner et al,58 2016 –1.18 (–1.54 to –0.82) CBT EDE-Q 12 Ljotsson et al,61 2007 –1.13 (–1.64 to –0.62) DBT EDE-Q 13 Masson et al,72 2013 –0.70 (–1.22 to –0.18) CBT EDE 12 Sánchez-Ortiz et al,59 2011 –1.24 (–1.73 to –0.75) Heterogeneity: τ2 = 0.05, I2 = 46.37%, H2 = 1.86 –0.96 (–1.26 to –0.67) Unguided self-help (eating disorder symptom severity) DBT EDE-Q 24 Carter et al,63 2020 –0.42 (–0.99 to 0.16) CBT EDE-Q 16 Grilo et al,53 2013 –0.24 (–0.81 to 0.33) CBT EDE-Q 26 Grilo et al,54 2014 –0.11 (–0.65 to 0.44) Unguided self-help (depression) CBT BDI 8 Carter et al,63 2020 –0.53 (–1.06 to 0.01) CBT BDI 16 Grilo et al,53 2013 –0.41 (–0.98 to 0.16) CBT BDI 26 Grilo et al,54 2014 –0.16 (–0.70 to 0.39) CBT EDE-Q 12 Schmidt et al,68 2008 –0.08 (–0.48 to 0.31) Guided self-help (depression) CBT BDI-II 12 Cachelin et al,52 2019 –0.39 (–1.02 to 0.24) CBT BDI 16 Wagner et al,58 2016 –0.53 (–0.87 to –0.19) CBT MADRS 12 Ljotsson et al,61 2007 –0.93 (–1.43 to –0.43) CBT HADS-Dep 12 Sánchez-Ortiz et al,59 2011 –1.09 (–1.57 to –0.60) CFT EDE-Q 8 Kelly and Carter,60 2015 –0.23 (–0.87 to 0.41) Body Projecta EDE-Q 8 Green et al,55 2018 –0.15 (–0.58 to 0.29) Heterogeneity: τ2 = 0.00, I2 = 0.00%, H2 = 1.00 –0.18 (–0.38 to 0.03) Heterogeneity: τ2 = 0.04, I2 = 41.88%, H2 = 1.72 –0.73 (–1.04 to –0.43) Heterogeneity: τ2 = 0.00, I2 = 0.00%, H2 = 1.00 –0.37 (–0.68 to –0.05) BDI indicates Beck Depression Inventory; CBT, cognitive behavioral therapy; CFT, compassion-focused therapy-based self-help; DBT, dialectical behavioral therapy; EDE, Eating Disorder Examination; EDE-Q, EDE Questionnaire; HADS-Dep, Hospital Anxiety and Depression Scale–Depression; KQ, key question; MADRS, Montgomery-Åsberg Depression Rating Scale. a Cognitive-dissonance-based intervention (http://www.bodyprojectsupport.org/). Clinical Review & Education US Preventive Services Task Force USPSTF Review: Screening for Eating Disorders in Adolescents and Adults 1076 JAMA March 15, 2022 Volume 327, Number 11 (Reprinted) jama.com © 2022 American Medical Association. All rights reserved.  Table 3. Summary of Evidence for Screening in Eating Disorders in Adolescents and Adults Topic No. of studies (No. of participants) Summary of findings Consistency and precision Study quality Limitations (including reporting bias) Overall strength of evidence Applicability KQ1: Benefits of screening 0 No eligible studies NA NA NA Insufficient NA KQ2: Accuracy of screening tests for detecting eating disorders SCOFF (≥2): 10 (3684) Pooled sensitivity, 84% (95% CI, 74%-90%); specificity, 80% (95% CI, 65%-89%) Consistent and precise for sensitivity; inconsistent and imprecise for specificitya 7 Good; 3 Fair Potential bias related to participant selection Reference standards varied across studies Moderate for adequate accuracy Studies enrolled adults and either limited to women or enrolled a majority of women Several studies enrolled from specialty clinics or college campuses SCOFF (≥3): 7 (2749) Pooled sensitivity, 69% (95% CI, 56%-80%); specificity, 90% (95% CI, 69%-98%) Inconsistent and imprecise for both sensitivity and specificityb 4 Good; 3 Fair Potential bias related to participant selection Reference standards varied across studies Low for adequate accuracy All studies enrolled adults and either limited to women or enrolled a majority of women Several studies enrolled from specialty clinics or college campuses EDS-PC (≥2): 2 (627) Sensitivity, 97% (95% CI, 88%-100%); 100% (95% CI, 90%-100%) Specificity, 40% (95% CI, 35%-46%); 71% (95% CI, 64%-77%) Consistent and precise for sensitivity; inconsistent and imprecise for specificity Good Studies used different reference standards and enrolled diverse populations Insufficient One study recruited females and males from primary care and college campuses in the UK (77% females), and the other recruited female US veterans KQ3: Harms of screening 0 No eligible studies NA NA NA Insufficient NA KQ4: Benefits and harms of interventions for screen-detected or previously untreated ED Benefits of pharmacotherapy Lisdexamfetamine (BED): 4 (900) Pooled mean difference for reduction in YBOCS-BE scores larger in lisdexamfetamine group vs placebo (−5.75 [95% CI, −9.32 to −3.17]) Other outcomes assessed by 1 trial each (depression, anxiety, QOL, and function) YBOCS-BE: consistent, precise Other health outcomes: unknown consistency and imprecise Fair Outcomes assessed over relatively short duration (11-12 wk) Moderate for benefit in eating disorder symptom severity; insufficient for other health outcomes Studies enrolled adults with BED and obesity recruited via study advertisements Topiramate (BED): 2 (465) Larger reduction in YBOCS-BE in topiramate groups vs placebo; difference between groups in score change, −6.40 (P < .001) and −2.55 (P = .004) Other outcomes assessed by 1 trial each (depression, anxiety) YBOCS-BE; consistent, imprecisec Other outcomes: unknown consistency, imprecise Fair Outcomes assessed over a relatively short duration (14-16 wk) Low for benefit in eating disorder symptom severity; insufficient for other outcomes Studies enrolled adults with BED and obesity recruited via study advertisements SSRIs (BED) 5 (208) Two reported on eating disorder symptom severity: fluoxetine (EDE-Q) SMD, −0.69 (95% CI, −1.30 to −0.08) and escitalopram (YBOCS-BE) SMD, −0.29 (95% CI, −0.83 to −0.24) Larger reduction in depression symptoms among SSRI groups vs placebo (5 trials): pooled SMD, −0.61 (95% CI, −0.90 to −0.33) Eating disorder symptom severity: unknown consistency, imprecised Depression: consistent, imprecise Fair Studies assessed different SSRIs and reported outcomes over 6-16 wk Study eligibility criteria varied in terms of body weight and duration/frequency of binge eating episodes Insufficient for eating disorder symptom severity Low for benefit in depression symptom severity Studies enrolled adults with BED, most recruited via advertisements. Two limited to populations that were obese, and 1 limited to those with concurrent depression (continued) USPSTF Review: Screening for Eating Disorders in Adolescents and Adults US Preventive Services Task Force Clinical Review & Education jama.com (Reprinted) JAMA March 15, 2022 Volume 327, Number 11 1077 © 2022 American Medical Association. All rights reserved.  Table 3. Summary of Evidence for Screening in Eating Disorders in Adolescents and Adults (continued) Topic No. of studies (No. of participants) Summary of findings Consistency and precision Study quality Limitations (including reporting bias) Overall strength of evidence Applicability Fluoxetine (bulimia nervosa): 3 (528) Two found larger reduction in EAT scores among fluoxetine group vs placebo; difference was statistically significant in 1 trial Two found larger reductions in HAM-D scores among fluoxetine vs placebo; difference was statistically significant in 1 triale Eating disorder symptom severity: consistent; imprecise Depression symptom severity: consistent; imprecise Fair Studies reported outcomes at different durations (8 and 16 wk) Low for benefit (eating disorder and depression symptom severity) All enrolled populations with bulimia nervosa recruited via advertisements; 1 limited to those with bulimia nervosa and recurrent binge eating Benefits of therapy interventions Guided self-help: 7 (431) Guided self-help reduced eating disorder symptom severity more than control (5 studies; n = 391): pooled SMD, −0.96 (95% CI, −1.26 to −0.67) Guided self-help reduced depression symptoms more than control (4 studies; n = 324): pooled SMD, −0.73 (95% CI, −1.04 to −0.43) Eating disorder symptom severity: consistent, precise Depression symptom severity: consistent, precise Fair Frequency and mode of delivering guidance varied (eg, emails, individual sessions); studies assessed eating disorder and depression symptoms using different measures over a relatively short duration (8-16 wk) Moderate for benefit (eating disorder and depression symptom severity) All enrolled adults with BED recruited primarily via advertisements; several limited to populations that were obese Unguided self-help: 7 (421) Pooled results (6 studies; n = 368) favored self-help for reduction in eating disorder symptom severity but difference was not statistic...