Original Article
Objective
:
The purpose of the present study was to develop a Korean version of the Beck Cognitive Insight Scale (BCIS-K) and to evaluate its reliability and factorial validity by using a sample of Korean patients with psychotic disorders.
Methods: The Beck Cognitive Insight Scale (BCIS) was translated into Korean and its psychometric properties were then examined. The linguistic equivalence of the translated form was verified by the back-translation method. The final, verified translation was administered to patients at Seoul National Hospital, and the responses of 78 adults were analyzed. We evaluated internal consistency, the relationship between the subscales and the composite index, and the correlation with both the Korean version of the revised insight scale for psychosis (KISP) and the 104-item Present State Examination (PSE-104).
Results: The study results demonstrated that the coefficient α and correlations were similar with those of the original BCIS scale and that the BCIS-K was moderately correlated with the KISP and it was not correlated with the PSE-104.
Conclusion: The findings verified the accuracy of the Korean translation of the BCIS and they confirmed the BCIS-K's capability to satisfy the developers' intent. The results also confirmed that the BCIS-K, like other measures of insight, measures different but related aspects of insight.
Correspondence :
Jong-Il Lee, MD, PhD, Department of Psychiatry, Seoul National Hospital, Junggok 3-dong, Gwangjin-gu, Seoul 143-711, Korea
Tel : +82-2-2204-0104, Fax : +82-2-2204-0392, E-mail : lji7542@hanmail.net
Introduction
A lack of insight has been found to be highly prevalent among schizophrenics. Severe and persistent unawareness of their illness by the patient is damaging not only to the afflicted person's relationships, but also to their chances of recovering and leading more satisfying and productive lives. Therefore, interventions aimed at either improving insight directly or that address the negative consequences of poor insight must be developed. A number of reviews have further paved the way for this research since Aubrey Lewis's first foray into this area in 1934.1 Most recent authors have described insight as a multidimensional construct that occurs on a continuum.2 According to Beck et al.,3 the current definitions of insight and the measures that are derived from them have contributed to the understanding of the prognosis and treatment of patients' illness, but these researchers have not addressed an important dimension of insight that they called cognitive insight. For example, when questioned about their symptoms, patients may repeat what they have been informed, along the lines of "I have a mental illness or schizophrenia. So I take the pills."The patients' explanations could be described as an expression of the insight under the view of the clinical phenomenology. However, upon an in-depth exploration, their explanations do not reflect their underlying delusional belief system. Even though patients may accept an explanation and agree that it makes sense, they may have a limited capacity to evaluate their anomalous experiences and their erroneous inferences. These cognitive deficiencies in psychotic patients include not only impairment of the objectivity about their distortions, but also resistance to corrective information from others and the patient's overconfidence in conclusions. This form of insight has been termed cognitive insight, and it may be related to the development and maintenance of patients' delusional beliefs and thinking, whereas clinical insight based on clinical phenomenology is essential for making the diagnosis and treatment. As an initial step in investigating these psychological process, Beck and colleagues developed the Beck Cognitive Insight Scale (BCIS), which focuses on self-reflectiveness about unusual experiences, the capacity to correct erroneous judgments and the certainty about mistaken judgments.3
The BCIS measure is a 15-item self-report checklist, and it assesses the patients' self-reflectiveness, as measured by their capacity to observe themselves at a distance from their anomalous experience and being open to external feedback, and also their self-certainty as measured by their confidence of their decision making. The patients who take this test rate the extent to which they agree with statements on a scale from 0 (do not agree at all) to 3 (agree completely). BCIS has two factors: selfreflectiveness (nine items) and self-certainty (six items). A composite index of BCIS is considered a measure of cognitive insight, and it is calculated by subtracting the score for the selfcertainty scale from that of the selfreflectiveness scale. A higher composite index indicates greater cognitive insight. The original BCIS version was used to examine the psychometric properties in a sample of 150 inpatients with and without psychotic disorders,3 and a follow-up study investigated a group of middle age and older adults with schizophrenia in an outpatient setting.4 These two previous studies supported a two-factor solution for the BCIS measure, i.e., self-reflectiveness and self-certainty, and the studies showed borderline internal consistency and moderate correlation with other insight scales. In addition, according to Warman and Martin,5 a normal population with diverse degrees of being prone to delusion responded differently on the BCIS as compared to patients with or without psychotic disorders.
The aims of the present study were to translate the BCIS into Korean and assess the validity of the Korean version of the Beck Cognitive Insight Scale (BCIS-K). In addition, we compared the psychometric properties of the BCIS-K with those of the BCIS that were observed by Beck et al.3 and Pedrelli et al.4 We intended for the results to confirm whether the translated BCIS-K version is suitable for the purpose of its developers and whether the BCIS-K measures insight, like the BCIS and other measures of insight, but if it emphasize different aspects of the construct. We analyzed the relationship among the BCIS-K, the KISP and the PSE-104. Further, we tested the relationship between the BCIS-K and the severity of the patients' condition.
Methods
Participants
The study sample was recruited at random from the subjects who were medicated patients from a variety of settings, including in-patient acute units, in-patient chronic units and out-patient clinics, at Seoul National Hospital. The survey was conducted from May 15 to July 19, 2006. The diagnosis was based on the Diagnostic and Statistical Mannual of Mental Disorders (DSM-IV-TR)6 criteria for schizophrenia, schizoaffective disorder and other psychotic disorders, a detailed clinical interview and a review of the prior records. Those patients with a history of brain damage or neurological disease that might cause cognitive impairment were excluded from this study. All study subjects gave us their written informed consent. Subjects were excluded from the analysis for providing incomplete answer and incomplete data, and if they didn't provide written consent. The resulting participants and whose data was analyzed consisted of 78 persons with a diagnosis of schizophrenia, schizoaffective disorder and other psychotic disorders. These results were not available to the personnel who were eliciting information on the subjects' mental state. The study's protocol was approved by the Seoul National Hospital Institutional Review Board.
The patients were assessed using the following protocol.
1) A clinical interview was done to collect the following information. (a) Basic demographic and clinical data to explore the correlation of the insight. (b) Standard clinical assessment of insight [the 104-item Present State Examination (PSE-104) PSE with a clinician-rated evaluation] as a validated comparative tool.7 Although the Korean version of the PSE has not yet been validated, this study uses it as a matter of convenience. (c) Global Assessment of Functioning Scale (GAF with a clinician-rated evaluation) to report a clinician's judgment of a patient's overall level of functioning. The current study used the Korean version of the GAF.8 (d) Clinical Global Impressions-Severity (CGI-S with a clinician-rated evaluation) to measure the overall severity of psychiatric disturbance.
2) The Korean version of the revised insight scale for psychosis (KISP), which was self-administered.
3) The BCIS-K, which was self-administered.
Instruments
We hypothesized that the previously identified psychometric properties of the BCIS3,4 would be replicated in this study and that the BCIS-K construct would be consistent with the BCIS. We also expected that the BCIS-K would be only moderately correlated, or not correlated at all, with KISP9,10 and PSE-1047 because these instruments, including the BICS-K, measure relatively different, but related aspects of insight. The KISP and PSE-104 were chosen for making comparison. The PSE-104 is based on clinical phenomenology and it is relatively suitable for the assessment of clinical insight. KISP has the benefit of being a short self-report measure, like the BCIS-K, and it has shown adequate reliability and validity.9 The KISP is not only based on the clinical phenomena of insight like the item 3 "I am ill", but it also reflects a patient's inference and attribution regarding the cause of their change. As well as fundamentally assessing the clinical insight, the KISP also reflects the cognitive process of the patients with psychotic disorder, thereby supporting our expectation for low correlation between the BCIS-K and the KISP and PSE-104.
The Korean version of the Beck Cognitive Insight Scale
Under license of the publisher, the 15-items of the BCIS were translated into Korean via two independent forward translations, reconciliation and two independent back-translations, and then via a pilot test on five schizophrenics. The two back-translations were sent to the author of the original scale to confirm their equivalence with the original scale. The final translation, approved by the author, was examined psychometrically in the following study. To preserve equivalence for the cross-cultural adaptation of the BCIS, we carefully followed the guidelines proposed by Guillemin et al.,11 including those on the back-translation techniques.12
The Korean version of the revised Insight Scale for Psychosis
The original insight scale for patients with psychosis was first published in 1992.13 The final version of the insight scale (30 items) has undergone some refinement and has had some items deleted, others added and several rephrased, and it has undergone re-standardization.11 The scoring of the items on the scale is dichotomous with a score of 1 given for insight and a score of 0 for no insight. Scores of 30 indicate full insight, and 0 is no insight. This scale is based on the concept of insight as a form of self-knowledge the patient has about his/her illness and how this might affect his/her ability to function and interact with the environment. This concept focuses on awareness and articulation of the changes in subjective experience rather than on more elaborated beliefs/attitudes about such changes.
Consequently, the instrument presented here is not only based on the wider concept of insight than is the traditional insight scales, but it is also able to measure insight while emphasizing different aspects of the construct along with the BCIS. Kim and colleagues translated the revised Insight Scale for Psychosis into Korean and they evaluated the KISP in a sample of 81 Korean schizophrenics.9
104-item Present State Examination
This measure is the standard clinical assessment of insight and it has been validated as a comparative tool.7 This is a single-item, clinician-rated evaluation that's based on answers to specific questions relating to psychotic symptoms and the answers are rated on a 4-point scale from 0 (full insight) to 3 (denied condition entirely). Patients assigned scores of 0, 1 and 2 report awareness of their illness, but with differing degrees of uncertainty on the examiner's part, while a score of 3 indicates no insight. The PSE-104 has value for determining the presence of mental illness. However, it does not directly address the patients' limited capacity for evaluating their anomalous experiences and their erroneous inferences. Consequently, the PSE-104 only emphasizes the patients' awareness of illness and it is inclined toward clinical insight.
Statistical analyses
Descriptive statistics were used to characterize the patients in terms of the socio-demographic and clinical data. Cronbach's alpha was used to determine the internal consistency of the BCIS-K. To estimate the test-retest reliability, the BCIS-K was administered to 20 patients on two different occasions that were separated by a 2-day interval. The Pearson correlation between the composite index scores on both occasions was calculated. The condition of the patient with psychosis and who was receiving treatment is changing all the time, and allowing a longer interval between test and retest might have revealed changes in their mental state that were unrelated to faults in the structure and content of the scale itself. The issue of what is the ideal interval duration to conduct the test-retest of such scales remains an important methodological challenge. Factor analysis and Pearson correlation between the BCIS-K and the other insight scales were conducted. Lastly, we calculated the Pearson correlation between the BCIS-K and the GAF and CGI-S scores to explore the cross-sectional relationships between the BCIS-K and the psychiatric conditions.
Results
Socio-demographic factors
The background and clinical characteristics of the patients are presented in Table 1. The patients showed a broad range of characteristics and all the patients were receiving antipsychotic medicines at the time of assessment.
Factor analysis
The results from the Varimax-rotated principal factor analysis indicated that the two-component solution displayed the simplest structure for all but two items (items 11 and 3) for the loading saliently (≥0.30) on one component as opposed to the other component. Table 2 shows the factor loadings for the 15 BCIS-K items. Despite the overall good fit, items 11 and 3 had lower factor loading than 0.3, but a higher loading than the opposite component. The first component addresses a patient's certainty about beliefs or judgments, and this corresponds to self-certainty. The second component is interpreted as an expression of introspection and the willingness to acknowledge fallibility and this is applicable to self-reflectiveness. In spite of the fact that the factor loading order was different from that of the BCIS as observed by Beck et al.,3 each item was comprised of the same components indicated on the BCIS and the result was also replicated on the follow-up study, which investigated patients with schizophrenia and schizoaffective disorder with using the BCIS.4
Subscale internal consistencies and Test-retest reliability (Table 3)
The coefficient α for the self-reflectiveness and self-certainty scores was 0.642 and 0.752, respectively. All of the item-total correlations of the BCIS-K items with their respective subscale scores were significant (p<0.01, one-tailed test).
The item-total correlation with self-reflectiveness ranged from 0.090 to 0.482, and the subscale for self-certainty had item-total correlation ranging from 0.159 to 0.667. Test-retest reliability was carried out on 20 patients. A significant Pearson correlation was found between the composite index of the BCIS-K on both occasions
(r=0.612, p<0.05).
Comparison of the Korean version of the Beck Cognitive Insight Scale and the other measures
The BCIS-K self-reflectiveness and self-certainty subscales, along with the composite index, were correlated with the KISP and PSE-104. Table 4 shows that significant correlations were limited to two: the BCIS-K composite index and the KISP
(r=0.525, p<0.01), and self-reflectiveness and the KISP (r=0.511, p<0.01). There was no significant correlation between the BCIS-K and the PSE-104.
Relationship between the Korean version of the Beck Cognitive Insight Scale and the patients' condition (Table 4)
Correlations were calculated between the BCIS-K and the patients' condition, as assessed by the GAF and CGI-S. None of the correlations was statistically significant.
Discussion
The primary goals of this study were to translate the BCIS into Korean and to assess the reliability and va-lidity of the resulting BCIS-K. The factor analysis of this study supported the validity of the two-factor structure reported by the developers of the BCIS. The BCIS-K measure showed adequate test-retest reliability, moderated internal consistencies, a moderate correlation with the KISP and no correlation with the PSE-104. Some consid-erations need to be taken into account when interpret-ing the current results. First of all, a fair amount of variability within the scales was observed, as indicated by the low factor loadings of two items. One item had a loading close to the cutoff of 0.30 (#11, "I cannot trust other people's opinion about my experiences."), while the other had a loading below the cutoff of 0.30 (#3, "Other people can understand the cause of my unusual experiences better than I can."). Further, one item (#12, "If somebody points out that my beliefs are wrong, I am willing to consider it.") had a similar loading for each factor. Although the loadings were below the cutoff of 0.30 for the two items and there was a similar loading for item 12, each item was included into the respective components. In the factor analysis by Beck et al.,3 one item also had a loading below the cutoff of 0.30 and the total variance was 32%. However, at this time, it does not seem appropriate to recommend a test of the two-factor solution model. This is because the follow-up analyses by Beck et al.3 and the follow-up study by Pedrelli et al.4 did not show better results than the original, two-factor model, and also because one of the aims of the current study was to produce an adequate translation of the BCIS that corresponded to the developers' intent. The BICS-K also exhibited less than ideal internal consistency. The α coefficients were similar to the ones found in Beck et al.3 and Pedrelli et al.4's studies. Beck et al. reported that the coefficient α for the self-reflectiveness and self-certainty scores was 0.69 and 0.59, respectively, for the BCIS, and these levels of internal consistency were considered to be acceptable for research purposes,14 even though both coefficient α values were less than the 0.70 level recommended by Nunnally.15 Therefore, the BCIS-K was also considered to be acceptable for the same purposes. Consequently, these results suggest that patients with psychotic disorders respond on the BCIS-K measure in a manner similar to that for the BCIS, and that the Korean translated BCIS-K version of this study adequately maintained the concept of the original BCIS.
Another aim of the study was to assess whether the BCIS-K provides a measure of insight that is distinct from the more traditional measures. The self-reflectiveness score and the composite index score were moderately correlated with the KISP. None of the BCIS-K subscales or the composite index was correlated with the PSE-104. Comparing these results with those of the previous studies that used the BCIS, the self-reflectiveness score and composite index were significantly correlated with the Scale to Assess Unwareness of Mental Disorder of (SU-MD-A) mental disorder and the SUMD-A delusion, respectively, in an investigation of patients with psychotic disorders.3 In addition, these scores were significantly correlated with the Birchwood Insight Scale and there was also no significant correlation with self-certainty.4 In the current study, when we compared the level of correlation corresponding to the KISP and PSE-104, the difference in the level of correlation may be attributed to the fact that the PSE-104 places a relative emphasis on clinical phenomenon. Consequently, the results of this study are consistent with the concept of insight as being multidimensional and also with our hypothesis that these instruments measure insight, but they place relatively differing emphasis on different aspects of the insight.
In our study, the BCIS-K did not correlate with the GAF or CGI-S, suggesting the absence of any correlation between cognitive insight and the patients' overall condition. Using the BCIS, Pedrelli et al.4 addressed the relationship between the BCIS composite index and the Positive and Negative Syndrome Scale (PANSS)-rated symptoms of schizophrenia, and they reported a few significant correlations. Several studies that have investigated the relationship between insight and symptoms have reported conflicting data.16,17,18,19,20 There are several possible explanations for these inconsistent findings. Different symptom types exhibit different aspects of insight, possibly because the insight for positive symptoms such as hallucinations and delusions is fundamentally different from the insight for negative symptoms such as affective blunting. Mintz et al.21 also showed that the relationship between symptomatology and insight varied across the different dimensions of insight. Therefore, it is not clear whether the findings of the current study can be generalized to the relationship between insight and symptoms across the spectrum of psychotic disorders. There are also some limitations when considering the relationship between the cognitive insight and patient's symptoms because the GAF and CGI-S address the patient's overall level of function and the severity of their illness, which is unlike traditional measures such as the PANSS. Generally, the studies focused on insight and schizophrenia has used research-rated insight scales rather than self-report instruments. Perhaps this is because it is assumed that a self-report instrument is incapable of capturing the way people with schizophrenia understand and feel about their illness. Although a previous study about this issue validated using a self-report insight scale as compared with a research-rated insight scale,22 an instrument measuring a cognitive process, like the BCIS, needs to be developed.
The cognitive insight model is a prominent concept of insight that encompasses continuity between psychotic and nonpsychotic phenomena. Assessing cognitive insight is expected to add valuable information to the information that's traditionally collected in psychiatry. We translated the BCIS measure into Korean to support further research. The current study confirmed the acceptable reliability of the BCIS-K and its correlation with other psychiatric measures for a Korean sample of schizophrenic patients.
David AS. "To see oursels as others see us". Aubrey Lewis's insight. Br J Psychiatry 1999;175:210-216.
Amador XF, David AS. Insight and Psychosis. 2nd ed. New York: Oxford University Press; 2004.
Beck AT, Baruch E, Balter JM, Steer RA, Warman DM. A new instrument for measuring insight: the Beck Cognitive Insight Scale. Schizophr Res 2004;68:319-329.
Pedrelli P, McQuaid JR, Granholm E, Patterson TL, McClure F, Beck AT, et al. Measuring cognitive insight in middle-aged and older patients with psychotic disorders. Schizophr Res 2004;71:297-305.
Warman DM, Martin JM. Cognitive insight and delusion proneness: an investigation using the Beck Cognitive Insight Scale. Schizophr Res 2006;84:297-304.
American Psychiatric Association: Diagnostic and Statistical Manual of Mental Disorders. Fourth Edition, Text Revision. Washington. DC: American Psychiatric Association; 2000.
Wing JK, Cooper JE, Sartorious N. Measurement and Classification of Psychiatric Symptoms. Cambridge, UK: Cambridge University Press; 1974.
Yi JS, Chung MH, Lee BW, Lee SI. Reliability and Validity of the Korean Version of Global Assessment of Functioning Scale. Korean J Schizophr Res 2003;6:118-124.
Kim JH, Yim SJ, Yoon HJ. Assessing the Reliability and Validity of the Korean Version of the Revised Insight Scale for Psychosis (KISP). Korean J Schizophr Res 2004;7:15-22.
Markova IS, Roberts KH, Gallagher C, Boos H, McKenna PJ, Berrios GE. Assessment of insight in psychosis: a re-standardization of a new scale. Psychiatry Res 2003;119:81-88.
Guillemin F, Bombardier C, Beaton D. Cross-cultural adaptation of health-related quality of life measures: literature review and proposed guidelines. J Clin Epidemiol 1993;43:1417-1432.
Brislin RW. Back-translation for cross-cultural research. J Cross Cult Psychol 1970;1:185-216.
Marková IS, Berrios GE. The assessment of insight in clinical psychiatry: a new scale. Acta Psychiatr Scand 1992;86:159-164.
Cortina JM. What is coefficient alpha? An examination of theory and applications. J Appl Psychol 1993;78:98-104.
Nunnally JC. Psychometric Theory (2nd ed). New York: McGraw-Hill; 1978.
Amador XF, Strauss DH, Yale SA, Flaum M, Endicott J, Gorman JM. Assessment of insight in psychosis. Am J Psychiatry 1993;150:873-879.
Amador XF, Flaum M, Andreasen NC, Strauss DH, Yale SA., Clark SC, Gorman JM. Awareness of illness in schizophrenia and schizoaffective and mood disorders. Arch Gen Psychiatry 1994;51:826-836.
Schwartz RC. Insight and illness in chronic schizophrenia. Compr Psychiatry 1998;39:249-254.
Smith TE, Hull JW, Santos L. The relationship between symptoms and insight in schizophrenia: a longitudinal perspective. Schizophr Res 1998;33:63-67.
Smith TE, Hull JW, Israel LM, Willson DF. Insight, symptoms, and neurocognition in schizophrenia and schizoaffective disorder. Schizophr Bull 2000;26:193-200.
Mintz AR, Dobson KS, Romney DM. Insight in schizophrenia: a meta-analysis. Schizophr Res 2003;61:75-88.
Marks KA, Fastenau PS, Lysaker PH, Bond GR. Self-Appraisal of Illness Questionnaire (SAIQ): relationship to research-rated insight and neuropsychological function in schizophrenia. Schizophr Res 2000;45:203-211.