Screening disordered eating in a representative sample of the German population: Usefulness and psychometric properties of the German SCOFF questionnaire

- Screenings are essential to detect people at risk of developing an eating disorder.
- Psychometric properties of German SCOFF were assessed in general population sample.
- Results revealed low sensitivity, satisfying specificity, and many false negatives.
- Factorial analyses revealed a unidimensional model with an excellent model fit.
- There are limitations using the SCOFF in heterogeneous general population samples.

The prevention of eating disorders and the identification of high-risk individuals are essential for the public health sector. There is need for sensitive and specific screening instruments of disordered eating that can be applied in universal samples as an initial step into disease prevention. The SCOFF is a screening instrument for disordered eating, frequently used in international and cross-cultural contexts to detect individuals at risk. The objective of this research is to evaluate whether the SCOFF can be used as a screening tool for disordered eating in universal samples. This is the first study which examined the psychometric properties of the German version of the SCOFF in a general population sample. A representative sample (N = 2527) of the German population, aged 14-95 years, was recruited. Psychometric properties were determined including reliability, concurrent and construct validity, and factor structure. The prevalence of disordered eating was assessed. The prevalence of disordered eating in the general population was 10%. Using the established cutoff point of ≥ 2, values for diagnostic accuracy were 26% (sensitivity), 97% (specificity), 80% (positive predictive value), and 74% (negative predictive value). Factorial analyses revealed an excellent model fit of a unidimensional model. Due to its low sensitivity and a high percentage of false negatives, there are limitations in using the German version of the SCOFF in general population samples with wide age ranges.