Prediction of long-term outcome in schizophrenia by prognostic scales.

The predictive validity of four prognostic scales described in the literature (Gittelman-Klein Scale, Goldstein Scale, Phillips Scale, Strauss-Carpenter Scale) and of four newly developed prognostic scores was tested in two long-term followup studies on schizophrenic patients. Among the prognostic scales of the literature, the Strauss-Carpenter Scale showed the best results; it reached a level that was attained by only three of the newly developed scales. The three other scales of the literature could also prove their predictive validity. It may be concluded from the results that a scale has a comparatively high predictive value when certain dimensions and aspects are considered but that a certain maximum cannot, in general, be exceeded--even if more items are added to the prognostic scale. A higher number of items may be able to guarantee a better stability of predictive power in different samples.