New US Preventive Services Task Force Recommendations on Screening for Eating Disorders

Eating disorders are conditions characterized by disturbances in eating behaviors that impair physical and psychosocial functioning and include, but are not limited to, anorexia nervosa, binge eating disorder, bulimia nervosa, avoidant/restrictive food intake disorder, and other specific feeding and eating disorders, including atypical anorexia nervosa (Table).¹ Lifetime prevalence estimates for eating disorders range from 0.5% to 3.5% in women and 0.1% to 2.0% in men,² although these may be underestimations, particularly because the prevalence of eating disorders has risen during the COVID-19 pandemic. Eating disorders have considerable short- and long-term consequences for mental and physical health.

Table.  Key Features of Different Eating Disorders^a

View LargeDownload

For the first time, the US Preventive Services Task Force (USPSTF) reviewed eating disorder screening in asymptomatic adolescents and adults and gave it an I statement for insufficient evidence,³ which was supported by an evidence report and systematic review.⁴ The USPSTF only reviewed evidence for adolescents and adults with no signs or symptoms of eating disorders and with a normal or high body mass index (BMI). It is important to note that the insufficient evidence statement is not a recommendation for or against screening, but rather that there is not enough evidence to make a recommendation either way.

Screening for eating disorders can be accomplished through questionnaires that ask about eating habits, feelings about eating, and perception of weight. The screening tool that had the most studies included in the systematic review was SCOFF.⁵ Among 10 trials in adults using a cut point of 2 or greater, SCOFF had a pooled sensitivity of 84% and specificity of 80%.⁴ Another screening tool included in the review was the Eating Disorder Screen for Primary Care (Box),⁶ which had a sensitivity of 97% to 100% and specificity of 40% to 71% in 2 studies of adults using a cut point of 2 or greater. Although the USPSTF determined that there was adequate evidence supporting the accuracy of SCOFF for screening for eating disorders in adult women, there was not adequate evidence for adolescents, men, and other populations. SCOFF was designed to detect anorexia nervosa and bulimia nervosa,⁵ but with the introduction of the Diagnostic and Statistical Manual of Mental Disorders (Fifth Edition)¹ there are other eating disorders such as avoidant/restrictive food intake disorder and atypical anorexia nervosa with a greater diversity in clinical presentations.

The USPSTF found no trials that directly evaluated the benefits or harms in participants who were screen detected from primary care, which ultimately led to the insufficient evidence recommendation.^3,4 While some trials assessed interventions for adults with recently detected or previously untreated eating disorders, the USPSTF only considered studies that link screening for eating disorders in primary care to health benefits or harms. One barrier limiting the link from primary care to treatment outcomes is that eating disorder care typically involves an interdisciplinary approach, including behavioral, medical, and nutritional specialists.⁷ Eating disorder services are often limited to specialty or academic centers and may have long wait lists. Therefore, easy and quick access to specialists from primary care or development of primary care interventions for eating disorders are needed to demonstrate a benefit for those who have positive screens for eating disorders in primary care. Without access to care, identifying eating disorders by screening could create anxiety while awaiting specialty care or the stigma associated with a presumed eating disorder diagnosis, which could be potential harms.

The USPSTF recommendation focuses on eating disorder screening in asymptomatic adolescents and adults, for which insufficient evidence was found.³ However, clinicians could consider assessing for eating disorders based on signs or symptoms. As recommended by several professional societies,^7-9 health care professionals should be able to recognize and identify signs of eating disorders, such as preoccupation with weight, food, eating, appearance, calories, and engaging in weight-control behaviors such as fasting, skipping meals, and vomiting to lose weight. In female patients, the American College of Obstetricians and Gynecologists notes that the presence of irregular menstrual cycles or amenorrhea may be presenting signs or symptoms of an eating disorder.⁸ Other risk factors for eating disorders include childhood adversity, trauma, perfectionism, social pressure related to appearance, the presence of other mental health conditions, and genetics.² Assessing weight, height, and BMI is already the standard of care in primary care settings.⁹ These anthropometrics can provide clinicians with key information about weight loss or weight change that might raise the index of suspicion for an eating disorder.⁹ However, there are limitations to the use of BMI, particularly with the inability to differentiate body composition (eg, fat mass vs lean mass).

Although there are common stereotypes about people with eating disorders, eating disorders can affect people of any BMI, gender, sexual orientation, race, ethnicity, or socioeconomic class. Part of the challenge in studying this topic is that disordered eating is a heterogeneous set of disorders for which no single screening tool may be effective and no single set of interventions may prove beneficial. For instance, people of any BMI can be affected by eating disorders, so BMI alone may not suggest an eating disorder. Atypical anorexia nervosa describes individuals who have lost a considerable amount of weight but whose weight may be in the normal or above normal range. Such individuals are at high risk for both medical and psychological complications.¹⁰ Engagement in disordered eating behaviors such as vomiting, skipping meals for weight loss, and using nonprescribed medicines for weight loss (eg, laxatives, diuretics, diet pills) are more common in adults with higher BMI compared with those with lower BMI.¹¹ Therefore, as part of obesity or weight management, clinicians should discourage disordered eating behaviors for weight loss.¹² Binge eating disorder is associated with higher BMI and metabolic disorders.¹³

Further challenging the study of this topic is the reality that the manifestation of specific eating disorders can present differently among different demographics. The USPSTF particularly noted a need for trials and studies addressing screening in men, adolescents, and populations of different sexual orientations, genders, races, and ethnicities. For instance, men may present with muscularity-oriented disordered eating (eg, protein overconsumption while restricting carbohydrates, use of steroids or other muscle-building supplements),¹⁴ which may not be detected by SCOFF or other traditional eating disorder screening measures. Transgender individuals may have concerns related to attainment of the body image norms for their preferred gender identity.¹⁵ These populations in particular are understudied and underdiagnosed, so validated screening tools applicable to these diverse populations are needed.

In summary, the USPSTF found insufficient evidence for screening for eating disorders in asymptomatic adolescents and adults.³ Research on benefits and harms of screen-detected populations from primary care settings is needed because no studies assessing this question were identified.⁴ In particular, studies of screening and treatment in adolescents, men, racial and ethnic minority groups, and sexual and gender minority groups are needed to move beyond the insufficient evidence recommendation.

Corresponding Author: Jason M. Nagata, MD, MSc, Division of Adolescent and Young Adult Medicine, Department of Pediatrics, University of California, San Francisco, 550 16th St, 4th Floor, Box 0110, San Francisco, CA 94158 (jason.nagata@ucsf.edu).

Published Online: March 15, 2022. doi:10.1001/jamainternmed.2022.0121

Conflict of Interest Disclosures: Dr Nagata reported grants from the National Institutes of Health and the American Heart Association during the conduct of the study. No other disclosures were reported.

Additional Contributions: We thank Khushi P. Patel of the University of California, San Francisco for editorial assistance and help with the literature review. No compensation was received for these contributions.