NCJ Number
170590
Journal
Substance Use and Misuse Volume: 33 Issue: 3 Dated: (1998) Pages: 765-791
Date Published
1998
Length
27 pages
Annotation
The use of Artificial Neural Networks (ANN) to study eating disorders was examined in an experimental study that used data from in 172 women at the Center for the Diagnosis and Treatment of Eating Disorders at the St. Eugenio Hospital in Rome, Italy.
Abstract
The participants were grouped into four categories based on the diagnosis made by the hospital specialist. These categories were Anorexia Nervosa (AN), Nervous Bulimia (NB), Binge Eating Disorders (BED), and Psychogenic Eating Disorders that are Not Otherwise Specified (PED-NOS). The research used 124 variables in the areas of general information, eating behavior, treatment and hospitalization, drug use, menstrual cycles, weight and height, laboratory analyses, and psychological evaluation. The research focused on the accuracy of ANN in identifying participants with AN and NB. Each of six experiments used different variables. Results revealed that ANN provided accuracy at an overall rate of 86.94 percent, even when the system did not receive data on menstruation and other factors known to indicate eating disorders. Findings suggested that the use of ANN may become useful for developing a diagnostic tool for use by physicians who lack training on eating disorders. Such an approach would increase the chance of early diagnosis, which is the main requirement for effective treatment. Further research is recommended. Tables, author biographies and photographs, and 13 references (Author abstract modified)