Cognitive insight and quality of life among psychiatric outpatients

Singapore is a multiethnic country in Southeast Asia with a population of 5.6 million; Chinese encompassed 74.3% of the resident population, followed by Malays (13.4%), Indians (9.0%), and other ethnic groups (3.2%) [17].

The current study utilized data from a larger cross-sectional study examining the pathways to care and illness perception among psychiatric outpatients (N = 400) seeking treatment at the Institute of Mental Health (IMH) or its satellite clinics between October 2015 and December 2016 (Refer to Jeyagurunathan et al. [18] for details on the main study). Participants were included in the study if they were between 21 and 65 years old, capable of providing written consent, able to read and understand English, Chinese, Malay or Tamil, had a clinical diagnosis of schizophrenia and related psychosis, any mood disorder (depression or bipolar disorder), or any anxiety disorder (generalized anxiety disorder, obsessive compulsive disorder, post-traumatic stress disorder or panic disorder) and had an illness duration of not more than 2 years as determined by a psychiatrist using International Classification of Diseases 9th Revision (ICD-9R) criteria. Following written informed consent, participants completed a set of questionnaires, after which they were reimbursed for their time. Individuals who had difficulty concentrating or reading had a research assistant read out the questions to them and clarify their queries. The study received ethics approval from the National Healthcare Group Domain Specific Review Board.

Table 1 shows the socio-demographic distribution of the sample. Of the total sample (N = 402), data from 400 patients were used for analysis; two were withdrawn due to the total duration of mental illness being more than 2 years. Of the 400 outpatients, 30% had anxiety disorders (n = 120), 35% had mood disorders (n = 140), and 35% had schizophrenia and related psychosis (n = 140). Majority of the sample was male (52.8%), of Chinese ethnicity (62.8%), had ‘A’ level/Poly/Other Diploma (i.e., qualifications obtained at a post-secondary/pre-university level) (39.0%), were employed (58.3%), and single (64.8%). The mean age of the sample was 32.66 years (SD = 11.38). The mean BCIS score obtained in this sample was 5.61 (range:-10 to 23).

A one-way ANOVA was conducted to examine CI and its subscales across diagnostic groups (Table 2). There was a statistically significant difference in SR scores across the diagnostic groups; F (2, 397) = 3.96, p = .020. Post-hoc comparisons indicated individuals with schizophrenia and related psychosis (M = 11.80, SD = 5.04) to have significantly lower SR mean scores than those with anxiety disorders (M = 13.93, SD = 8.56). No significant differences were found for SC scores and overall CI scores with respect to diagnosis.

Multiple linear regression analyses were conducted to examine the socio-demographic correlates of CI and its subscales (Table 3). With regards to SR scores, those with anxiety disorders were more likely to have higher SR scores than those with schizophrenia and related psychosis (β = 1.827, p = .003); whereas those between the age 50–65 years (β = − 2.288, p = .011) were less likely to report higher SR scores compared to those between 21 and 34 years old. Only educational level was significantly associated with SC scores. In particular, those with no formal education/Primary school leaving examination (PSLE) (β = 2.025, p = .023) were more likely to report higher SC scores than those with A level/polytechnic diploma. In relation to overall CI scores, those between the age of 50–65 years (β = − 2.368, p = .020) were less likely to have higher overall CI scores than those between 21 and 34 years old. In addition, those of Malay ethnicity (β = − 1.412, p = .039) tended to report lower CI scores than the Chinese.

A significant positive correlation was only found between overall CI scores and the environmental domain of QoL (r = .153, p < .01). Multiple linear regression analyses after adjusting for socio-demographic variables and diagnostic groups found higher CI scores to be significantly associated with higher QoL in the environmental domain (β = .480, p = .003) (Table 4). With regards to the subscales, after controlling for socio-demographic and diagnostic correlates, higher SR scores were associated with higher QoL in the environment domain (β = .404, p = .030) whereas, higher SC scores were associated with lower scores on the social relationships domain of QoL (β = −.577, p = .045).

Current findings, however, should be considered in view of its limitations. Firstly, the cross-sectional nature of the study does not allow for causal inferences between CI and QoL. Secondly, individuals with bipolar disorder and depression were included under the broader diagnostic group of mood disorders. Past studies have not only shown CI to vary between individuals with bipolar disorders and psychosis but also noted differences among individuals with bipolar disorders, such that bipolar patients with a more recent episode of mania have less CI than those with bipolar disorder with a most recent episode of depression [11]. Not making this distinction might have masked some essential differences in CI across the diagnostic groups. Thirdly, conceptions of mental illness, its symptomatology and insight are likely to be strongly influenced by cultural interpretations [27]. Kao et al. [13] for instance, found the Taiwanese version of the BCIS to be similar in factor structure to the original BCIS but the items in the Taiwanese version to be distributed differently from the original scale following a principal components analysis (PCA). Hence, it is possible that cultural differences in endorsement of items may exist and this scale might have to be validated in the current context. In the current study, CI was found to differ across ethnicities; while this was partly attributed to differences in the levels of prejudice and understandings of mental illness found among the local population as reported in past studies [25], these differences were not examined in this study. Measuring the level of stigma among the different ethnicities might be important given that past studies such as Lysaker et al. [26] have found insight (clinical insight) to interact with internalized stigma to predict levels of hope, self-esteem, as well as functioning.

Despite some supporting evidence for the link between CI and QoL in this study, it is likely that this relationship is moderated by other variables. Past studies have suggested an interaction between symptoms and the level of insight to predict QoL, particularly among individuals with schizophrenia [16, 28, 29]. For instance, while Phalen et al. [6] found SR to have a direct positive association with QoL, the effect of SC on QoL was moderated by symptom severity whereby greater SC had a positive impact on QoL when the severity of symptoms was high but a negative impact on QoL when symptom severity was low. Similarly, Kim et al. [10] found that while higher levels of CI was associated with lower subjective QoL, this relationship was attenuated after controlling for symptoms. Yet other studies have suggested a positive correlation between delusional thinking and SC scores [16, 28, 30, 31]. While the lack of difference in SC and CI scores in our study population could be reflective of a relatively stable sample in terms of symptomatology, using a standardized measure to measure symptomatology and its severity would have helped to clarify if the absence of significant findings in terms of SC and CI across the diagnostic groups was due to the low levels or severity of symptoms.

In addition, studies have suggested clinical insight as a moderator between CI and QoL [4, 6, 32]. It is likely that an individual’s level of awareness of their symptoms and illness (i.e., clinical insight) will impact the extent to which one considers it necessary to observe their thought processes with regards to their illness or alter their beliefs (i.e., cognitive insight) [4, 14]. Also, individuals with high clinical insight have been shown to be more compliant with medication [29, 33, 34] which may in turn result in them being more able to manage their symptoms. This could subsequently promote greater CI (i.e., more willing to alter beliefs and receive feedback) which could then result in better functioning.

Only information regarding the primary diagnosis for the patients was collected in the current study. Given that the presence of co-morbid diagnoses is likely to have a greater detrimental effect on QoL compared to those without co-morbid diagnoses [35], not including this information on co-morbid diagnoses represents a potential limitation. Lastly, studies examining insight in relation to QoL have used both objective [16, 26] and subjective measures of QoL [10]; while observer-rated quality of life scales may be more adept in capturing functional capacity and illness severity, self-report scales are likely to incorporate affective states and subjective level distress [16]. As such, it would have been better to incorporate an objective measure of QoL in addition to the subjective QoL scale used in the current study for a better estimation of quality of life outcomes [16].