How similar are correlates of different subclinical eating problems and bulimia nervosa?

Background: This study examined the physical, parental, and psychological correlates of subtypes of high EAT-26 scores and bulimia nervosa (BN) in 155 girls/young women seen at two times over a 6-year period (ages 16 and 22). Participants were from white, upper-middle-class families. Methods: In young adulthood, participants were interviewed for lifetime history for an eating disorder; participants completed questions assessing subclinical problems and correlates at both times of assessment. Based on their EAT-26 scores and diagnostic status for BN at each time, girls were classified into the following groups: True Positive (BN and high EAT-26 scores), False Positive (high EAT-26 scores and no BN), and True Negative (low EAT-26 scores and no BN). Results: In adolescence, 4% of girls were classified as True Positives BN and 23% were classified as False Positives. Similar distributions were found at young adulthood (i.e., 5% True Positives BN and 21% False Positives). By using cluster analysis, two types of subclinical problems (False Positives) were identified at each time with some girls having only high Eating symptoms and others also having high psychopathology. The 'high Eating symptoms only' cluster was more similar to the True Negative group than was the 'high Psychopathology' cluster at each time. Conclusions: Individuals in the latter group experienced co-occurrence with depression, resulting in pervasive impairments in psychosocial functioning during both adolescence and young adulthood.