ARTICLE
Differential diagnosis of eating disorders with the use of classification trees (decision algorithm)
 
More details
Hide details
1
Instytut Psychologii Stosowanej Uniwersytet Jagielloński w Krakowie
 
2
Instytut Psychologii Uniwersytet Śląski w Katowicach
 
 
Submission date: 2016-07-28
 
 
Final revision date: 2016-10-27
 
 
Acceptance date: 2016-10-28
 
 
Publication date: 2016-12-05
 
 
Corresponding author
Bernadetta Cecylia Izydorczyk   

Instytut Psychologii Stosowanej Uniwersytet Jagielloński w Krakowie, Grażyńskiego 53, 40-126 Katowice, Poland
 
 
Arch Psych Psych 2016;18(4):53-62
 
KEYWORDS
TOPICS
ABSTRACT
Aim of the study:
The aim of the study was to establish whether it is possible to make a correct diagnosis of various types of eating disorders, on the basis of measurements of only several variables.

Subject or material and methods:
A group of 213 females in the age range of 20 to 26 years of age underwent the study with the use of Eating Disorder Inventory (EDI), Socio-cultural Attitudes Towards Appearance Questionnaire-3, as well as the questionnaire for behaviour towards the body. Selection of dependent variable predictors of disease has been made. The classification tree has been developed.

Results:
Differential diagnosis in line with decision algorithm allows to establish a correct diagnosis for healthy individuals, in 100% of the cases, bulimia - in 78.72% of the cases, binge eating disorder - in 93.33% of the cases, and anorexia - in 86.36% of the cases.

Discussion:
Results indicate that the application of classification trees enables development of a decision algorithm and differential diagnosis of eating disorders. However, large sample is required for high predictive value of the decision tree. Therefore future research are necessary and decision algorithm should be verified on independent sample

Conclusions:
Psychological predictors of eating disorders may be arranged in the form of a classification tree. It is possible to make an accurate differential diagnosis of eating disorders on the basis of results of measurements of six variables.

eISSN:2083-828X
ISSN:1509-2046
Journals System - logo
Scroll to top