Mj. Muller et al., Evaluation of standardized rater training for the Positive and Negative Syndrome Scale (PANSS), NERVENARZT, 71(3), 2000, pp. 195-204
The present analysis comprises 3 studies on the interrater reliability of t
he Positive a nd Negative Syndrome Scale for Schizophrenia (PANSS;German ve
rsion). To our knowledge this is the first empirical report on interrater r
eliability and on results of rater training of the German Version of the PA
NSS despite the widespread use of the scale. In a total of 47 training part
icipants with different clinical experience standardized PANSS rater traini
ng was carried out and subsequently evaluated. Therefore, concordance rates
with an expert standard (C) and weighted coefficient kappa (kappa(w)) were
calculated. As a main outcome of the studies, at least 3 training sessions
were necessary but also sufficient to reach acceptable interrater reliabil
ity of the PANSS (C>80%, kw>0,60). In training participants with low psychi
atric experience the level of interrater reliability of schizophrenic negat
ive symptoms did not reach the results of positive symptoms after the train
ing.
Despite some conceptual limitations with respect to negative symptoms, the
German Version of the PANSS seems highly suitable to assess a broad spectru
m of schizophrenic psychopathology in a reliable and economic manner. The p
resent results also underline the practicability of our recommendations for
conducting PANSS rater training in the clinical and scientific field as pa
rt of quality control and quality assurance in psychopathological assessmen
t.