Introduction

The goal of this guideline is to improve the quality of care and treatment outcomes for patients with eating disorders, as defined by the Diagnostic and Statistical Manual of Mental Disorders, 5th Edition (DSM-5; American Psychiatric Association 2013). As described in “Scope of Document,” we focus primarily on anorexia nervosa (AN), bulimia nervosa (BN), and binge-eating disorder (BED) rather than other DSM-defined feeding and eating disorders. Since publication of the last American Psychiatric Association (APA) practice guideline (American Psychiatric Association 2006) and guideline watch on eating disorders (Yager et al. 2012), there have been many studies on psychotherapies for individuals with these diagnoses as well as some studies on pharmacotherapies. Despite this, there are still substantial gaps in the availability and use of evidence-based treatments for individuals with an eating disorder (Kazdin et al. 2017). This practice guideline aims to help clinicians improve care for their patients by reviewing current evidence and providing evidence-based statements that are intended to enhance knowledge, increase assessment, and optimize treatment of eating disorders.

The lifetime prevalence of eating disorders in the United States is approximately 0.80% for AN, 0.28% for BN, and 0.85% for BED (Udo and Grilo 2018), although estimates can vary depending on the study location, sample demographic characteristics, case finding, and diagnostic approaches (Galmiche et al. 2019; Santomauro et al. 2021; Wu et al. 2020). For example, the prevalence of an eating disorder appears to be higher in LGBTQ+ individuals than in cisgender heterosexual peers (Kamody et al. 2020; Nagata et al. 2020b). Furthermore, data suggest an increasing incidence of eating disorders and inpatient care for eating disorders, particularly AN, during the COVID-19 pandemic (Agostino et al. 2021; Asch et al. 2021; Otto et al. 2021; Taquet et al. 2021). Importantly, the lifetime burdens and psychosocial impairments associated with an eating disorder can be substantial because these illnesses can persist for decades, and they typically have an onset in adolescence or early adulthood (Udo and Grilo 2018).

In the United States, for the 2018–2019 fiscal year, the total economic costs of eating disorders were estimated to be $64.7 billion, with an additional $326.5 billion attributable to reductions in well-being associated with eating disorders (Streatfeild et al. 2021). Of total economic costs, AN represented 17% of costs, BN 18% of costs, and BED 30% of costs (Streatfeild et al. 2021). Evidence from other countries is consistent with the United States’ findings and emphasizes the high economic burdens associated with eating disorders (Jenkins 2022; Tannous et al. 2021; van Hoeken and Hoek 2020).

Eating disorders are associated with increases in all-cause mortality and in deaths due to suicide (Auger et al. 2021; Nielsen and Vilmar 2021; Tith et al. 2020; van Hoeken and Hoek 2020). With AN, the increases in risks of mortality and premature death are substantially greater in men than in women (Edakubo and Fushimi 2020; Fichter et al. 2021; Iwajomo et al. 2021; Quadflieg et al. 2019), although the absolute numbers of deaths associated with an eating disorder are greater in women. Rates of suicide attempts are also increased in individuals who have an eating disorder (Keski-Rahkonen 2021; Smith et al. 2018; Udo et al. 2019). Morbidity and mortality among individuals with an eating disorder are heightened by the common co-occurrence of health conditions, such as diabetes, and of other psychiatric disorders, particularly depression, anxiety, posttraumatic stress disorder (PTSD), obsessive-compulsive disorder (OCD), attention-deficit/hyperactivity disorder (ADHD), and substance use disorders (Ahn et al. 2019; Cliffe et al. 2020; Gibbings et al. 2021; Keski-Rahkonen 2021; Udo and Grilo 2019).

Accordingly, the overall goal of this guideline is to enhance the assessment and treatment of eating disorders, thereby reducing the mortality, morbidity, and significant psychosocial and health consequences of these important psychiatric conditions.

This practice guideline focuses on evidence-based pharmacological, psychotherapeutic, and other nonpharmacological treatments for eating disorders in adolescents, emerging adults, and adults. In addition, it includes statements related to assessment and treatment planning, which are an integral part of patient-centered care.

The scope of this document is shaped by the diagnostic criteria for eating disorders and by the available evidence as obtained by a systematic review of the literature through September 2021. In particular, it focuses on AN, BN, and BED as defined by DSM-III, DSM-III-R, DSM-IV, DSM-IV-TR, DSM-5, or ICD-10. Some of the studies included individuals whose symptoms were below the threshold for a diagnosis of AN, BN, or BED, but these data were rarely analyzed separately in a way that would permit unique recommendations to be crafted for this group of patients. Nevertheless, some guideline statements may also be relevant to individuals with unspecified or other specified feeding or eating disorders.

Our systematic review attempted to include literature on avoidant/restrictive food intake disorder (ARFID); however, rigorous clinical trial data were not available due to the relative recency of the introduction of this diagnosis. Consequently, none of the guideline statements are related to the treatment of ARFID. However, we have included some discussion of ARFID in the implementation sections of this document, particularly as it relates to assessment and treatment planning.

We specifically excluded rumination disorder and pica from our search of the literature due to their typical age of onset in infancy or childhood and the limited evidence on their treatment. We also excluded treatment of obesity from the scope of this guideline because obesity is not categorized as an eating disorder. Although obesity is common among individuals who are treated in psychiatric practice, literature on obesity is already summarized by practice guidelines from other organizations and professional societies.

Most studies reported including a preponderance of women, typically adolescents or young adults, but participants’ genders were not described more fully. Most studies also enrolled predominantly white participants or did not specify the racial, ethnic, or cultural characteristics of the sample. These limitations of the evidence should be considered in terms of the document scope and the compelling need for additional research in more representative samples. In addition, as evidence accrues and as social norms change, terminology will likely evolve as well (Flanagin et al. 2021; OHSU Center for Diversity and Inclusion 2021).

Data are also limited on individuals with eating disorders and significant physical health conditions or co-occurring psychiatric conditions, including substance use disorders. Many of the available studies of eating disorders did not analyze data separately for these patient subgroups or excluded individuals with these comorbidities. Nevertheless, in the absence of more robust evidence, the statements in this guideline should generally be applicable to individuals with co-occurring conditions.

Our systematic review did not include studies for preventive interventions (Harrer et al. 2020; Stice et al. 2021; Watson et al. 2016) or risk factors for eating disorders, such as frequent dieting behaviors, childhood abuse, or bullying (Emery et al. 2021; Hooper et al. 2021; Lie et al. 2019; Solmi et al. 2021; Yoon et al. 2020). It also did not include search terms to identify literature on stigma and discrimination, either as risk factors for eating disorders, contributors to symptoms, or barriers to seeking treatment (Ali et al. 2017; Brelet et al. 2021; Bristow et al. 2020; Foran et al. 2020; Hamilton et al. 2022; O’Connor et al. 2021). Each of these topics is important but would warrant a distinct systematic review from one focused on treatments for eating disorders. Cost-effectiveness considerations are also outside of the scope of this guideline. Although treatment-related costs are often barriers to receiving treatment, costs of treatment typically differ by country and geographical region and vary widely by health system and payment model. In addition, few high-quality studies exist on the cost-effectiveness of treatments for eating disorders that could be used to inform health care policy.

Although we discuss studies of specific psychotherapies that were delivered via a Web-based approach, we do not discuss telehealth as a specific intervention because there were no direct comparisons of telehealth and in-person care prior to 2020. The rapidly expanding literature on the use of telehealth (Anderson et al. 2017; Blalock et al. 2020; Levinson et al. 2021; Matheson et al. 2020; Raykos et al. 2021; Stewart et al. 2021; Waller et al. 2020), Web-based interventions (Barakat et al. 2019), and mobile applications (Anastasiadou et al. 2018; Linardon et al. 2020; Wasil et al. 2021) in the treatment of eating disorders will help to inform future practice guidelines.

Since the publication of Clinical Practice Guidelines We Can Trust (Institute of Medicine 2011a), a report of the Institute of Medicine (now known as National Academy of Medicine), there has been an increasing focus on using clearly defined, transparent processes for rating the quality of evidence and the strength of the overall body of evidence in systematic reviews of the scientific literature. This guideline was developed using a process intended to be consistent with the recommendations of the Institute of Medicine (2011a) and Principles for the Development of Specialty Society Clinical Guidelines of the Council of Medical Specialty Societies (2012). The parameters used for the guideline’s systematic review are included with the full text of the guideline; the development process is fully described in the following document, which is available at the APA website: www.psychiatry.org/psychiatrists/practice/clinical-practice-guidelines/guideline-development-process.

Development of guideline statements entails weighing the potential benefits and harms of the statement and then identifying the level of confidence in that determination. (See Appendix G for detailed descriptions of the potential benefits and harms for each statement.) This concept of balancing benefits and harms to determine guideline recommendations and strength of recommendations is a hallmark of GRADE (Grading of Recommendations Assessment, Development and Evaluation), which is used by multiple professional organizations around the world to develop practice guideline recommendations (Guyatt et al. 2013). With the GRADE approach, recommendations are rated by assessing the confidence that the benefits of the statement outweigh the harms and burdens of the statement, determining the confidence in estimates of effect as reflected by the quality of evidence, estimating patient values and preferences (including whether they are similar across the patient population), and identifying whether resource expenditures are worth the expected net benefit of following the recommendation (Andrews et al. 2013).

In weighing the balance of benefits and harms for each statement in this guideline, our level of confidence is informed by available evidence (see Appendix C), which includes evidence from clinical trials as well as expert opinion and patient values and preferences. Evidence for the benefit of a particular intervention within a specific clinical context is identified through systematic review and is then balanced against the evidence for harms. In this regard, harms are broadly defined and may include serious adverse events, less serious adverse events that affect tolerability, minor adverse events, negative effects of the intervention on quality of life, barriers and inconveniences associated with treatment, direct and indirect costs of the intervention (including opportunity costs), and other negative aspects of the treatment that may influence decision-making by the patient, the clinician, or both.

Many topics covered in this guideline have relied on forms of evidence such as consensus opinions of experienced clinicians or indirect findings from observational studies rather than research from randomized trials. It is well recognized that there are guideline topics and clinical circumstances for which high-quality evidence from clinical trials is not possible or is unethical to obtain (Council of Medical Specialty Societies 2012). For example, many questions need to be asked as part of an assessment, and inquiring about a particular symptom or element of the history cannot be separated out for study as a discrete intervention. It would also be impossible to separate changes in outcomes due to assessment from changes in outcomes due to ensuing treatment. Research on psychiatric assessments and some psychiatric interventions can also be complicated by multiple confounding factors such as the interaction between the clinician and the patient or the patient’s unique circumstances and experiences. The GRADE working group and guidelines developed by other professional organizations have noted that a strong recommendation or “good practice statement” may be appropriate even in the absence of research evidence when sensible alternatives do not exist (Andrews et al. 2013; Brito et al. 2013; Djulbegovic et al. 2009; Hazlehurst et al. 2013). For each guideline statement, we have described the type and strength of the available evidence as well as the factors, including patient preferences, that were used in determining the balance of benefits and harms.

The authors of the guideline determined each final rating, as described in the section “Guideline Development Process” (see Table 1). A recommendation (denoted by the numeral 1 after the guideline statement) indicates confidence that the benefits of the intervention clearly outweigh harms. A suggestion (denoted by the numeral 2 after the guideline statement) indicates greater uncertainty. Although the benefits of the statement are still viewed as outweighing the harms, the balance of benefits and harms is more difficult to judge, or either the benefits or the harms may be less clear. With a suggestion, patient values and preferences may be more variable, and this can influence the clinical decision that is ultimately made. Each guideline statement also has an associated rating for the strength of supporting research evidence. Three ratings are used: high, moderate, and low (denoted by the letters A, B, and C, respectively) and reflect the level of confidence that the evidence for a guideline statement reflects a true effect based on consistency of findings across studies, directness of the effect on a specific health outcome, precision of the estimate of effect, and risk of bias in available studies (Agency for Healthcare Research and Quality 2014; Balshem et al. 2011; Guyatt et al. 2006).

The APA Practice Guidelines are assessments of current (as of the date of authorship) scientific and clinical information provided as an educational service. The guidelines 1) should not be considered as a statement of the standard of care or inclusive of all proper treatments or methods of care; 2) are not continually updated and may not reflect the most recent evidence, as new evidence may emerge between the time information is developed and when the guidelines are published or read; 3) address only the question(s) or issue(s) specifically identified; 4) do not mandate any particular course of medical care; 5) are not intended to substitute for the independent professional judgment of the treating clinician; and 6) do not account for individual variation among patients. As such, it is not possible to draw conclusions about the effects of omitting a particular recommendation, either in general or for a specific patient. Furthermore, adherence to these guidelines will not ensure a successful outcome for every individual, nor should these guidelines be interpreted as including all proper methods of evaluation and care or excluding other acceptable methods of evaluation and care aimed at the same results. The ultimate recommendation regarding a particular assessment, clinical procedure, or treatment plan must be made by the clinician directly involved in the patient’s care in light of the psychiatric evaluation, other clinical data, and the diagnostic and treatment options available. Such recommendations should be made in collaboration with the patient whenever possible and should incorporate the patient’s personal and sociocultural preferences and values, which can enhance the therapeutic alliance, adherence to treatment, and treatment outcomes. For all of these reasons, the APA cautions against the use of guidelines in litigation. Use of these guidelines is voluntary. APA provides the guidelines on an “as is” basis and makes no warranty, expressed or implied, regarding them. APA assumes no responsibility for any injury or damage to persons or property arising out of or related to any use of the guidelines or for any errors or omissions.

For ease of clinical use, only the guideline statements and the discussion of their implementation are included in the print version. Discussion of the research evidence and a detailed description of the guideline development process are available online at https://psychiatryonline.org/doi/book/10.1176/appi.books.9780890424865.