Abstract
The past decade has seen a renewal of interest in the concept of insight in psychosis [1]. Lower levels of insight in schizophrenia have been associated with poorer long-term outcome [2–4], worse executive function deficits [5–9] and more persistent positive symptoms [10–13]. Psychotic patients with poor awareness of having a mental illness also show poor compliance with both pharmacological [2],[14–17] and psychosocial [18] treatments.
The explanation for these interrelationships is likely to be complex, with multiple causal interrelationships. For example, poor compliance, as well as being a component of poor insight [15], may also result in inadequately treated positive symptoms that will further hinder the achievement of good insight.
Poor insight has also been related to increased risk of violence [19], [20]. It has even been hypothesized that the increased risk of interpersonal violence in patients with schizophrenia [21] is largely due to a subgroup characterized by poor insight, poor compliance, neuropsychological deficits and substance abuse [22]. The HCR-20 [23], a well validated assessment tool for assessing the risk of violence in psychiatric patients, includes a measure of insight as one of the ‘clinical’ risk factors for future violence. However, a recent prospective study of patients with schizophrenia who were in remission found that poor insight failed to predict future violence [24]. There has thus far been a paucity of research examining insight in patients with psychotic illnesses who have previously committed serious criminal offences.
Another strand of research links higher levels of insight in schizophrenia with an increased risk of lowered mood [25–27] and even suicide [28], [29]. There are various explanations given for these findings, the simplest being that poor insight or denial of illness serves a protective function with respect to self-esteem [30]. Most previous studies of this issue have used global measures of depression. However, it seems more likely that any relationship with increased insight would be with cognitive symptoms of the depressive syndrome, such as loss of optimism for the future (hopelessness) due to awareness of having a chronic disabling illness, rather than with biological symptoms of depression.
The practical value of investigating poor insight as a separate variable is that it may be a potential target for specific interventions, with consequent improvement in global outcome [31–33]. However, improvement in insight does not always result in overall clinical gains [34].
Various scales have been developed to quantify insight [2],[35–37]. It is clear that insight is not a dichotomous variable, but that it can be subdivided into several domains. These include the awareness of having a mental illness, compliance with treatment and the ability to relabel psychotic symptoms as pathological [15].
The current study quantified the level of insight in a group of patients with schizophrenia who were, or had previously been, subject to custody in a high security forensic hospital. This allowed comparison with data from non-forensic populations, and investigation of the relationship of insight to symptoms and to a history of violence.
Another aim was to examine the relationships between specific domains of insight and hopelessness. It was hypothesized a priori that awareness of illness and compliance would correlate with higher levels of hopelessness. This is because both awareness of illness and compliance may relate to the acceptance by the subject that they are suffering from a chronic mental illness in need of medication, which may in turn impact adversely on hopes for the future. The ability to relabel symptoms on the other hand, may confer a sense of control over the distressing experience of psychosis, and so was not expected to have such a relationship.
Methods
Sample
Participants were recruited from the inpatient rehabilitation wards and the outpatient service of the Victorian Institute of Forensic Mental Health (VIFMH). This is a publicly funded statewide service providing forensic mental health services for the Australian state of Victoria.
The inclusion criterion for the study was a DSM-IV [38] diagnosis of schizophrenia made by the current treating consultant psychiatrist. The exclusion criteria were: inability or unwillingness to give consent; inadequate English language skills; and intellectual disability.
Diagnosis, demographic details, criminal and psychiatric history were obtained by a combination of file review and interview with the subject.
Measures
The study was cross-sectional in nature. Using structured questions over the course of one or two interviews, the principal researcher (AC) obtained measures of psychopathology using the Positive and Negative Syndrome Scale for Schizophrenia (PANSS) [39] and of insight using the Schedule for Assessment of Insight (SAI) [15]. The SAI comprises three subscales: ‘awareness of having a mental illness’; ‘ability to relabel symptoms as pathological’; and ‘compliance with treatment’. Higher scores on this measure indicate better levels of insight.
Participants were asked to rate themselves on the Beck Hopelessness Scale (BHS) [40], which comprises a checklist of 20 true/false items. Higher scores reflect a higher degree of hopelessness.
Consent and ethics
The local ethics committee approved the project. Potential participants were given a leaflet outlining the project and given several days to consider their response before being approached again to obtain consent or refusal.
Statistical analyses
All analyses were carried out on SPSS for Windows V10.0 software [41]. Where data were not normally distributed, non-parametric statistical tests were used.
Results
Sample characteristics
Fifty-seven potential participants were approached, of whom 44 consented to participate. Of these, 28 met the criteria for this study. The sample comprised 25 men and three women, with a mean age of 38.2 years (SD 9.0). Twenty-three were born in Australia and 25 spoke English as their first language. They had a mean duration of illness of 12.4 years (SD 8.7). There was a wide range in terms of previous psychiatric hospital admissions: eight participants had no prior admissions, while nine had four or more previous admissions.
All participants had a significant forensic history: 23 had been found not guilty of a serious offence (homicide in 22 cases) on grounds of insanity, and the other five were ‘security patients’ transferred from prison because of mental illness. Twenty-six had initially been placed in custody because of a violent crime; 12 also had a history of violent crime prior to the index offence. The mean time elapsed since index offence was 97.6 months (SD 97.8).
Eight participants were residing outside hospital at the time of the study, either in a halfway house situated in the grounds of the forensic hospital, or in their own residence in the community. The mean time since their discharge into the community was 41.1 months (SD 42.1).
Twenty participants were inpatients at Mont Park Psychiatric Hospital in Melbourne, which was the hospital arm of the VIFMH at the time. Ten of these participants were on the intensive rehabilitation ward, which was locked but of generally low security with many patients having unescorted leaves. Ten were on the slow stream rehabilitation ward, which was of a higher level of security.
Insight and schizophrenic symptomatology
The overall mean insight (SAI) score was 8.39 (SD 4.88). Both positive (Spearman's ρ= −0.54; p = 0.003) and negative (Spearman's ρ= −0.46; p = 0.014) symptom scores on the PANSS were inversely correlated with level of insight. Those with a higher level of psychopathology thus demonstrated lower levels of insight.
Insight and forensic variables
There was no significant relationship between total insight scores and time elapsed since the index offence (Spearman's ρ= 0.23; p = 0.239). Insight scores did not differ between those 12 subjects who had prior convictions for violence before their index offence and the 14 subjects who did not (F-statistic =2.81; p = 0.11; df = 27).
Insight and hopelessness
The overall mean ‘Hopelessness’ (BHS) score was 3.50 (SD 3.01).
To test the hypothesis that certain components of insight were positively correlated with hopelessness, one-tailed non-parametric tests of correlation were carried out. These revealed no significant correlations between hopelessness scores and ‘compliance’ (Spearman's ρ=0.18; p = 0.18).
Hopelessness scores were positively correlated with ‘awareness of illness’ scores (Spearman's ρ= 0.36; p = 0.028), indicating that a higher level of awareness of having a mental illness was related to feeling more hopeless about the future.
Composite total insight scores showed a non-significant trend to be correlated with higher hopelessness scores (Spearman's ρ= 0.28; p = 0.078).
Discussion
Insight in forensic patients
Although previous work has examined the relationship between poor insight and violence [20], [24], this is the first article to quantify insight in a group of patients who are, or have been, subject to custodial orders because of serious offending.
The mean score obtained on the Schedule for Assessment of Insight was 8.39. This is similar to the mean score of 8.48 obtained in a Canadian study of general psychiatry outpatients with schizophrenia [11] and better than the mean of 7.40 found in an outpatient sample in Taiwan [24]. The inpatients in the current study were generally stable and almost all would have been treatable in the community but for their forensic background. It is thus unsurprising that the sample in this study showed a similar level of insight to general psychiatry outpatients. Level of insight was not related to history of previous violent offending prior to the index offence.
This finding of similar levels of insight to general psychiatric patients, combined with the lack of a relationship with extent of past violence, may appear to be at odds with the notion that violence in psychosis is largely attributable to a particularly insight less subgroup of patients. The participants in this study however, were assessed a considerable time, usually several years, after their offences. In the interim period, they had received copious therapeutic input, both pharmacological and psychological. It is likely that, if assessed closer to the time of offending, they would have had more active positive symptoms. Given the clear association between positive symptoms and poor insight [12], it is also likely that their insight would have been correspondingly poorer. If measured closer to the time of offending therefore, it is possible that level of insight would more closely correlate with history of violence.
Another possible reason for the comparatively high levels of insight found is that the group in this study consisted mainly of homicide offenders. It is possible that the characteristics of such offenders generally differ from those of other psychotic patients with a history of repeated, but less extreme, acts of violence. Homicide is a rare event, and most of the homicide offenders in this study had no prior history of interpersonal violence.
Insight and hopelessness
The relationships found between insight and schizophrenic symptoms are consistent with other studies. The finding that poor insight is related to active positive symptoms is well replicated [11], [12], [26], [27], [42], [43] and unsurprising. The relationship of poor insight with negative symptoms is less well established, with no clear consensus emerging [11], [12], [27], [42].
Previous research has demonstrated that patients with schizophrenia who have poor insight into their illness also tend to underestimate problems in their objective life conditions [44], social functioning [45] and some of their own psychosocial needs [46]. A possibly related link with mood, whereby poor insight appears to offer some protection against depressed mood, is now fairly well established [26], [27].
A priori, it was hypothesized that awareness of illness and compliance would correlate with higher levels of hopelessness. This study found no overall correlation between hopelessness and insight. However, as predicted, there was a significant relationship with awareness of illness: those who had a poor awareness of their illness generally had more hope for their future, while those who had accepted that they suffer from a mental illness were more pessimistic. From this study, it is not possible to determine whether such pessimism reflects a more or less accurate appraisal of future possibilities.
A recent study of a non-forensic population [47] using simpler measures of insight and hopelessness found no relationship between the two constructs. It is possible that the relationship between awareness of illness and hopelessness is more marked in forensic patients because of the added burden of guilt about offending and awareness of the likelihood of prolonged institutional detention. Surprisingly however, the mean hopelessness score in this study (3.50) compares favourably with scores in a previous American study [48] using the same scale, which found mean scores of 6.62 and 6.04 for general and forensic psychiatric inpatients, respectively.
Implications for treatment
Whatever the explanation, the finding does suggest the need for careful, measured approaches when educating long-stay patients about their mental health. Psychoeducational and psychotherapeutic approaches that aim to enhance the awareness of relapse risk and of potential long-term deficits should be tempered by a judicious level of optimism. They should also be combined with interventions that address disability and social disadvantage. It is important to be aware of the possibility of exacerbating a sense of hopelessness, and hence suicide risk [47].
The hypothesis that compliance would correlate with hopelessness was not supported. The result may be due to the fairly small sample size. If valid however, it is an encouraging finding, since it suggests that improved compliance may be achievable without negatively impacting on patients’ hopes for the future.
Limitations
The sample size was small because of the number of available patients with a serious offending history. Participants were not representative of forensic patients, and that those who declined may have included a high proportion of more antisocial, insight less patients. Ethical approval to examine this possibility by checking the case notes of refusers was not forthcoming.
The interviewer and rater (AC) was not blind to the hypotheses, which may have led to bias. Finally, the time elapsed since the index offence is an important limitation.
Conclusions
We found that forensic patients with schizophrenia showed similar levels of insight to mainstream psychiatric outpatients. There was a relationship between the acceptance of having a mental illness and hopelessness. This has important implications for psychoeducation, particularly for patients with good premorbid functioning who may be placed in institutional care for long periods after committing a serious offence.
The notion of insight as predictive in forensic patients is likely to appeal to courts and tribunals [49] and has intuitive appeal, but lacks empirical basis. This study did not support such a notion. Future research could examine whether the domains of insight do have predictive value with respect to reoffending by adopting a longitudinal design.
Footnotes
Acknowledgements
We thank patients and staff of the Victorian Institute of Forensic Mental Health, Paul Mullen and James Ogloff (VIFMH and Monash University) for their assistance.