[Validation of the SCOFF questionnaire for screening of eating disorders in university women]. / Validación de la encuesta SCOFF para tamizaje de trastornos de la conducta alimentaria en mujeres universitarias.

BACKGROUND: The prevalence of eating disorders has been steadily increasing in recent years. This has led to a growing sense that instruments are necessary for early detection in high risk populations. OBJECTIVE: Herein the SCOFF (sick, control, one, fat, food) questionnaire was examined for its internal consistency, reliability, and criterion validity in screening eating disorders among university women. MATERIALS AND METHODS: A sample of 385 students was selected at random from six schools in the Universidad Autonóma de Bucaramanga, Colombia, and were evaluated independently with the SCOFF questionnaire and a CIDI (Composite International Diagnostic Interview). RESULTS: A total of 149 (38.7%) students were positive for eating disorder syndromes according to the SCOFF questionnaire, whereas 82 (21.2%) students met eating disorder using the CIDI. SCOFF questionnaire sensitivity was 78.4% (95%CI 64-88%), specificity 75.8% (95%CI 69-82%), ROC curve area 0.823 (95%CI 0.76-0.89), Cronbach's alpha 0.480, Cohen's kappa 0.433 (95%CI 0.32-0.55), and test-retest concordance 91.6% (95%CI 86-96). CONCLUSION: These data indicate the SCOFF questionnaire to be a good scale in screening for eating disorders among university women. Its added advantages are that the test is simple and brief.