Patients diagnosed with Binge Eating Disorder (BED) overeat food but, as opposed to patients with bulimia nervosa, they do not compensate for their increased food intake and therefore their body weight increases. It is hypothesized that the speed of eating has increased in BED and that body weight will decrease if the speed of eating is decreased. The speed of eating is measured using Mandometer, an eating disorder conditioning tool, class 2 device cleared by the FDA for the treatment of eating disorders. Mandometer is a scale connected to a computer, patients eat food from a plate on the scale and the computer stores the weight loss of the plate, thus recording the speed of eating. Patients decrease their speed of eating by following training curves on the computer screen while eating. The emergence of their own speed of eating on the screen makes this possible. Patients use Mandometer for lunch and dinner over one week at home to estimate their speed of eating and their food intake as the first step of clinical pratice. Mandometer is then programmed with how much and how quickly to eat and patients practice eating using Mandometer at home over the subsequent treatment. Data from 166 BED-patients using Mandometer at home and from a total of 354 patients who have been treated have been collected. 30 normal weight, healthy subjects will be recruited to test the hypothesis that their speed of eating is lower than that of the BED-patients. These control subjects will eat lunch and dinner using Mandometer at home over one week, but they will, obviously, not participate in treatment.

Background: control of the speed of eating and the amount of food eaten is a major intervention in Mandometer treatment of anorexia nervosa (AN), bulimia nervosa (BN), and eating disorders not otherwise specified (EDNOS), including binge eating disorder (BED), which is an EDNOS. The control is achieved by Mandometer, which is a scale connected to a computer. The patient eats food from a plate on the scale and the computer records the weight loss of the plate over the course of the meal. Thus, the amount of food eaten and the duration of the meal are measured. This yields a measure of the speed of eating. Patients also estimate their feeling of fullness at regular intervals on a scale on the Mandometer computer screen and the computer stores the ratings. Thus, the development of satiety is also measured. AN-patients eat only little food over a long period of time. They practice eating more food at a higher speed by following training curves on the computer screen while eating. This is possible because the patients can see their own speed of eating emerging on the screen while eating and so they can adjust their speed of eating to the training curves. The patients learn to feel full in a normal manner by practicing in the same way, assisted by feedback from the Mandometer computer screen. Mandometer treatment of AN and BN has been evaluted in a randomized controlled trial, outcome of treatment of 1428 patients with eating disorders in six clinics in four countries has been reported, and the treatment has been reported to be more effective than a standard treatment in reducing body weight and improving health in obese adolescents in a randomized controlled trial. Mandometer has been cleared as an eating disorder conditioning tool, class 2 device by the FDA for the treatment of eating disorders.

Hypotheses: it is hypothesized that BED-patients eat more food at a higher speed than normal weight, healthy people. Because they do not compensate for their increased intake of food, BED-patients gain weight and become obese. It is also hypothesized that BED-patients will lose weight by practicing eating less food at a reduced speed using Mandometer.

Aims of trial: 166 BED-patients have practiced eating less food at a reduced speed at home over one week and a total of 354 BED-patients (including the 166 patients) have practiced eating in the same manner in Mandometer treatment of eating disorders. This trial aims at testing the hypothesis that BED-patients eat at a higher speed than normal weight, healthy control subjects. The trial also aims at examining whether BED-patients lose weight by practicing eating less food at a reduced speed. Normal weight, healthy control subjects will be recruited and tested for food intake and speed of eating at home over one week to meet these aims.

Patient admission and treatment: patients are continuously admitted to Mandometer treatment, which is a standard of care for eating disorders since 1997 in the Stockholm City Council. Hence, patients are treated according to the clinical practice of the Mandometer Clinic. 20% of the patients are admitted through referral within the Swedish Health Care System and 80% through self-referral, which is an option within the system. 85% of the 354 patients are women, their age and BMI were similar to those of men, their mean (SD) age at admission was 38 (15.4) years and their BMI was 37.1 (7.5) kg/m2, for women and men combined. The marked sex difference is expected, most patients with eating disorders are women. The patients first come to the clinic for instruction on how to use Mandometer at home and they return after one week for an eating examination. While eating at home over the first week using Mandometer is now part of clinical practice, this was not the case when Mandometer treatment for BED was launched. Thus, the first 354-166=188 patients who were admitted to Mandometer treatment did not eat using Mandometer at home before they entered treatment. Instead, these patients ate using Mandometer as part of the initial examination at admission. At that examination, which is referred to as an eating examination, all patients fill in the the Eating Disorders Inventory, the Comprehensive Psychopathological Self-Rating Scale, and questionnaires of the quality of life. As part of the eating examination, the patients are also examined by a physician, including blood pressure, heart and lung examination, and palpation of the stomach, and they are interviewed concerning their medical history and history of eating disorders. The patients´ body weight and composition are examined (Tanita BC-418 MA) and a record of their recent 24-hour food intake and eating habits is also taken. They are then diagnosed with BED using the Diagnostic and Statistical Manual of Mental Disorders.

Intervention: the patients are instructed in the use of Mandometer at their first visit to the clinic and they are given a Mandometer for use at home over the following week. They are encouraged to eat the foods they normally eat for lunch and dinner on five days of the week. 75% of the 166 patients had up to six meals with Mandometer in these tests. The average amount of food and the time to eat thus collected on Mandometer are loaded onto Mandobase, a custom made database used in the clinical practice of the Mandometer Clinic. The data are used to program Mandometer with training curves for the control of eating behavior, i.e., amount of food eaten and duration of the meal. The aim of Mandometer training is that patients should eat about 300-350 g of food for both lunch and dinner at a speed of about 25-30 g/min. These values correspond to those of previously tested normal weight, healthy control subjects, who ate using Mandometer in a research laboratory, but never at home. Patients are also provided with a schedule for 24-hour food intake, in which regularity of eating is stressed and between meal snacking is discouraged. Patients return to the Mandometer Clinic every sixth week for measuring their body weight and their eating behavior using Mandometer without training curves. The 188 patients who did not use Mandometer at home had their training curves similarly prepared but based on one test with Mandometer only, which was administered as part of their eating examination when they were admitted to the clinic. The patients are treated as out-patients, they come once/week over the first eight weeks and then once/second week. The average number of visits to the clinic is 30 over about a year.

Quality assurance plan: trained clinicians enter data from Mandometer into Mandobase. Mandometer yields quantitative, valid data, e.g., amount of food eaten (g) and duration of meal (min). These data can be accessed by researchers and clinicians who get log in information after they qualified for using Mandometer. Staff are trained at Mando Academy and certified before they can treat eating disorder patients. Mandometer Clinics are ISO 9001 and 14001 certified at regular audits.

Data checks: the data are not checked against data in other registers.

Source data verification: because Mandometer data are quantitative and valid, they are not in need of external verification. Speed of eating among obese subjects has been examined before, but using questionnaires rather than quantitative methods.

Data Dictionary: all measures are standard metrics, e.g., kg, g, min.

Standard Operating Procedures: Mandobase is designed for ease of data entry and data collection. Data in Mandobase is easily exported to standard statistical programs for analysis. A system for reporting adverse events is in operation and part of clinical practice, but not part of Mandobase.

Sample size, number of participants necessary to demonstrate an effect: The female BED-patients ate with a median speed of 44 (quartile range: 35-53) g/min and the male patients ate with a median speed of 47 (41-61) g/min during the week at home. Please not the large variation, making the slight sex difference statistically insignificant. Previous tests in the research laboratory have demonstrated that healthy, normal weight women and men eat with an average speed of approximately 30 g/min. Because the 25 pecentile speed of eating among female and male BED-patients is well above the mean speed of eating among the previouslty tested normal, healthy subject, it is necessary to recruite only 20 age-matched, healthy women and 10 men to demonstrate an effect. The effect will appear after one week of testing. The subjects will be recruited by advertisment at the local train station next to the campus of the Karolinska University Hospital, Huddinge. This manner of recruitment has been used in several experimental studies on eating behavior, using Mandometer, but normal weight, healthy subjects have never been tested at home. Also, previously tested subjects have been about 22 years old and the subjects to be recruited in the present trial should be matched for age with the BED-patients, i.e., they should be about 38 years old. A relatively extensive literature indicates, however, that eating behavior is similar among 20 and 40 year old humans. Even so, data from these control subjects are necessary for testing the hypotheses of the present trial.

Plan for missing data: data were obtained for all 166 BED-patients using Mandometer at home and for all 354 BED-patients subesequently using Mandometer in treatment; missing data is not a problem and inconsistent data or out-of-range results have not been obtained. While it is thought that BN- and BED-patients can eat very large amounts of food, an intake of 1900 g of food has been obtained once in one patient only, the average intake is approximately 430 g in hundreds of test meals on BED-patients.

Statistical analysis: Variations in food intake, meal duration, and speed of eating will be displayed as box plots, BMI will be shown as mean (SD), and changes in these measures and measures of secondary outcome variables over time will be analyzed using ANOVA.

Inclusion Criteria:

• patients with binge eating disorder and normal weight, healthy controls

Exclusion Criteria:

• ongoing drug abuse and/or other psychiatric diagnosis than binge eating disorder