ABSTRACT
The paper deals with the evaluation of a computer-based cognitive group training programme for schizophrenic patients. The evaluation of the training results of 100 patients who participated in an earlier training scheme showed ceiling effects in up to 46% of the cases. In the remaining cases the improvements varied in a wide range between 1.6 and 20%. These results implied that the training had to be adapted more closely to the specific needs of the individual patient. To this end we developed criteria for each task to decide whether training in this task is necessary or the performance is sufficient. This approach limits training to those tasks in which patients display deficits. Satisfactory improvement terminating training for the specific task, is defined as an increase in correct responses of at least 10%. For those tasks in which reaction times are used as performance measure, the criterion is a 10% increase in percentile rank. An analysis of the new training programme with the data of 20 patients revealed that 27% of the tasks showed ceiling performance with no further training need. In the remaining tasks the patients' performance attained the training criterion in 68%, at instances mostly even after one additional session. These preliminary results indicate that the new training scheme is able to provide more flexibility, individualization and an enhanced focus on the patients' deficits.