Clinical and subthreshold eating disorders: case detection in adolescent schoolgirls

The point prevalence of eating disorders was investigated in schoolgirls from lower socio-economic classes and a method of case detection was tested. A two-stage procedure (self-report measures and interviews) was followed. All girls meeting at least one of the following three criteria were recruited as possible cases: EAT > or = 30, BMI < 17.5, BMI < 19 or > 24 and EAT > or = 20. The Mann-Whitney's two-tailed test and the chi-square test were used to assess the significance of differences between the possible cases and the others and between the false positives and the cases. A diagnosis of eating disorder was made for 24 subjects (8.1%): 5 (1.7%) with bulimia nervosa and 19 (6.4%) with eating disorders not otherwise specified (EDNOS). Of the 24 girls with eating disorders only 16 had high EAT40 scores (> or = 30). The other 8 cases were identified by the help of other criteria that proved to be useful. Furthermore we found data on five self-report instruments (EAT40, EDI, SCANS, SCL90, SEI). On almost all the scales, the possible cases obtained scores significantly different from the others. On the contrary, we did not find any instrument able to discriminate between false positives and cases.