BACKGROUND & AIMS: Several questionnaires to screen for eating disorders have been validated in Malaysia. However, these tools are lengthy, and require specialist interpretation. The sick, control, one stone, fat, food (SCOFF) is easy to administer by non-specialists, but has not been validated in Malaysia. Therefore, the aim of our study was to validate the SCOFF on a non-clinical sample of tertiary students to determine if it could identify individuals with an eating disorder.
METHODS: We recruited second year tertiary students from five faculties in a university in Malaysia, from June-November 2014, who could understand English. The SCOFF and the EAT-26 were administered at baseline. Two weeks later, the SCOFF was re-administered to assess for reliability.
RESULTS: A total of 292 students were approached, and all agreed to participate (response rate = 100%). There was moderate correlation between the total SCOFF score with the EAT-26's dieting domain (spearman's rho = 0.504, p < 0.001), bulimia and food preoccupation domain (spearman's rho = 0.438, p < 0.001), and total score (spearman's rho = 0.483, p < 0.001). The internal consistency of the SCOFF was low (Cronbach alpha = 0.470). At retest, kappa scores ranged from 0.211 to 0.591. The sensitivity of the SCOFF was 77.4%, and its specificity was 60.5%. The positive predictive value was 18.9%, and its negative predictive value was 95.8%.
CONCLUSIONS: The SCOFF was found to have adequate convergent validity and stable reliability. However, its internal consistency was low. The SCOFF can still be used in clinical practice. However, its positive results should be interpreted with caution due to its low positive predictive value.
KEYWORDS: Eating disorder; Malaysia; SCOFF; Sensitivity; Specificity; Validation