Consistencies and discrepancies in self- and observer-rated anxiety scales. A comparison between the self- and observer-rated Marks-Sheehan scales

The Marks-Sheehan anxiety scales are the only scales where self-ratings and observer ratings are perfectly matched by the number, the content and the scaling of the items. Therefore these scales are an excellent tool to investigate the compatibility and to study different structures in self- and observer ratings. This was done by using the data material on the Marks-Sheehan scales of the Cross National Collaborative Panic Study. In this study 1168 outpatients who met the DSM-III criteria for panic disorder were randomly allocated either to alprazolam, imipramine or placebo treatment. Our results show that the Marks-Sheehan scales are highly comparable to other established rating scales. Both scales have a similar stable and consistent factor pattern for somatic symptoms but not for psychic symptoms of anxiety. Our results provide empirical evidence that the low consistencies between self- and observer ratings reported so far can be improved by using a rating scale which matches up item for self- and observer ratings. However, other sources of disagreement can only be solved by more elaborated item descriptions and training of patients as well as raters to obtain a better compatibility between self- and observer rating scales.