Cognitive Insight in Alzheimer’s Disease

Abstract

Background/rationale:

We investigated the cognitive insight profile of patients with Alzheimer’s disease (AD) using the Beck Cognitive Insight Scale (BCIS).

Methods:

This study involved 30 patients with probable AD and 15 healthy participants (ie, the controls). All individuals completed the BCIS, the Hamilton Rating Scale for Depression (HAMD), and the Hospital Anxiety and Depression Scale (HADS).

Results:

Mean scores of the HADS-depression subscale, HAMD, BCIS-self-reflectiveness (BCIS-R), and BCIS-self-certainty (BCIS-C) subscales were significantly different between the patients and the controls. However, there was no significant difference in BCIS reflectiveness–certainty index scores between the patients and the control groups. Regression analyses showed a moderately positive correlation between hallucinations and BCIS-C scores.

Conclusion:

This study is the first to investigate cognitive insight in patients with probable AD. The BCIS-R and BCIS-C scores were significantly lower in patients than in control group.

Introduction

Various authors have generally described insight as the knowledge of facts such as a patient’s awareness of having a mental illness, experiencing symptoms of that illness, awareness of needing treatment, and awareness that the illness is the cause of the symptoms. ^{1 -3} Clinical insight is defined as a state of awareness about whether one has a mental disorder. ^1,2 However, insight has more recently been described as a multidimensional construct with multiple factors. ² Cognitive insight is defined as a patient’s current capacity to evaluate his or her anomalous experiences and atypical interpretations of events. ⁴

It is difficult to discriminate between cognitive and clinical insight, but it can be concluded that clinical insight constitutes only a part of cognitive insight and is invaluable for the formulation and treatment of patients. ⁴ The Beck Cognitive Insight Scale (BCIS) ⁵ is a new scale for assessing insight. It is used with the goal of understanding how individuals understand their own judgment and reasoning processes. ^{5 -7} The insight evaluated by BCIS is cognitive insight.

It is well known that clinical insight is affected in patients with dementia. ⁸ Numerous studies have investigated the cognitive insight in individuals with psychotic disorders and other psychiatric disorders. ^{9 -12} However, how cognitive insight is affected in patients with probable Alzheimer’s disease (AD) has not been investigated previously. To know the cognitive profile of a patient with AD would be helpful for the clinicians while managing the patient’s medical therapy and while making rational suggestions for the patient’s daily life. In this study, we investigated the cognitive insight profile of patients with probable AD using the BCIS, and we evaluated the correlation of insight scales with depression and anxiety scales.

Methods

Participants

Participants were recruited from the Specified Outpatient Dementia Clinic at Faculty of Medicine, Pamukkale University (Denizli, Turkey). The patient group consisted of 30 patients who met the criteria for probable AD, and the control group consisted of 15 age- and gender-matched healthy participants. Diagnosis of AD was made according to the National Institute of Neurological and Communicative Disorders and Stroke and the Alzheimer’s Disease and Related Disorders Association (NINCDS-ADRDA) Alzheimer’s Criteria. ¹³ Exclusion criteria were (1) lack of informed consent of the participant; (2) dementia caused by other neurodegenerative diseases; (3) existence of treatable dementia causes; and (4) current substance abuse or the presence of any psychiatric disorders other than AD. The patients were randomly selected by disposing of many from the patient list provided by the Specified Outpatient Dementia Clinic. Demographic and clinical data were collected. Dementia severity was assessed by the Mini-Mental State Examination (MMSE) in which 21 to 27 points indicated mild cognitive impairment; 11 to 20 points indicated moderate cognitive impairment; and 10 points or less indicated severe cognitive impairment. The patient’s caregiver was questioned about whether the patient had hallucinations and/or delusions, and the caregiver’s answers were recorded. In addition, the answers of the questions that the patients could not understand were taken from their primary caregivers.

Control group was selected from age- and gender-matched voluntary, healthy caregivers of nondemented patients who visit the outpatient clinic frequently. Both written and oral informed consents were obtained from the patients and healthy participants before inclusion in the study.

Procedure

This study was conducted for the Student Scientific Study Module of the Faculty of Medicine, Pamukkale University, which is an employment of our faculty in which graduate students perform a scientific study under the supervision of an academic member. All measures were performed by an educated clinician with the students in a quiet, empty room containing a table, 5 chairs, and an examination table. We allowed only one caregiver in the study room.

Measures

Clinical Insight

A simple questionnaire about acknowledgment of having dementia was used for measuring clinical insight. The patients were asked about their illness, and the patients who acknowledged having dementia were classified as “patient with normal clinical insight.”

The BCIS

The BCIS is a 15-item self-report measure designed to assess cognitive insight. Participants rate on a scale from 0 points (do not agree at all) to 3 points (agree completely), the extent to which they agree with statements on the BCIS. The BCIS comprises 2 subscales: self-reflectiveness (BCIS-R that contains 9 items) and self-certainty (BCIS-C that contains 6 items). A composite reflectiveness–certainty (R-C) index score is obtained by subtracting the total score of the BCIS-C subscale from the total score of the BCIS-R subscale. The R-C index is considered a measure of cognitive insight. The higher the R-C index scores, the greater the cognitive insight. ⁵ We administered a validated Turkish version of the BCIS.

Hamilton Rating Scale for Depression

The Hamilton Rating Scale for Depression (HAMD) is widely used and is considered a gold standard for measuring depressive symptoms. The scale contains 17 variables. Each question has 3 to 5 possible responses in increasing severity. A score of 0 to 7 points is considered normal; a score of 8 to 13 points indicates mild depression; a score of 14 to 18 points indicates moderate depression; and a score of 19 or more points indicates severe depression. ¹⁴

The HADS

Hospital Anxiety and Depression Scale (HADS) was developed to identify possible anxiety disorders and depression among patients in nonpsychiatric hospital clinics. It is divided into an anxiety subscale (ie, HADS-A) and a depression subscale (ie, HADS-D), each of which contains 7 intermingled items. A score of 0 to 7 points is considered normal; 8 to 10 points indicates mild depression or anxiety; 11 to 14 points indicates moderate depression or anxiety; and 14 to 21 points indicate severe depression or anxiety. ¹⁵

Statistical Analyses

All statistical analyses were performed using SPSS 15.0 for Windows (SPSS Inc, Chicago, Illinois). The Mann-Whitney U test and chi-square test were used where appropriate. To determine the significance of correlations between continuous variables, Pearson’s correlation analyses were performed. We used a linear regression model for the regression analyses. The level of statistical significance was defined as P < .05.

Results

The patients were predominantly female (83.3%) with a mean age of 71.1 years (and a standard deviation (SD) of ± 8.9 years). The control group participants were also predominantly female (73.3%) with a mean age of 74.1 years (and an SD of ± 6.8 years). Gender distribution and mean age values were not significantly different between the patient and the control groups (the P values were .454 and .246, respectively). There was no significant difference in the education level and the socioeconomic status between the patient and the control groups (the P values were .507 and0.291, respectively; Table 1).

Table 1.

Demographics of the Patient and the Control Groups.

Demographics	Patients	Controls	P value
Gender (n, %)	Female (n = 25, 83.3%); Male (n = 5, 16.7%)	Female (n = 11, 73.3%); Male (n = 4, 26.7%)	.454
Age (years, mean ± SD)	71.1 ± 8.9	74.1 ± 6.8	.246
Education level	No literacy (n = 11, 6.6%); Primary school (n = 19, 63.4%)	No literacy (n = 6, 40.0%); Primary school (n = 9, 60.0%)	.507
Socioeconomic status	Low (n = 3, 10.0%); Moderate (n = 25, 83.3%); High (n = 2, 6.7%)	Low (n = 2, 13.3%); Moderate (n = 10, 66.7%); High (n = 3, 20.0%)	.291

Abbreviation: SD, standard deviation.

The clinical severity of dementia was mild in 8 (26.7%) patients, moderate in 14 (46.7%) patients, and severe in 8 (26.7%) patients. The mean disease duration was 3.17 ± 2.27 years (the minimum disease duration was 1.0 year and the maximum disease duration was 11.0 years). A total of 22 (73.3%) patients reported difficulties in finding directions, 13 (43.3%) patients reported hallucinations, and 12 (40.0%) patients reported delusions.

The mean scores of the HADS-D and HADS-A subscales, HAMD, BCIS-R, and BCIS-C subscales, and R-C index were compared between the patient and the control groups. The mean score of the HADS-D subscale was 9.4 ± 5.5 points in the patient group and 3.8 ± 4.9 points in the control group, and the difference was significant (P = .002). The mean score of the HADS-A subscale was 6.6 ± 4.7 points in the patient group and 5.0 ± 4.9 points in the control group, and the difference was insignificant (P = .287). The mean score of the HAMD was 11.3 ± 7.4 points in the patients and 6.2 ± 6.1 points in the control group, and the difference was significant (P = .027). The mean score of the BCIS-R subscale was 8.9 ± 3.5 points in the patient group and 12.4 ± 2.0 points in the control group, and the difference was significant (P =.001). The mean score of the BCIS-C subscale was 6.1 ± 2.5 points in the patient group and 8.5 ± 2.0 points in the control group, and the difference was significant (P = .003). The mean R-C index score was 2.8 ± 3.4 points in the patient group and 3.8 ± 1.8 points in the control group, and the difference was insignificant (P = .283). Table 2 shows the mean scores of the HADS-D and HADS-A subscales, HAMD, BCIS-R, and BCIS-C subscales, and R-C index in the 2 groups.

Table 2.

Mean Scores of HADS-Depression and -Anxiety, HAMD, BCIS-Self-Reflectiveness (BCIS-R), Self-Certainty (BCIS-C), and R-C Index in 2 groups.

Scales	Patients (mean ± SD)	Control (mean ± SD)	P value
HADS-depression	9.4 ± 5.5	3.8 ± 4.9	.002
HADS-anxiety	6.6 ± 4.7	5.0 ± 4.9	.287
HAMD	11.3 ± 7.4	6.2 ± 6.1	.027
BCIS-R	8.9 ± 3.5	12.4 ± 2.0	.001
BCIS-C	6.1 ± 2.5	8.5 ± 2.0	.003
BCIS-R-C index	2.8 ± 3.4	3.8 ± 1.8	.283

Abbreviation: BCIS, Beck Cognitive Insight Scale; HADS, Hospital Anxiety and Depression Scale; HAMD, Hamilton Rating Scale for Depression; R-C index, reflectiveness-certainty index; SD, standard deviation.

In the patient group, using BCIS scores in the regression analyses of clinical parameters—including clinical severity, difficulty in finding direction, existence of hallucinations and delusions, and loss of clinical insight—showed a significant and moderately positive correlation between hallucinations and the BCIS-C subscale score (P = .012; Table 3).

Table 3.

Results of Regression Analyses of Clinical Parameters With BCIS Scores.

Clinical Parameters	BCIS-R r and P values	BCIS-C r and P values	BCIS-R-C index r and P values
Clinical severity: 1 = mild (n = 8); 2 = moderate and severe (n = 22)	.091-.633	.091-.634	.165-.385
Difficulty in direction finding: 1 = yes (n = 22); 2 = no (n = 8)	.112-.555	.120-.527	.210-.266
Hallucinations: 1 = yes (n = 13); 2 = no (n = 17)	.160-.397	.455-.012*	.177-.349
Delusions: 1 = yes (n = 12); 2 = no (n = 18)	.162-.393	.117-.537	.081-.670
Loss of clinical insight: 1 = yes (n = 21); 2 = no (n = 9)	.068-.731	.234-.214	.249-.184

Abbreviations: BCIS, Beck Cognitive Insight Scale; BCIS-C, Beck Cognitive Insight Scale self-certainty; BCIS-R, Beck Cognitive Insight Scale self-reflectiveness; BCIS-R-C index, Beck Cognitive Insight Scale reflectiveness-certainty index. * = p < 0.05.

Correlation analyses were also performed between HADS-D and HADS-A scores, HAMD scores, with the BCIS scores (ie, BCIS-R, BCIS-C, and R-C index). There was no significant correlation between the HADS-D and HADS-A scores, HAMD scores, with the BCIS scores (Table 4).

Table 4.

Results of Correlation Analyses Between HADS-Depression and -Anxiety, HAMD and BCIS-Self-Reflectiveness (BCIS-R), Self-Certainty (BCIS-C), and R-C Index.

Scales	BCIS-R r and P values	BCIS-C r and P values	BCIS-R-C index r and P values
HADS-depression	.144 to .447	.234 to.214	−.099 to .601
HADS-anxiety	−.053 to .779	.177 to.349	−.196 to .300
HAMD	−.099 to .601	.180 to.342	−.336 to .070

Abbreviations: BCIS, Beck Cognitive Insight Scale; HADS, Hospital Anxiety and Depression Scale; HAMD, Hamilton Rating Scale for Depression; R-C index, reflectiveness-certainty index.

Discussion

The aim of the present study was to investigate the cognitive insight in patients with probable AD. We found that both BCIS-R and BCIS-C scores were significantly lower in the patient group than in the control group. In a recent study by Orfei et al, the self-reflectiveness score (as indicated by the BCIS-R subscale) was correlated with reduced gray matter volume in the right ventrolateral prefrontal cortex. ³ It is well known that metabolic and structural abnormalities of the ventrolateral prefrontal cortex can exist in patients with frontotemporal dementia and AD. ¹⁶ Self-reflectiveness (BCIS-R) is defined as a person’s ability to consider several types of information, perspectives, and alternative hypotheses simultaneously, which allows the person to come to a judgment about the self. This ability reflects verbal working, memory, and decision-making processes. ³ In light of this knowledge, our results gave objective data about the patients’ poor ability to consider the perceptions in probable AD.

In this study, we found that the BCIS subscale scores were significantly lower in the patient group than in the control group, but the R-C index was not significantly lower in the patient group. In various studies that were performed in patients with schizophrenia and healthy participants, the BCIS-R and BCIS-C subscale scores and the R-C index were not significantly different between the 2 groups. It is concluded that the cause of the differing results could be cultural differences in the way individuals understand questions on the scales and insufficient sample size in these studies. ^6,7 However, high scores on the BCIS-R subscale and low scores on the BCIS-C subscale are regarded as normal. ^6,17 In our study, both BCIS-R and BCIS-C subscale scores were lower in the patient group than in the control group. This explains why the R-C index was not significantly different between the patient group and the control group. We believe the global neurocognitive profile of the patients with AD would explain the similar R-C index in the patient and the control groups. It may be that the clinical view of cognitive insight in patients with probable AD is similar to healthy individuals because of low self-reflectiveness; however, assessing cognitive insight in large groups of patients with dementia by comparing them with healthy individuals would be helpful in understanding our results.

Buchy et al report that the BCIS-C subscale and R-C index scores were significantly correlated with verbal learning, memory, and attention in patients with first-episode psychosis. ¹⁸ The BCIS-R subscale score was also correlated with verbal memory. A high BCIS-C score has been reportedly correlated with reduced hippocampal volume, and a low R-C index score was correlated with reduced left hippocampal volume. ¹¹ In our study, the R-C index was not significantly different between the patient and the control groups. We believe neuroimaging studies in patients with different types of dementia, such as frontotemporal dementia, would give additional neuroanatomical correlates of the BCIS scores.

In our study, 70% of the patients had a loss of clinical insight. Administering the BCIS to these patients may be helpful in finding the patients with real misperceptions and biases. In this situation, having objective data about the insight of the patients would be helpful in managing the patients’ disease. In addition to find the patients with perception problems before existence of an apparent awareness of illness would be helpful while informing caregivers of the patients. It has been reported that including cognitive insight when evaluating patients may enhance the prediction of behavioral problems. ¹⁹ Therefore, we believe that using the BCIS can provide relevant information in identifying patients who are most vulnerable to having a lack of insight and can help develop effective cognitive therapeutic strategies.

The BCIS was developed to evaluate cognitive insight in patients with psychosis. Studies in which the BCIS was used to evaluate cognitive insight demonstrated that healthy people show better cognitive insight through higher self-reflectiveness and lower self-certainty compared to people with psychosis. ²⁰ In our study, we found that both BCIS-R and BCIS-C scores of the patients were higher in the patient group than in the control group. This result could be attributed to the different characteristics of the control groups used in various studies, since the control groups of most previous studies consisted of healthy young people. However, our control group consisted of older healthy participants.

In AD, neuropsychiatric symptoms such as delusions, agitation, depression, euphoria, disinhibition, irritability, aberrant motor behaviors, sleep, appetite, hallucinations, apathy, and anxiety are very common. Hallucinations, apathy, and anxiety are associated with global functional impairment. ²⁰ In this study, we performed correlation analyses of some neuropsychiatric symptoms, (including depression, anxiety, hallucination, and delusion) with cognitive insight; however, the correlation between clinical parameters and the BCIS scores only showed a positive correlation between hallucination and the BCIS-C scores. Insight is inversely proportional to the severity of illness and cognitive flexibility (which is also affected by psychopathology). Poor insight was associated with severe mental illness, particularly acute psychosis. ²¹ Therefore, this result in our patients may explain the correlation of insight with psychotic symptoms, rather than depression. Warman and Martin report that healthy people who are higher in delusion proneness evidenced more certainty in their beliefs and judgment than did people who were lower in delusion proneness, and healthy people who were higher in delusion proneness were more open to external feedback and more willing to acknowledge fallibility than people who were lower in delusion proneness. ⁷ In our study, we did not find any significant correlation between delusion and the BCIS scores. This result could be attributed to the profile of the participants we used in our study. Warman et al used undergraduate students with no history of psychotic disorder; in contrast, the patient group in our study consisted of older patients with probable AD. Furthermore, as a general point of view, the BCIS does not exclusively assess judgment in relation to delusional beliefs. Items are constructed to determine how individuals assess their judgments in general, not specifically to unusual beliefs. However, evaluating the relationship of delusions with cognitive insight by specific and objective scales of delusion in patients would give additional results.

In a study by Engh et al, delusions in patients with schizophrenia were associated with low self-reflectiveness and high self-certainty, irrespective of the presence or absence of hallucinations. ²² In addition, Engh et al stated that solitary hallucinations may also be associated with high cognitive insight, which is contrary to our results. The small patient sample and the lack of a more detailed evaluation of hallucinations and delusions may have caused the incompatible results. It is difficult to find an exact explanation for our results, but further studies on cognitive insight in patients with dementia would give additional information.

The BCIS-R subscale score was linked to heightened levels of emotional discomfort symptoms in patients with schizophrenia. ²³ However, studies that report contradictory results associated between depression and cognitive insight and between anxiety and cognitive insight exist. ^12,24 We did not find any significant correlation between the scores of HADS-D and HADS-A subscales, HAMD, and BCIS. This result could be attributed to the characteristics of our patient group, which consisted of older patients with dementia whose capability of self-expression about emotional states would have affected the depression and anxiety scales.

Insight among patients with dementia does not correlate with cognitive impairments, and the mechanism for loss of insight is reported to be anosognosia or inability to recognize one’s own illness or deficits. ²⁵ In patients with frontotemporal dementia, lack of concern for proper self-appraisal was defined as “Frontal anosodiaphoria,” and there is evidence that the ventromedial prefrontal cortex, especially on the right, and its adjacent connections are involved in this process. ^26,27 However, these studies focused on the clinical insight, and there is no knowledge in the literature about the cognitive insight and its anatomical correlates. It is well known that specific brain areas are more affected in AD, so results of our study may throw a new light on this area.

Clinical insight is only a part of cognitive insight, and it is invaluable for the formulation of treatment plan and medical treatment of patients. ⁴ Therefore, it is difficult to claim that patients with normal clinical insight have enough insight about themselves, their beliefs, and their behaviors. With this point of view, to evaluate a patient’s insight in detail would give more information about treatment options and suggestions for daily life.

Psychiatric patients with higher cognitive insight are more likely to be living independently, ¹² but how cognitive insight affects patients with dementia is still unknown. All studies evaluating cognitive insight have involved patients with psychiatric diseases. The present study is the first study to investigate the cognitive insight of the patients with probable AD. Furthermore, neuroimaging studies evaluating different dementia types and studies evaluating the association of patients’ cognitive insight with treatment and quality of life would add very comprehensive knowledge to the literature.

In conclusion, compatible with the hypothesis of our study, BCIS-R and BCIS-C scores were significantly lower than the controls in patients with AD. Our results emphasize the fact that knowing the cognitive profile of a patient with AD would be helpful for the clinicians while managing the patient’s medical therapy and while making rational suggestions for the patient’s daily life.

Limitations

This study was performed using a small sample consisting of 30 patients who met the criteria for probable AD. Further studies using large patient groups would provide more informative knowledge.