Comparing self- and expert rating: a self-report screening version (SIAB-S) of the Structured Interview for Anorexic and Bulimic Syndromes for DSM-IV and ICD-10 (SIAB-EX)

Abstract

Objective: Carrying out structured interviews in larger numbers by well-trained interviewers is costly and time consuming. Therefore, we developed parallel to the existing Structured Interview for Anorexic and Bulimic Syndromes (SIAB-EX) a similarily designed questionnaire for symptoms of disordered eating and related areas (SIAB-S). Method: 377 treated eating disordered patients were assessed within a two-week time period using both the SIAB-EX and SIAB-S. Results: Generally, self-ratings based on the SIAB-S were quite similar to expert ratings. Cohen's kappa showed good agreement between self- and expert ratings. Factor structure based on principal component analyses of expert ratings or self-ratings led to rather similar results confirming the robustness of the subscales in self- and expert ratings. Using expert rating as a criterion, the self-rating (SIAB-S) – which can more easily be used for screening purposes – had a sensitivity of 0.70, a specificity of 0.80 and a PPV = 0.91 for the DSM-IV diagnoses of AN and/or BN (worst ever condition). Diagnostic sensitivity (79/73%) and specificity (66/63%) were in an acceptable range (past/current). If we focus on the differences between the two approaches the following was found: self-rating (compared to expert-rating) resulted in lower scores for items inquiring about binges and inappropriate compensatory behaviour, attitudes towards food and eating, and social interaction. On the other hand, self-rating (compared to expert-rating) led to higher scores for items measuring general psychopathology and atypical bingeing. Conclusion: compared to the “gold standard” of data obtained with investigator-based standardised or structured interviews, data based on self-rating with items formulated clearly and concisely can lead to reliable and valid results. While complex issues (what is a binge) are difficult to assess in self-ratings, some (very personal) questions may even be better asked in a self-report questionnaire.