Insight and its relationship to clinical outcomes in patients with schizophrenia or schizoaffective disorder receiving long-acting risperidone

Abstract

Poor insight is common in schizophrenia, predictive of non-compliance, and an impediment to effective patient management. We hypothesized that long-acting risperidone would be associated with enhanced insight, contributing to improved quality of life measures. In an international, open-label 50-week study, stable patients received long-acting risperidone every 2 weeks. Assessments included the Positive and Negative Syndrome Scale [PANSS; item G12 rated 'impaired insight and judgment' from 1 (no impairment) to 7 (severe impairment)]; Clinical Global Impressions-Severity (CGI-S); and the Medical Outcomes Study Short-form 36-item Health Survey (SF-36) (patient-rated quality of life). Correlation and regression post-hoc analyses examined associations between insight and other measures. At baseline, 314 (51.1%; N=614) patients had impaired insight (G12=3-7) and 83 (26.4%) achieved normal or near normal ratings at endpoint (G12=1-2). Symptom severity corresponded with insight: baseline mean+/-SD PANSS total scores were 56.0+/-14.4 in patients without impaired insight (G12=1-2); 73.4+/-15.7 with mild-moderate impairment (G12=3-4); and 86.0+/-17.4 with severe impairment (G12=5-7). These scores improved significantly in each group at endpoint (P<0.001). Improved insight ratings correlated with improvements in CGI-S (r=0.37); PANSS disorganized thought (r=0.46); negative symptoms (r=0.32); and anxiety/depression (r=0.24; P<0.001 all comparisons), but not quality of life ratings. The change in insight did not contribute significantly to the variance in SF-36 scores; however, changes in negative symptoms, anxiety/depression and CGI-S scores did contribute significantly. Long-acting risperidone was associated with improvements in insight, symptom domains, clinical status and quality of life measures. Associations were noted between patient-rated quality of life and specific symptom domains, but not insight.