Cognitive insight and acute psychosis in schizophrenia


Emre Bora, MD, Melbourne Neuropsychiatry Center, University of Melbourne, Alan Gilbert Building, NNF level 3, Vic. 3053, Australia. Email:


Abstract  Cognitive insight is a new concept. There are very few data regarding the relationship between cognitive insight and positive symptoms. The goal of the present study was to investigate the impact of acute psychosis (delusions and hallucinations) on overconfidence in judgments and self-reflectiveness of patients with schizophrenia. The Beck Cognitive Insight Scale was used to compare the cognitive insight of schizophrenia patients with (n = 93) and without (n = 45) current psychotic symptoms. Clinical symptoms and clinical insight of the patients were also assessed. The present findings suggest that both overconfidence in judgments and impaired self-reflectiveness are associated with acute psychosis. Only diminished self-reflectiveness seem to be improved following hospitalization. Although overconfidence of schizophrenia patients in their judgments was more severe in schizophrenia patients with psychotic symptoms, self-certainty of schizophrenia patients may be a relatively persistent characteristic that is also present after recovery of psychosis. Studies with larger samples involving follow up for longer periods will be valuable to understand the nature of the relationship between cognitive insight and clinical symptoms of schizophrenia.


Lack of insight has been accepted as one of the most important features of schizophrenia. Previously, insight in psychosis had been regarded as a categorical concept; the patient may possess or lack insight. However, more recent authors define insight as a multidimensional concept that occurs on a continuum.1,2

The unawareness of illness in schizophrenia is generally evaluated by a variety of clinically oriented insight scales.3–5 Although the concept of clinical insight has contributed to the understanding of prognosis and treatment of schizophrenia, this concept also has some limitations. Clinical insight ratings may sometimes reflect the superficial beliefs of the patients or explanations that they received from their doctors or from other treatment staff. The patients can accept the explanation that they have a mental disorder without being convinced of this. Having a real and enduring awareness of illness may be more essential for helping the patients; for example, to increase their treatment compliance.

Recently, a new concept, cognitive insight, was proposed by Beck et al.6,7 Patients with schizophrenia, especially deluded patients, were proposed to have limited capacity for evaluation of their erroneous inferences and they seem to be relatively resistant to corrective feedback. Cognitive insight is concerned with distancing from and reassessing distorted beliefs and misinterpretations. Even though patients may accept an illness explanation for their symptoms and agree that this explanation make sense (intellectual insight), still this may not produce any significant change in their underlying belief system (cognitive insight).8 The Beck Cognitive Insight Scale (BCIS) was developed to evaluate the patients' reflectiveness and their overconfidence in their interpretations of their experiences. This scale was shown to be a valid and reliable instrument for inpatients and outpatients with schizophrenia.6,7 Cognitive insight and clinical insight scales seem to measure at least partly different aspects of the concept of insight. The BCIS ratings were reported to be only moderately correlated with the Birchwood insight scale in outpatients.7

Cognitive insight in schizophrenia patients may be related to the current symptoms and stage of the illness. The relationship between cognitive insight and clinical symptoms of schizophrenia has not been adequately studied. Pedrelli et al. reported only mild correlations between positive, negative symptoms and the self-certainty scale of the BCIS in outpatients.7 Only Warman et al. investigated the impact of active delusions on cognitive insight in schizophrenia.9 They compared cognitive insight levels of 33 deluded and 11 non-deluded patients with a control group. Their findings, compared to the controls, showed that although the patients with active delusions had higher self-certainty scores, non-deluded patients had higher self-reflectiveness scores. However, because the sample size was extremely small for non-deluded patients, this finding can be accepted as a preliminary finding. Furthermore, the direction of relationship between self-certainty and delusions is unclear. The confidence in errors may be one of the causes for delusions as suggested by some authors.10 If overconfidence is a temporary characteristics of schizophrenia patients that is seen only during psychotic exacerbations, overconfidence should be decreased following symptom recovery. However, it is also possible that certainty about judgments may be relatively permanent characteristics of delusion-prone individuals. For example, even schizotypic individuals have been reported to be more confident about their judgements.11 Only longitudinal studies can definitely answer this question.

The primary aims of this study were (i) to compare the cognitive insight impairments of schizophrenia patients with and without psychotic symptoms; and (ii) to investigate the relationship between symptom recovery and change in cognitive insight.



The subjects were 138 schizophrenia patients who were treated at Ege University Department of Psychiatry. The patients were diagnosed according to DSM-IV criteria. The patients diagnosed with schizophrenia (age range, 16–55 years) and who were able to complete a self-report form were included in the study. The patients with other schizophrenia spectrum disorders were excluded. The study was explained to the patients and their informed consent was obtained. The patients were recruited from two different settings: (i) the inpatient unit (n = 77); and (ii) the outpatient Psychosis Unit of Ege University Department of Psychiatry (n = 61). The outpatients were relatively stable; they did not experience significant symptom and treatment change in the last 4 months. Symptoms were assessed with the Positive and Negative Symptoms Scale (PANSS).12,13 Sixty-seven of 77 inpatients (87%) and 26 of 61 outpatients (43%) had scored ≥3 on the delusions or hallucinations items. These 93 patients were classified as having psychotic symptoms (currently psychotic group). The remaining 45 patients were classified as currently non-psychotic. Eighty-eight of the patients had delusions and 48 of them had reported hallucinations (only five of them had only hallucinations). Except one of the outpatients, all other patients were receiving antipsychotic treatment (74% only atypical neuroleptics, others only typical, or an atypical–typical neuroleptic combination). Because the current health-care regulations in Turkey assist government employees to access university hospital services, the socioeconomic level of current sample was relatively higher than general schizophrenia population of İzmir. A significant number of the patients were children, spouses of caregivers who were working in community sector, or community employees themselves.

Thirty-four consecutive patients were initially included into the longitudinal part of the present study. All of these patients were psychotic at admission. Four of the patients who were discharged from the inpatient unit without the consent of the staff or who refused to be reinterviewed at discharge were excluded from the longitudinal part of the study. The final longitudinal sample consisted of 30 patients. These patients were interviewed as soon as possible, in the first week of their admission (mean time, 3.3 ± 1.9 days after admission) and they were reinterviewed at discharge (average time between two interviews was 31.2 ± 11.5days). The PANSS, the Scale of Unawareness to Mental Disorders (SUMD) and the BCIS were readministered to the patients during the second interview.

Insight measures

Beck Cognitive Insight Scale

The BCIS is a 15-item self-report measure designed to assess cognitive insight in patients with psychosis.6 Participants rate the statements from 0 (do not agree at all) to 3 (agree completely). The BCIS consists of two factors, Self-Reflectiveness (nine items) and Self-Certainty (six items). A composite index is also calculated by subtracting the score of Self-Certainty from the score of Self-Reflectiveness and is considered as a measure of cognitive insight. The Self-Reflectiveness items of the scale were written for capture of patients' objectivity, reflectiveness and openness to feedback (e.g. ‘Some of the ideas I was certain were true turned out to be false’). Self-certainty items were written to address decision-making regarding mental products: certainty about being right and jumping to conclusions (e.g. ‘I cannot trust other people’s opinion about my experiences').6,7 The BCIS was also reported to be a reliable and valid instrument for Turkish patients with psychosis.14

Scale of Unawareness to Mental Disorders

The SUMD4 was used to assess the clinical insight of the patients. The Turkish version of the scale was shown to have good interrater reliability (interclass correlation coefficients of the subscales were between 0.72 and 0.93).15 For the purpose of the present study we reported only the total SUMD score, which was calculated by summing three global items of the SUMD: (i) current awareness of having a mental disorder (SUMD1); (ii) current awareness of the response to medication (SUMD2); and (iii) current awareness of the social consequences of the disorder (SUMD3).

Data analysis

Between-group differences were analyzed on t-test. Paired-t-tests were used to analyze improvement of insight and clinical symptoms. Spearman's correlation was used to investigate the relationship between cognitive insight, clinical symptoms and changes in these variables. The threshold for statistical significance was determined as P < 0.05. For statistical analysis, SPSS 11.0 (SPSS, Chicago, IL, USA) was used.


There was no significant group difference for the demographic variables. The patients with psychotic symptoms had significantly more positive symptoms and they also had more negative symptoms (Table 1). The patients without psychotic symptoms had better clinical insight than psychotic patients on SUMD.

Table 1.  Between-group differences
 Currently psychotic (n = 93)Currently non-psychotic (n = 45)TP
  1. PANSS, Positive and Negative Symptoms Scale; SUMD, Scale of Unawareness to Mental Disorders.

Education (years)10.83.711.
Duration of illness (years)
Hospital admissions2.
PANSS positive19.<0.001
PANSS negative19.
Gender male : female54:39 29:16   
Schizophrenia subtype, n (%)
 Paranoid61 (65) 33 (74)   
 Undifferentiated29 (32) 6 (13)   
 Disorganized3 (3) 1 (2)   
 Residual  5 (11)   
Cognitive insight
 Composite index3.
Clinical insight
 Total SUMD9.<0.001

Spearman's correlation was applied to all 138 patients. Positive symptoms were inversely related to Self-Reflectiveness (ρ = −0.24, P = 0.005) and composite index (ρ = −0.29, P = 0.001). They were also associated with Self-Certainty (ρ = 0.29, P = 0.001). Negative symptoms were related only to Self-Reflectiveness (r =−0.25, P = 0.004). There were strong negative relationships between the SUMD total score and BCIS Self-Reflectiveness (ρ = −0.56, P < 0.0001) and composite index (ρ = −0.55, P < 0.0001). The SUMD total was also, although less strongly, related to Self-Certainty (ρ = 0.28, P = 0.002). The BCIS factors were not related to duration of illness, education, age or number of hospitalizations.

The patients without psychotic symptoms demonstrated significantly more Self-Reflectiveness than patients with psychotic symptoms (Table 1). The patients with current psychotic symptoms were significantly more confident about their judgments than currently non-psychotic patients. Active psychosis was also related to poorer cognitive insight (composite index). The BCIS factors were not significantly different between the patients with only delusions and those with delusions and hallucination.

Changes in insight ratings and their association with improvement in symptoms

There was a significant improvement of symptoms and clinical insight deficits after the inpatient treatment (Table 2). While self-certainty scores did not change at all during treatment, there was improvement in Self-Reflectiveness and composite index. Spearman's correlation was also used to investigate the relationship between symptom and cognitive insight change (n = 30). The change in positive or negative symptoms was not significantly correlated with change in any of the cognitive insight ratings. There was no significant relationship between change in the SUMD total and the BCIS factors.

Table 2.  Changes in symptoms and insight measures
 Admission (n = 30)Discharge (n = 30)tP
  1. BCIS, Beck Cognitive Insight Scale; PANSS, Positive and Negative Symptoms Scale; SUMD, Scale of Unawareness to Mental Disorders.

PANSS positive23.37.512.34.97.8<0.001
PANSS negative21.99.715.97.35.4<0.001
 Composite index4.
 Total SUMD9.<0.001


The present study investigated cognitive insight in schizophrenia patients with and without current psychotic symptoms. It also investigated the relationship between acute symptom recovery and change in cognitive insight. The patients with current psychotic symptoms showed overconfidence in their judgment and they were also impaired in self-reflectiveness. While impairment in self-reflectiveness was decreased following inpatient treatment, overconfidence of patients in their judgments did not reduce.

Previously, individuals with schizophrenia have been demonstrated to have overconfidence in their judgments compared to depressed patients, and patients with delusions also had been shown to have higher self-certainty than healthy controls.6,9 The present findings also suggested that schizophrenia patients with current psychotic symptoms are more overconfident of their judgments. Some previous work suggested that making decisions without gathering sufficient information and being overconfident of these decisions are characteristics of deluded patients.10,16 These finding raised the possibility that self-certainty may be a state-dependent variable that can have a role in acquisition of delusions. However, the present findings did not demonstrate a decrease in overconfidence of schizophrenia patients after symptomatic improvement. Self-certainty was also shown to be high in normal subjects who are accepted to be delusion prone.11 Recently, the findings of Peters and Garety also demonstrated that individuals with delusions required the same amount of information to make a decision even after their delusions remitted.17 These findings suggest that overconfidence of schizophrenia patients with delusions maybe a more persistent feature of these patients that extends beyond active psychotic states. Being overconfident in judgments is an outcome that can be influenced by many other factors such as personality. Poor insight was also shown to be related to better self-esteem and coping styles, which may be associated with self-certainty.18 As a result, overconfidence in own judgments may be a stable characteristics of delusion-prone schizophrenia individuals rather than a temporary feature in acute psychosis. However, we should be cautious about our results, for the following reasons. One month of treatment and acute recovery of psychotic symptoms may not mean that the patient has totally recovered from the underlying psychotic state. Reassessing cognitive insight after a longer symptomatic recovery can help to answer this question. Also, the power of the present longitudinal sample may be insufficient to show a possible modest decrease in self-certainty.

Previous research demonstrated that patients with schizophrenia are less self-reflective.6,9 However, only one study compared self-reflectiveness of deluded and non-deluded schizophrenia patients. Interestingly, Warman et al. reported, in a small sample, that schizophrenia patients with delusions (n = 33) are more self-reflective than patients without delusions (n = 11).9 The present study showed the opposite: schizophrenia patients with current psychosis are less self-reflective than currently non-psychotic patients. We also found that self-reflectiveness is strongly associated with better clinical insight. Unlike overconfidence in judgments, both clinical insight and self-reflectiveness improved following reduction of psychotic symptoms. Clinical insight was reported to be improved especially following short-term hospitalization and in early phases of recovery, but there were no data regarding cognitive insight.19–21 The present findings suggest that self-reflectiveness, compared to self-certainty, may be closer to clinical insight and more open to modulation by the state variables. Further studies are needed to clarify the nature of the relationship between psychotic symptoms and self-reflection. Another interesting issue might be investigation of the relationship between hallucinations and cognitive insight. As far as we know no study has investigated this issue. In a post-hoc analysis we could not find a difference in cognitive insight items between patients with and without hallucinations. However, because most of the present patients with hallucinations also had delusions, the sample is not suitable for investigating this question. Future studies investigating cognitive insight in patients with only hallucinations will be useful.

The strengths of the present study were the relatively large sample size for cross-sectional comparison and the simultaneous application of cross-sectional and longitudinal approaches. One limitation of the study was the relatively small sample size for the longitudinal data. Another limitation may be the self-report nature of the BCIS; but assessment of cognitive distortion of the patients with a more experimental approach can overcome this difficulty.

In conclusion the present findings suggest that both overconfidence in judgments and impaired self-reflectiveness are associated with acute psychosis. Only diminished self-reflectiveness seems to be improved following hospitalization. Studies with larger samples that will follow up patients for a longer period will be valuable to understand the nature of the relationship between cognitive insight and clinical symptoms of schizophrenia.