2026 eating disorder statistics

The U.S. figures on this page come from household surveys summarized by the National Institute of Mental Health and from a June 2020 cost study by Deloitte Access Economics for the Strategic Training Initiative for the Prevention of Eating Disorders (STRIPED) at the Harvard T.H. Chan School of Public Health and the Academy for Eating Disorders. The surveys interviewed people from 2001 to 2004. The cost study covers 1 October 2018 through 30 September 2019. A clinician makes a diagnosis. These figures describe groups of people in published research.

This page puts the year next to each figure and says what a shortened version leaves out. For an overview of the illnesses, see what is an eating disorder.

Those three figures are modeled estimates from the Deloitte report. The report applied prevalence to a U.S. population it cites as 330.3 million.¹

One-year prevalence

In 2018-19, an estimated 1.66% of people in the United States had an eating disorder, about 5.48 million people. The rate was 2.62% among females (4.39 million) and 0.67% among males (1.09 million).¹

By diagnosis, the one-year rates for all ages were 0.12% for anorexia nervosa, 0.19% for bulimia nervosa, 0.62% for binge-eating disorder, and 0.73% for other specified feeding or eating disorder (OSFED). OSFED was the most common of the four. Prevalence was highest in the 20 to 29 age group for both males and females.¹

The report left out avoidant/restrictive food intake disorder (ARFID), pica, and rumination disorder. It did not have economic data it judged sufficient for those conditions.¹

Lifetime prevalence

The same report gives three lifetime-style counts, and they answer different questions.

Lifetime prevalence among people alive in 2018-19 was estimated at 8.60% for females and 4.07% for males. About 21.0 million people had already had an eating disorder (14.4 million women and 6.6 million men).¹

A further projection says 28.8 million Americans who were alive in 2018-19 will have an eating disorder at some point in life, counting past, current, and future cases. Of that total, 7.8 million were expected to develop an eating disorder later. About 1.9 million of those future cases were expected before age 20.¹

The 28.8 million figure is the one most often shortened to a single percentage. It counts past, current, and future cases. The number of people who had an eating disorder in 2018-19 is 5.48 million.¹

The most common diagnoses

In the 2018-19 model, OSFED accounted for the largest one-year share (0.73% of people), then binge-eating disorder (0.62%), bulimia nervosa (0.19%), and anorexia nervosa (0.12%).¹

Adult interview data measure something else. They leave out OSFED, and they measure lifetime risk rather than one year. Among U.S. adults in the National Comorbidity Survey Replication (NCS-R), lifetime prevalence was 2.8% for binge-eating disorder, 1.0% for bulimia nervosa, and 0.6% for anorexia nervosa. Binge-eating disorder was the most common of those three.²³

Past-year adult rates from the same survey were 1.2% for binge-eating disorder (1.6% of females, 0.8% of males) and 0.3% for bulimia nervosa (0.5% of females, 0.1% of males). The survey did not produce a past-year estimate for anorexia nervosa. Lifetime anorexia was 0.9% among females and 0.3% among males.²

Among adolescents ages 13 to 18 in the National Comorbidity Survey Adolescent Supplement (NCS-A), lifetime prevalence of anorexia, bulimia, or binge-eating disorder together was 2.7% (3.8% of females, 1.5% of males). The rate was 2.4% at ages 13 to 14, 2.8% at ages 15 to 16, and 3.0% at ages 17 to 18.²⁴

Both surveys used DSM-IV criteria. ARFID was not a separate diagnosis in that manual. The surveys published sex as female and male, and the figures above follow that.

For the illnesses themselves, see binge eating disorder, bulimia nervosa, anorexia nervosa, OSFED, and ARFID.

Age of onset

In the NCS-R adult sample, the median age of onset was 18 years for anorexia nervosa, 18 years for bulimia nervosa, and 21 years for binge-eating disorder.²³

That median is the middle of the adult sample, not the only age an eating disorder can start. The adolescent survey found cases from ages 13 to 18, and the cost report estimated eating disorders in children and in adults over 40.²¹ See eating disorders in teens.

Other mental illnesses

In the NCS-R, 56.2% of adults with anorexia nervosa, 94.5% of adults with bulimia nervosa, and 78.9% of adults with binge-eating disorder also met criteria for at least one other core DSM-IV disorder assessed in the survey. For all three eating disorders, the highest overlap was with an anxiety disorder: 47.9% in anorexia, 80.6% in bulimia, and 65.1% in binge-eating disorder.²

Mood disorders, impulse-control disorders, and substance use disorders were also elevated. Another mental illness can occur alongside an eating disorder. It does not replace the eating disorder, and it does not make the eating disorder milder.²

Past-year impairment, measured with the Sheehan Disability Scale, was common. Among adults with binge-eating disorder, 62.6% had any impairment and 18.5% had severe impairment. Among adults with bulimia nervosa, 78.0% had any impairment and 43.9% had severe impairment.²

Treatment seeking

In the NCS-R lifetime data, 33.8% of adults with anorexia nervosa, 43.2% of adults with bulimia nervosa, and 43.6% of adults with binge-eating disorder had sought treatment specifically for the eating disorder.²

The pattern differed by sex and by diagnosis. Females sought eating-disorder treatment more often than males for bulimia (47.0% versus 29.1%) and for binge-eating disorder (50.8% versus 28.9%). Males with anorexia nervosa sought treatment more often than females (50.2% versus 29.8%).²

A larger share, 50.0% to 63.2% depending on the diagnosis, had received treatment for emotional problems at some point. Treatment for depression or anxiety is separate from treatment for the eating disorder.²

These percentages come from household interviews in 2001 to 2003. They describe treatment seeking in that survey. The eating-disorder questions were asked of a subsample of 2,980 people inside a survey of 9,282 adults.²

Mortality

A 2011 meta-analysis of 36 studies, cited in the cost report, found standardized mortality ratios of 5.86 for anorexia nervosa, 1.93 for bulimia nervosa, and 1.92 for eating disorder not otherwise specified (the DSM-IV category that preceded OSFED). A ratio of 5.86 means the death rate in the anorexia studies was 5.86 times the rate in the general population of similar age and sex.¹⁵

The cost report applied those ratios and estimated about 10,200 deaths associated with eating disorders in the United States in 2018-19. The range around that estimate is 5,500 to 22,000. More of the modeled deaths were tied to OSFED (about 3,400, or 33%) than to any other single diagnosis, because OSFED was more common in the model, not because OSFED has the highest death rate per person.¹

The report notes that recorded deaths can miss an eating disorder. A death certificate may list heart failure or another immediate cause when an eating disorder was the underlying illness. The mortality ratios also do not all adjust for every other illness a person may have.¹

Dividing 10,200 by the minutes in a year (525,600) comes to about one death every 52 minutes. That line restates the point estimate. The report does not publish it. The range of 5,500 to 22,000 is wide enough that a single "every 52 minutes" line is more exact than the estimate.

Cost

Financial costs in 2018-19 were estimated at $64.7 billion, or $11,808 per person with an eating disorder that year. Health system spending was $4.6 billion, 7% of the financial total. Of that, $363.5 million was paid out of pocket. The rest of the financial total is mostly lost productivity, unpaid care from family, and efficiency losses.¹

The report also valued lost wellbeing at $326.5 billion, from 1.3 million disability-adjusted life years multiplied by a value of a statistical life year. That wellbeing figure is a valuation of lost health, not money spent. Adding it to the $64.7 billion would mix those two estimates.¹

Statistics to use with care

Citing this page

Where these figures come from

Two public sources carry the numbers on this page.

The diagnosis-specific adult and adolescent rates are the figures published by the National Institute of Mental Health. Adult rates come from the National Comorbidity Survey Replication, a household survey interviewed between February 2001 and April 2003. Eating disorders were assessed in a subsample of 2,980 of 9,282 respondents, with DSM-IV criteria.²³ Adolescent rates come from the National Comorbidity Survey Adolescent Supplement, ages 13 to 18, interviewed between February 2001 and January 2004 (10,123 adolescents).²⁴

The one-year counts, the 28.8 million lifetime projection, the death estimate, and the cost figures come from Deloitte Access Economics, Social and economic cost of eating disorders in the United States of America (June 2020), prepared for STRIPED and the Academy for Eating Disorders. The report combines survey estimates with modeling. It is a model, not a count of diagnosed cases in medical records. ARFID is outside its scope.¹

Standardized mortality ratios are the ratios the report cites from Arcelus and colleagues (2011), a meta-analysis of 36 studies.⁵

For levels of care, see levels of care. A private screen, not a diagnosis, is the self-assessment.