[A new rating for the detection of dyskinesias. Clinical and metric problems]

In context with a study on tardive dyskinesia in a psychiatric clinic, two ward psychiatrists rated patients with respect to extrapyramidal dyskinetic reactions. (Age, sex, type of neuroleptic treatment and diagnosis were taken into account.) Few multiple dyskinesias have been found. When they occur at all then predominantly in the area of head and limbs; dyskinesia of the eyes and joints of knee and elbow are rare. Although three factors can be extracted by a factor analysis (head, trunk, and limbs) the insufficient reliability of single subscales becomes evident. In spite of these metric problems the use of the rating in the existing version is justified for clinical and educational reasons. It is recommended to secure statistical reliability mainly with the total score which provides a good interrater reliability (.88 and .82) and a sufficient internal consistency (.70). For analyses of the course we suggest a data screening at the level of single items or subscales.