For Families > Disorders and Treatments > A-Z Disorder Guide > Eating Disorders

Seminal Issues in the Treatment of Eating Disorders

Katherine Halmi, M.D. spoke with us about the current state of research on eating disorders and their treatment. Dr. Halmi is a Professor of Psychiatry at Weill Medical College of Cornell University and the Director of the Eating Disorder Program at New York Presbyterian Hospital - Westchester Division.

AOK: What causes eating disorders?
Dr. Halmi: My work is based on a multidimensional model of causal and maintaining factors in anorexia nervosa (AN) and bulimia nervosa (BN). In this model, three factors contribute to the development of eating disorders after dieting begins: biological vulnerability, psychological predisposition, and social climate. Dieting leads to weight loss and malnutrition, causing psychological changes that reinforce dieting. At the same time, dieting can cause starvation effects and physiological changes, which also contribute to psychological changes that further reinforce dieting behavior.

Do certain personality characteristics make a person more likely to develop an eating disorder?
Perfectionism seems to be linked to eating disorders. In conjunction with several other sites nation-wide, I conducted a study examining this relationship. We looked at four different groups, consisting of 163 anorectic restrictors (people with anorexia who restrict their food intake), 120 anorectic purgers (people with anorexia who also engage in purging behavior), 62 anorectic binge purgers (people with anorexia who engage in both bingeing and purging behaviors), and 44 people with no eating disorder diagnosis. Participants in all groups completed a perfectionism scale, an eating disorders scale, and a scale measuring obsessions and compulsions. We found all three subgroups of individuals with anorexia to have higher rates of perfectionism than controls. There was also a significant, but less robust correlation between the trait of perfectionism and the presence of obsessions and compulsions.

We decided to further explore the relationship between obsessions and compulsions and eating disorders within this participant group. The participants completed checklists measuring the presence of eating disorders, obsessions, and compulsions. Their scores were then compared to those of an age-matched group of female patients with obsessive-compulsive disorder drawn from another study. The results indicated that 70 to 80% of patients diagnosed with AN had experienced some obsessions or compulsions in their lifetime. Over half of those diagnosed with AN had aggressive, symmetry, and contamination obsessions, though only for symmetry obsessions did AN patients score as highly as those with OCD.

What other factors put a person at risk to develop an eating disorder?
There are several important risk factors associated with eating disorders. Familial factors include having a family member with anorexia nervosa, bulimia nervosa, or obesity. To a lesser degree, having a family member with a history of alcohol or drug abuse or depression can also increase risk. Individual biological factors can also play a role. Being mildly overweight at puberty, or having an early onset of menarche increases one's chances of being diagnosed with an eating disorder. According to my research, psychological factors, such as perfectionism, can increase risk. Other psychological risk factors include low self-esteem, a sense of ineffectiveness, affective disorders (depression, for example), and drug and alcohol abuse. In addition, risk is increased by individual behaviors, such as dieting or involvement in activities or professions that demand and emphasize weight control for performance, such as gymnastics. Finally, the culture one lives in affects one's chance for developing an eating disorder. Living in an industrialized country, living in a society which emphasizes thinness as beauty, and living in a country where there has been a weight increase in the population all contribute to risk.

Has the profile of a typical eating disorder patient changed over time?
According to statistics gathered from our Eating Disorder Inpatient Unit at New York Presbyterian Hospital - Westchester Division, the years from the eighties to the nineties saw a marked increase in admissions of patients over 30, as well as an increase in admissions of patients younger than thirteen. Non-Caucasians, who constituted 4.5% of admissions in 1984, made up 11% of admissions in 1998. There has also been a significant increase in male admissions.

How does a male patient with an eating disorder differ from a female patient?
Our statistics actually revealed few differences between males and females with eating disorders. Males did have a later age onset of eating disorders by an average of two years, but this may be explained by males' later age of onset for puberty. Other differences we found were that males were more likely to be involved in a sport or occupation requiring weight control, were less likely to use diet pills and laxatives as weight control methods, and were more likely to have parents who were divorced. Males were also more likely to be readmitted than females.

Are there specific warning signs that a person is suffering from an eating disorder?
If you're concerned that someone might have an eating disorder, you should be alert to the following signs:

  • an increase or decrease in weight not due to a medical condition
  • extreme eating behaviors such as severe dieting, secretive bingeing, peculiar or ritualized eating behavior at mealtime, or refusal to eat with family or friends
  • progressive isolation from friends and social activities
  • intense preoccupation with weight and body image
  • compulsive or excessive exercising
  • self-induced vomiting
  • use of laxatives, diuretics, or diet pills
  • mood changes of depression, irritability, or an inability to concentrate

What factors affect the course and severity of the disorder?
For anorexia nervosa, poorer outcomes are found in cases where the age of onset is either under twelve or over eighteen. The presence of bulimic or purging symptoms, lower weight at treatment, repeated hospitalizations, poor social functioning, and a more chronic pattern of illness are all indicators of a worse prognosis.

How should a professional best assess a person with a possible diagnosis of anorexia nervosa?
The assessment interview process should include discussions with the patient's family and friends. This provides a more accurate view of the situation, as most people suffering from anorexia nervosa resist being diagnosed and may withhold or provide misleading information. During these discussions, the interviewer should establish any family history of eating disorders, affective disorders, alcohol, or drug abuse. The interviewer should also try to assess the parents' response to the patient's dieting behaviors and weight loss as well as the parents' relationship with each other. Familial relationships are important to the treatment process - controlled studies show that children with eating disorders who receive family counseling in addition to individual counseling do significantly better than those who receive individual counseling alone. The interviewer should ask about any previous treatment for eating disorders and gauge the patient's response to those treatments.

The patient should undergo a full mental status evaluation. This evaluation should include an assessment of the patient's attitude about food, eating, and weight. The interviewer should ask the patient their ideal weight as well as their culinary activities (e.g. any interests in cooking, grocery shopping, collecting recipes, etc.). The patient's ability to concentrate, her attitudes about sex and peer relationships, her attitudes about family members, and the presence of depression or sleep disturbance should also be assessed at this time.

Finally, the patient should have a complete medical assessment, including a cardiac examination and measurement of nutrient levels in order to fully assess the physiological consequences of the disorder.

What is the current state of treatment?
In order to understand treatment outcomes, it's helpful to compare some statistics we gathered at the New York Presbyterian Hospital - Westchester Division's Eating Disorder Inpatient Unit in 1984 with those we gathered in 1998. During this time, there was a great increase in first-time admissions to the inpatient unit, coupled by a sharp decrease in the average length of stay. In 1984, patients stayed in the unit for an average of 3 ½ months, while in 1998, the average stay was only 22 days. The length of stay decreased most significantly for patients with anorexia nervosa. This may be explained by the sharp rise in the number of patients with managed care insurance -- from 0% in 1984 to 58.4% in 1998.

There was also a dramatic increase in patients readmitted, most of who had anorexia nervosa. I would argue that the large number of readmissions is reflective of the fact that patients are being discharged too soon. Our statistics show that from the 1980's to the 1990's, there was a significant drop in body mass index scores (a calculation in which a person's weight is corrected for their height) at the time of discharge. Research shows that if a patient is discharged with a body mass index below the normal range, the chance of relapse is significantly greater.

What are the chances of recovery?
One-quarter of patients with anorexia nervosa fully recover, another ¼ remain chronically ill, and ½ have partial improvement. Overall, there is a 40% rate of relapse. One-half of patients with bulimia nervosa fully recover. Twenty percent of these patients stay chronically ill and 1/3 relapse.

After 10 years of illness, the mortality rate for anorexia nervosa is 6.6%. At thirty years, the mortality rate rises to between 18 and 20%.

What makes the treatment of eating disorders so difficult?
A patient's eating disordered behavior can be reinforcing, thereby making treatment difficult. There are two primary reinforcing factors for anorexia nervosa. First, an eating disorder can serve as a coping mechanism for adverse experiences. Examples of such experiences include developmental transitions, fears of maturity and autonomy, distressing life events, feelings of ineffectiveness, helplessness, poor self-esteem, and disturbed relationships with family and friends. Second, a patient may use her eating behaviors in an attempt to increase her self-esteem and feeling of control. For example, people with eating disorders often believe that by remaining sick, they will be able to keep a troubled family intact. Family, friends, or other adults may also unwittingly reinforce the behaviors by commenting on weight or looks or requiring a particular weight for an activity or career.

What is the best method of treatment?
Fewer than 10 randomly assigned, controlled studies on the treatment of anorexia nervosa exist. Almost all of these have methodological problems, including small sample sizes, high dropout rates, and interfering treatment variables. The only existing controlled studies examine drug therapy and cognitive behavioral therapy (CBT).

How can CBT be applied to the treatment of anorexia nervosa?
The process of CBT for the treatment of anorexia nervosa rests on two core assumptions. First, patients with anorexia nervosa believe that food avoidance is necessary to maintain a low weight. Second, anorexia serves a positive function for the patient by providing an escape from unpleasant feelings and distressing life events. Based on these core assumptions, it is deduced that anorexia nervosa is reinforcing and the prospect of relinquishing safe and routinized AN behaviors is terrifying for the patient.

In CBT, the practitioner works towards two treatment goals, problem solving and cognitive restructuring, while closely monitoring food intake, as well as any dysfunctional thoughts, feelings, and emotions.

The problem-solving component uses the following steps. The patient:

  1. identifies specific problems
  2. generates as many solutions to these problems as possible
  3. considers the effectiveness and feasibility of each strategy
  4. selects a coping strategy
  5. defines the steps needed to carry out that strategy
  6. carries it out
  7. evaluates the process and results with the therapist

Cognitive restructuring involves a similar series of steps. The patient:

  1. writes down her thoughts and judgements about food and weight
  2. writes down arguments and evidence for the validity of these thoughts and judgements and lists support for the arguments
  3. writes down arguments and evidence doubting the validity of these thoughts and judgements and lists support for these counter-arguments
  4. generates a realistic and appropriate interpretation
  5. reaches a reasoned conclusion for use in governing weight and weight-related behaviors

Patients are told that they are not necessarily expected to believe these conclusions, but that they should view this process as a cognitive exercise.

How does CBT compare to drug therapy in terms of effectiveness?
We participated in a multi-center study that examined the relative effectiveness of CBT and the drug fluoxetine in the treatment of anorexia nervosa. Anorectic restrictors and anorectic binge-purgers were randomly assigned to one of three treatment groups. The first group received CBT and medical management. The second group received fluoxetine and medical management. The third group received both CBT and fluoxetine in addition to medical management. The group receiving only fluoxetine had the highest dropout rate (66%) of any group, leading us to conclude that most people afflicted with anorexia nervosa will not accept medication alone. When patients received fluoxetine in combination with therapy, they seemed more willing to accept the medication; the dropout rate in this group was only 53%. The only significant difference between the CBT group and the CBT with fluoxetine group was that medication seemed to add to the effectiveness of CBT in improving self-esteem.

What conclusions have you made about the effectiveness of drug therapy in the treatment of anorexia nervosa?
I would argue that medication is only a useful adjunct in the treatment of anorexia nervosa. Cyproheptadine, the safest of the drugs used to treat anorexia, has been shown to facilitate weight gain in anorectic restrictors, though not in anorectic binge-purgers. Cyproheptadine also improves symptoms of depression.
Only two controlled studies have been done on any of the major tranquilizer drugs and these yielded only minor effects. Tranquilizers may reduce severely obsessive, compulsive, and agitated behaviors, and has the often desirable side effect of weight gain. I recommend the use of these drugs primarily in cases where the patient is very agitated and is constantly exercising.

Studies like the one I just described suggest that fluoxetine and other selective serotonin reuptake inhibitors may reduce the relapse of eating disorder behaviors. These drugs are not effective in emaciated anorectics.

Are there any effective methods of eating disorder prevention?
Of the few well-controlled studies on prevention of eating disorders that exist, most have insignificant or contradictory findings. Without specific data, we are only able to pinpoint some general areas of concern on which we can focus our prevention efforts. These concerns include:

  1. the strong negative influence of peer group pressure
  2. a lack of developmental experiences to develop an individual's sense of confidence and effectiveness
  3. a general reduction of physical activity in children leading to a corresponding increase in weight
  4. the "cult of the super-performers", resulting in children who don't like to participate in sports or other athletic activities if their performance can not be perfect.

About the Authors

Katherine Halmi, M.D. is a Professor of Psychiatry at Weill Medical College of Cornell University and the Director of the Eating Disorder Program at New York Presbyterian Hospital - Westchester Division

Jennifer Rosenblatt is the Project Assistant for