Cognitive insight and clinical insight in schizophrenia : methodological, neurocognitive and psychopathological aspects and comparison to bipolar disorder

Abstract

Poor clinical insight (i.e. insight of illness) is considered a characteristic feature of schizophrenia. It is central for determining the presence of mental disorder and its prognosis, as well as for prescribing appropriate treatment and management. Patients with schizophrenia and other psychotic disorders suffer markedly from distorted beliefs. Yet, insight of illness does not specifically address the patient’s capacity to evaluate these disruptions. Evaluation of own anomalous beliefs and recognition of incorrect conclusions can be conceptualized as cognitive insight and assessed with Beck Cognitive Insight Scale (BCIS).<br> Lack of clinical insight may also be a major problem in psychotic disorders other than schizophrenia. Studies comparing levels of insight across diagnostic groups have not found substantial differences between patients with bipolar disorder and schizophrenia. Notably, the insight questionnaires used in these studies were not validated for patients with bipolar disorders. This also applies to the scale measuring clinical insight in this study, the Birchwood Insight Scale (IS). The first part of this thesis focuses on methodological issues regarding these questionnaires. In study I acceptable psychometric properties were found for the IS when applied to patients with schizophrenia and bipolar I disorder. However, for patients with bipolar II disorder the scale seemed to work poorly. Findings in the original BCIS study (Beck et al., 2004) suggested that the scale was able to differentiate between depressive patients with and without psychosis. We found that the psychometric properties were acceptable for the bipolar disorder group as well as the schizophrenia group. The scores of healthy controls, however, cannot be compared to patient scores without excluding items referring to psychotic experiences. Furthermore, the two BCIS subscales, self-reflectiveness and self-certainty, showed low or moderate correlation for all the three groups, indicating that they represent two different dimensions of cognitive insight. High scores on selfreflectiveness and low scores on self-certainty were considered as normal.<br> In the second part of the thesis the relationships between cognitive insight and clinical characteristics were explored. An association was found between self-certainty and delusions, in line with findings in the few previous studies. Furthermore, we explored the relationship between cognitive insight and different constellations of psychotic symptoms (delusions and/or hallucinations). Delusions irrespective of the presence or absence of hallucinations were associated with low self-reflectiveness and high self-certainty, reflecting low cognitive insight. In contrast, a subgroup of patients with solitary hallucinations showed high cognitive insight, which is quite interesting for understanding hallucinations but needs replication. The relationship between cognitive insight and neurocognition was also studied within the framework of this thesis. Consistent with a study on patients with first episode psychosis we found that low self-certainty is associated with high IQ and verbal learning. Verbal learning made a specific contribution in explaining self-certainty also when potential confounders were taken into account.<br> Taken together, the present thesis suggests that both clinical and cognitive insight could be measured with self-report questionnaires in schizophrenia and bipolar I disorder. Further, it suggests that hallucinations are related to better cognitive insight, and delusions with worse cognitive insight, and that low cognitive insight is associated with high IQ and low verbal learning. This indicates that cognitive insight is an important concept in understanding the psychopathology of psychotic symptoms.