Evaluation of standardized rater training for the Positive and Negative Syndrome Scale (PANSS)

The Positive and Negative Syndrome Scale (PANSS) is employed increasingly for the evaluation of therapeutic outcome in studies on schizophrenia. Rater training is important to improve the concordance and accuracy of ratings; however, there are no established guidelines for carrying out such training. We conducted rater training, under clinical conditions, of psychiatrists and clinical psychologists who were rather unfamiliar with the PANSS. Based on videotapes of PANSS interviews, all participants were trained during five successive standardized weekly sessions. The results were analyzed with respect to conventional criteria of concordance with standard expert ratings and interrater reliability. The main objective was to evaluate the number of training sessions which are necessary and sufficient to achieve acceptable PANSS rating results. Additionally, differences in training outcome for positive, negative and general psychopathological symptoms and between subgroups of different clinical and PANSS experience were considered. After three weekly sessions, satisfactory concordance of about 80% of clinicians on the PANSS total scale was obtained. However, in comparison with the positive and general psychopathological subscales, the PANSS negative-symptom subscale yielded somewhat less satisfactory results, with about 70% of the raters achieving sufficient accuracy. Intraclass correlations corroborated these findings. No substantial differences in training outcome were found between subgroups of different clinical and PANSS experience. We conclude that at least three standardized PANSS training sessions are recommended to obtain satisfactory accuracy of ratings.