Methodological problems in polydiagnostic research.

Abstract

The growing application of the polydiagnostic approach makes it necessary to examine the methodological problems associated with the simultaneous assessment of multiple competing diagnoses. This paper contrasts the method of nonstandardized consecutive judgement of nonintegrated criteria lists with the method of a structured polydiagnostic interview with integrated criteria lists. The comparison of two polydiagnostic studies using both methods of assessment confirms that the unstructured use of nonintegrated and consecutively judged criteria lists is biased by a halo effect. This halo effect leads to a reduction of differences between the classifications of competing operational diagnoses and influences the type of patient distribution along the diagnostic dimension. This finding is interpreted as an argument to switch over to integrated criteria lists applied on the basis of an unstructured clinical examination or a structured interview.