Introduction

Lack of insight or the lack of awareness of one’s mental health problems affects 30 to 50% of patients with a schizophrenia spectrum disorder [1] and is often debilitating [2, 3]. Lack of insight has been defined as a continuous and multidimensional construct that includes the following aspects: (1) awareness of having a mental illness, (2) an understanding of the need for treatment, (3) awareness of the social consequences of mental disorders, (4) awareness of symptoms, and (5) attribution of symptoms to a mental disorder [4]. Research has shown that the relationship between the lack of awareness and the various aspects of a serious and persistent mental disorder is complicated and affected by both aspects of the illness (e.g., symptom severity) and psychosocial factors (e.g., years of education, age, etc.). Lack of insight has been postulated to impair functioning [4,5,6,7], quality of life [8], and increases the resistance and lack of compliance to treatment (particularly psychotropic medications; [8,9,10]. Lack of compliance with treatment is often followed by relapse and re-hospitalization [11, 12].

Lack of insight of the behaviors and symptoms associated with a psychiatric disorder has been found to have a significant effect on clinical presentations. One study showed that the presence of a high level of insight has been associated with impaired functioning, perception of a lower quality of life, hopelessness, and depression [13]. It is possible that having an awareness of having a serious mental illness results in resignation to a lower standard of living and the giving up of one’s goals and aspirations, which may cause or exacerbate depression. Depression in schizophrenia is a serious clinical problem as depression increases the risk of suicide. Suicide rates in people living with schizophrenia are eight times higher than in the general population, and 20–40% of people with schizophrenia attempt suicide [14].

The assessment of insight has been a subjective aspect of mental status examinations, based on a clinical interview of undefined depth and intensity. To place more rigor on the assessment of insight, Amador and colleagues developed the Scale to Assess Unawareness of a Mental Disorder (SUMD; [15]). Some evidence shows that patients with bipolar disorder did not differ significantly in the level of insight when compared to patients with schizophrenia [2, 16] and that insight deficits were worse in psychotic and bipolar disorders compared to those with a depressive disorder [17]. The results from the Pini, et al. [16] study showed that the assessed psychosocial functioning of a patient using the Global Assessment of Functioning scale (GAF) had a direct and significant effect on the level of insight. Thus, those with a psychotic and bipolar disorder are likely to have worse insight and overall functioning compared to those with a depressive disorder.

Because of its length (i.e., 17 items), however, the SUMD is seldom utilized in clinical settings and practice [2]. A shorter version of the SUMD (i.e., 9 items) that can be done within the time constraints of clinical settings was developed by the same group of authors. The abbreviated version, the SUMD-A, has been shown to be a valid and reliable instrument to assess insight in patients with schizophrenia [1]. This instrument has been used by others to explore insight and medication adherence [8] and in a large clinical study by the developers [2]. The authors concluded that the SUMD-A can be used in a clinical setting for reliable and valid assessment of insight.

A review of the literature has shown that the research in which the SUMD has been used has excluded participants whose psychotic symptoms were the result of intoxication related to a substance use disorder, have a diagnosed organic disorder, or neurodevelopmental disorder; some studies excluded all individuals not meeting criteria for a type of schizophrenia (i.e., paranoid type; [13]). In addition, the different samples used have been individuals that have been residing in the community for an unspecified duration [13, 18], recently discharged from the hospital [19], in remission [20], at-risk outpatients [21], attending a day treatment program [8], or a mix of outpatients and individuals admitted to a public university teaching hospital [22]. Studies that utilized a sample that reflects a “standard” urban, public psychiatric facility where the individuals were assessed shortly after being admitted are lacking. Pini, et al. [16] did use consecutively hospitalized patients but the assessments occurred one week prior to discharge when the patients were stable in their symptoms and medications.

Studies have shown that there is a higher prevalence of substance abuse among persons who are homeless and have a mental illness [23,24,25]. Co-occurring disorders can be difficult to diagnose due to the complexity and severity of symptoms. One report cited that two million people with mental illness are booked into jails each year, and that nearly 15% of male and 30% of female inmates recently admitted to jail have a serious mental illness [26]. The high prevalence of substance use and homelessness in a psychiatric population may interact with lack of insight.

There are several aims to the study. First, we wanted to examine the relationship between diagnosis, symptom severity, and insight on variables relevant for this clinical population, such as substance use and homelessness. We hypothesized that insight and symptom severity would have significant associations with many of these clinical variables, and that substance use and homelessness would be associated with worse symptom severity and insight. Second, we aimed to look at the unique impact of diagnosis on clinical variables, including insight, symptom severity, depression, and quality of life. We hypothesized that individuals with a psychotic disorder would show greater problems on all measures compared to those with a depressive or bipolar disorder after controlling for demographic variables. Finally, we hypothesized that insight would act as a mediator between having a psychotic disorder and symptom severity.

Method

Participants and Setting

The data collection took place at a state psychiatric facility located in a major metropolitan city in the southwest area of the country. The city in which the hospital is located is culturally diverse and one of the most diverse cities in the nation. The hospital has approximately 120 beds. The average length of stay is less than 21 days. Patients were all transferred to the hospital from an emergency room. All consecutive admissions on one 40 bed unit consisting of two interdisciplinary teams were assessed. All participants were 18 years of age or older.

Patients were assessed within the first three to five days after being admitted. Exclusionary criteria were limited to exclude persons with severe decompensated neurodevelopmental disorders (i.e., autism spectrum disorder with intellectual impairment) and moderate to severe intellectual impairments. Patients were excluded if the person refused to participate or complete the assessment. All participants provided written informed consent upon admission. All clinical assessments were performed in routine practice during treatment team meetings and results were reported back to the members of the team as part of the routine treatment team review that was done for each working day (Monday through Friday).

Upon admission into the unit, each participant’s chart was reviewed for admitting diagnoses and demographics were collected if available. Demographic information that was uncertain or unavailable during the chart review was obtained during the initial interview after admission or through collateral information from family members. One of the two authors attended the initial treatment review and interview by the interdisciplinary team of each new admission. The interdisciplinary team consisted of a board-certified psychiatrist, a licensed psychologist, two licensed social workers and at least one registered nurse.

Measures

Participants were asked about socio-demographic information including ethnicity, living arrangement, education, existing legal problems, and history and type of incarceration. Information was gathered by interviewing the participant and available medical records. Participants were asked clarifying questions regarding clinical characteristics if needed, including substance use history and awareness of substance use when admitted into the hospital. DSM 5 diagnoses were obtained from the admitting paperwork.

Clinical Global Impressions Scale (CGI)

The Clinical Global Impressions Scale (CGI; [27]), is a clinician-rated measurement of overall dysfunction due to illness. The measurement uses a seven-point scale (1 = normal, 7 = extremely dysfunctional). It has been shown to have good convergent validity with other assessments of functioning. Participants were assessed in terms of functioning of several domains: cognitive status, perceptual experiences, mood, and social functioning.

Calgary Depression Scale for Schizophrenia (CDSS)

The Calgary Depression Scale for Schizophrenia (CDSS; [28]) is a nine-item scale specifically designed for patients with schizophrenia that evaluates depression independently of extra-pyramidal and negative symptoms. It has shown good predictive validity in evaluating depression in individuals with schizophrenia compared to other instruments focused on depressive disorders.

World Health Organization Quality of Life (WHOQoL-BREF)

The World Health Organization Quality of Life (WHOQoL-BREF; [29]) is a widely used, 26 item self-report measure that assesses aspects of daily life in four domains: physical health (activities of daily living, dependence on medical treatment, energy and fatigue, mobility, pain, sleep and work capacity); psychological health (body image and appearance, negative and positive feelings, self-esteem, spirituality, and concentration); social relationships (personal relationships, social support, sexual activity); and environment (finances, physical safety, access to health services, home environment, opportunities to acquire new information, leisure activities, physical environment, and transport). In addition to the domain scores, an overall score also is calculated that includes questions regarding self-perceptions of overall well-being and health. The WHOQoL-BREF has been used in assessing perceived quality of life in patients at risk in developing a serious mental illness [21].

Subjective Unawareness of a Mental Disorder- Abbreviated (SUMD-A)

The Subjective Unawareness of a Mental Disorder- Abbreviated (SUMD-A) is a standardized nine item scale based on a patient interview. The SUMD-A assesses current awareness of the following states: a) having a mental disorder, b) consequences of a mental disorder, c) the effects of psychotropic drugs, d) hallucinatory experiences, e) delusional ideas, f) disorganized thinking, g) blunted affect, h) anhedonia, and i) lack of sociability. Higher scores represent less insight. SUMD-A data was obtained during the initial interview or during another interview within 5 days after admission to the unit.

Results

Sample demographics are presented in Table 1. Mean age was 37.05 years. Most participants were male (56.8%) and nearly half were Caucasian (49.2%); over half of the participants were never married (66.8%) and nearly one third (31.6%) had some education after graduating from high school. Forty-five percent were homeless, compared to a report that 33 % of homeless individuals had a serious mental illness [30]. In addition, nearly two thirds either had past or present substance abuse (65.8%) and more than half (56.6%) had a positive drug screen for some type of illegal substance (including cannabis) when admitted. More than half (54.8%) had a history of being incarcerated; of those reported having been incarcerated, 71% reported having only been in jail, and not prison.

Table 1 Sample demographics (N = 199)
Full size table

Diagnoses were coded into three general categories listed in the DSM 5. The groupings that were included were Schizophrenia Spectrum and Other Psychotic Disorders, Depressive Disorders, and Bipolar and Related Disorders.

Correlations between the sociodemographic data and clinical variables are presented in Table 2. Table 3 presents the correlations between the clinical characteristics, and Table 4 presents the correlations between demographic variables and clinical variables.

Table 2 Correlation matrix for examined variables
Full size table
Table 3 Correlation matrix for clinical measurements
Full size table
Table 4 Correlation matrix for clinical measurements with demographics
Full size table

Correlation analyses show relationships between the psychosis category and non-Caucasian race (r = .29, p < .01), less education (r = .15, p < .01), and younger age (r = −.14, p < .05). Having a diagnosis within the psychotic category was positively correlated with all domains of subjective well-being and life satisfaction (rs ranging from .20 to .38, ps < .05), and less awareness of having taken an illegal substance on either the current or a prior admission (r = .21, p < .05). Having a psychotic disorder was associated with higher total scores on the SUMD-A (i.e., less insight; r = .47, p < .01) and the CGI (i.e., more severe impact on functioning; r = .40, p < .01) and negatively associated with the CDSS (i.e., less depression; r = −.28, p < .01).

Having a diagnosis of Bipolar disorder was associated with less subjective satisfaction in three domains on the WHOQoL: physical health (r = −.24, p < .01), psychological health (r = −.21, p < .05), and social relationships (r = −.18, p < .01). A bipolar diagnosis was also associated with more insight (r = −.52, p < .01) and less severe symptoms (r = −.14, p < .05). A depressive diagnosis was associated with worse psychological health (r = −.26, p < .01), greater insight (r = −.34, p < .01), and less severe symptom severity (r = −.30, p < .05).

Having a depressive disorder was associated with lower quality of psychological health (r = −.26, p < .01), greater awareness of symptoms (r = −.34, p < .05), and less symptom severity (r = −.30, p < .01). Current depressive symptoms were associated with older age (r = .19, p < .01), greater insight (r = −.21, p < .01), and with lower quality of life in all four domains, as well as the total score (rs ranging from −.24 to −.54, ps < .01).

Being homeless was associated with being older (r = −.16, p < .05) and with non-Caucasian race (r = .15, p < .05). Having a positive urine drug screen for illegal substances upon being admitted was negatively associated with poorer medication compliance (r = −.18, p < .05). Being unaware of current or past substance use was associated with less awareness of one’s mental health problems (r = .36, p < .05), greater symptoms severity (r = .23, p < .01), and decreased depressive symptoms (r = −.18, p < .05). Having a history of any type of incarceration was only positively correlated with symptoms severity (r = .17, p < .05). Individuals who reported not being homeless (i.e., living with family, in a managed care facility, or independently by oneself) were more likely to report greater psychological (r = .24, p < .05) and environmental health (r = .21, p < .05), and overall quality of life (r = .22, p < .05).

Correlations between the clinical measures show positive correlations among the scores of all the domains of the WHOQoL (rs from .58 to .71, ps < .01). Domain 2 of the WHOQoL (Psychological Health) was correlated with less insight on the SUMD-A (r = .20, p < .01). Less insight was strongly correlated with overall symptom severity (r = .59, p < .01) and negatively correlated with depression symptoms (r = −.21, p < .01). Medication compliance was also associated with fewer depression symptoms (r = −.15, p < .01).

To test for the effect of diagnosis, a MANCOVA was conducted. The three diagnostic groups (i.e., psychotic disorders, bipolar disorders, and depressive disorders) were specified as independent variables in the model. Symptom severity, awareness of symptoms, depression, and quality of life were specified as dependent variables. Education level, race, marital status, living situation, age, gender, and history of incarceration were specified as covariates. Results revealed a significant effect of diagnosis on symptom severity (F(2, 145) = 5.05, p < .01), awareness of symptoms (F(2, 145) = 16.43, p < .001), and depression (F(2, 145) = 6.26, p < .01). Diagnosis did not have an effect on quality of life (F(2, 145) = .89, p = .41).

Separate ANOVAs with planned contrasts were conducted to follow up the results of the MANCOVA. Diagnosis was identified as the independent variable in all analyses. Contrasts were conducted using a Bonferroni correction. For symptom severity, contrasts revealed a significant difference between psychotic disorders (M = 5.08) and both bipolar disorders (M = 4.72, p < .01) and depressive disorders (M = 4.34, p < .001). Results showed that individuals with psychotic symptoms showed more symptom severity compared to individuals with depressive and mood disorders.

Contrasts of the awareness analysis revealed a significant difference between psychotic disorders (M = 13.18) and both bipolar disorders (M = 7.46, p < .001) and depressive disorders (M = 7.07, p < .001). Results showed that patients with psychotic disorders were rated as being less aware of their symptoms and their effects than patients with either bipolar disorders or depressive disorders. Contrasts of the depressive symptoms analysis revealed a significant difference between depressive disorders (M = 9.57) and psychotic disorders (M = 4.71, p < .001). Results showed that individuals with depressive and mood disorders were more likely to report symptoms associated with depression than by patients with psychotic disorders.

Awareness of psychiatric symptoms was tested as a mediator between having a psychotic disorder and symptom severity. Separate correlations between the three variables were found to be significant (see Table 2), satisfying the first three conditions of a mediator as defined by Baron and Kenny [31]. Hierarchical multiple regression analysis was used to test the strength of the relationship between having a psychotic disorder and symptom severity when in the presence of awareness. In the first step, having a psychotic disorder was associated with greater symptom severity (beta = .47, p < .001). In the second step, awareness was added a predictor to the model. When awareness was entered into the model, having a psychotic disorder was no longer a significant predictor (beta = .08, p = .18). Awareness was significantly associated with symptom severity in the model (beta = .55, p < .001), indicating full mediation.

Discussion

This study sought to explore the link between diagnosis, insight into one’s own mental health symptoms, and symptom severity in an urban inpatient psychiatric setting.

Patients who had less awareness of their own psychiatric symptoms were more likely to have worse overall symptom severity, regardless of diagnosis. Symptom unawareness was also related to being less aware of having taken illegal substances either in a previous or current admission. However, having a history of substance abuse was not associated with symptom severity, indicating that substance use is not a driving force behind unawareness and severe mental illness.

While having little insight was associated with less awareness of having taken substances (either in a past or present admission), it was not associated with substance use itself. Additionally, substance use was only associated with a history of incarceration, having less education, and being white. The lack of connection between substance use and both insight and symptom severity suggests that substance use is related more to demographic factors in this population rather than psychopathology. The lack of connection between substance use and both insight and symptom severity indicate that substance use is not a driving force behind why severely mentally ill patients struggle with insight and debilitating symptoms.

When insight and symptom severity were examined by diagnosis, patients with any psychotic disorder exhibited greater unawareness of their illness and greater symptom severity compared to patients with non-psychotic disorders after controlling for age, gender, race, substance use history, education, and homelessness. This is in line with previous research that found individuals with schizophrenia had worse insight into their symptoms than those with bipolar disorder [32]. This finding speaks to the unique and debilitating impact of psychotic symptoms. Complex psychosocial variables can greatly hinder recovery and stability for a variety of reasons, but successful treatment of psychotic symptoms and fostering insight may be a crucial first step in addressing issues of substance use and homelessness in this population.

Awareness was found to fully mediate the relationship between having a psychotic disorder and symptom severity. Individuals with psychotic disorders may be more prone to having their symptoms debilitate their functioning as they may attribute their lack of functioning to other sources (e.g., delusions, hallucinatory content). Emphasis on psychoeducation may be especially important for those who have a psychotic disorder.

Quality of life was negatively associated with depressive symptoms. After controlling for psychosocial variables, quality of life was not associated with insight, general symptom severity, or diagnosis. It is possible that individuals with psychotic disorders are so out of touch with reality (as indicated by their high unawareness of symptoms and general symptom severity) that they may not perceive their quality of life as being diminished. Additionally, our population was comprised of individuals often receiving treatment against their will. A population of patients with psychotic disorders that have shown treatment adherence and stability may show differences in perspective of quality of life. With regard to patients with bipolar disorder, it is likely that individuals in an active manic phase may also overinflate their quality of life compared to those in a depressive state.

Quality of life was negatively associated with depressive symptoms, but those with a depressive disorder did not overall endorse a worse quality of life after controlling for demographic variables. The lack of connection between quality of life and having a depressive disorder implies that other factors (e.g., homelessness, substance use) may better explain the link between quality of life and depression in an inpatient population. The basic needs of this sample were complex, and stress from a variety of areas may be reducing quality of life significantly more than current depression.

Depression is often seen in schizophrenia. However, individuals with a psychotic disorder in our sample were less likely to report depressive symptoms than those with other diagnoses. It is possible that individuals with a psychotic disorder have fewer depressive symptoms when they are in an episode of care that requires inpatient hospitalization. This is congruent with research that found that greater insight increases depression, self-stigma, hopelessness [33] and suicidality [34] in those with a schizophrenia diagnosis. Those with a psychotic disorder who are in an inpatient episode of care should be given skills to manage depression and safety planning that are necessary after insight increases.

These results have several important implications. Patient education may be an important part of recovery, as increasing insight into mental health symptoms may reduce overall symptom severity, regardless of diagnosis and specific symptoms. Patients who are unaware of their diagnosis and symptoms will likely go longer without treatment and disagree about needing treatment in the first place. Some patients may not have insight into when symptoms get worse, and may not get treatment until well into an episode (e.g., not recognizing early signs of mania or depression). Education of signs and symptoms of how severe episodes start may decrease overall symptom severity and reduce the number of hospitalizations. However, it is important to note that greater insight may result in greater depression and thoughts of suicide in those with a psychotic disorder, and so coping skills for depression are crucial during an inpatient episode of care for this population.

This study also suggests a diminished role of substances and insight. Despite their mind-altering effects and interactions with psychopathology, this sample did not show a significant association between substances and either insight or symptom severity. However, there were significant correlations with being white, homeless, and having been incarcerated. While substance abuse should remain an important target of treatment, substance use seems to be tied in with systemic and demographic problems that are often beyond the scope of treatment for an inpatient psychiatric hospital. This suggests that referrals to quality resources, especially around homelessness and being an ex-offender, are crucial for discharge planning.

There were several limitations to this study. This study used cross-sectional data, and so we cannot say how insight changed over time. Participants were interviewed quickly after admission due to the high census turnaround, so participants may have given different responses if they were interviewed closer to discharge when they have regained some stability and had a satisfactory discharge plan.