The Efficacy of Metacognitive Skills Training in the Context of Forensic Psychiatric Care

Based on their multi-professional treatment plan in the Vanha Vaasa hospital, participation in the MCT group intervention is offered to patients who could benefit from it. The intervention under investigation is part of the standard care of the hospital. Participation does not prevent participation in other forms of rehabilitation, and patients can participate in the group even if they don't participate in the study. Being part of the control group does not prevent participation in the intervention group, but being part of the control group might delay participation. When these are in conflict, treatment takes precedence over research. Data is collected until the sample size is satisfactory (at least 20 to 25 patients). Patients in the MCT condition are compared to patients (n=30) and controls (n=30) measured with a psychological test battery not completing the group. The test battery is the same for all the groups. It consists of valid tasks measuring neurocognition, social cognition, and psychiatric symptoms and a task for measuring the tendency to jump to conclusions. Patients in the group condition are tested before the group and nine months after the group has concluded. For both control groups testing interval is one year. The comparison between the groups (intervention group, patient controls, and non-patient controls) is done by comparing the rate of change in the tendency to jump to conclusions. This comparison is done with regression analysis. If minor differences and equal variances are assumed (delta of slope 0.1), the power of the comparison is around 0.57. If larger differences are assumed (delta of slope 0.5), the power of the comparison approaches 1. In a previous study, a medium-sized comparative difference between patients in MCT condition and patients in cognitive remediation condition was found. To avoid problems with multiple testing, the differences in magnitudes of mood symptoms, delusions, paranoia, and psychological flexibility after the delay are assessed with MANOVA. In a recent meta-analysis, the observed effect of MCT on delusions was high medium (g=0.69). The observed effect on negative symptoms was small but significant (g=0.23). Consequently, the expected power of the MANOVA ranges from 0.48 to 0.99. Direct comparisons are made with discriminant analysis with identical power estimates. Univariate ANOVAs can also be used, but with much worse power estimates, when controlling for multiple testing.