Abstract
There is considerable evidence to suggest that the prevalence of anxiety disorders in general, and social anxiety specifically, among individuals with psychosis is relatively common [1]. Cosoff and Hafner [2] examined 100 consecutive inpatients with psychosis using the Structured Clinical Interview for DSM-III-R [3]. The prevalence of social anxiety disorder (17%), obsessive-compulsive disorder (13%), and generalized anxiety disorder among individuals diagnosed with schizophrenia was relatively high, as was the prevalence of obsessive-compulsive disorder (30%) and panic disorder (15%) among individuals with bipolar disorder. The proportion of participants with an anxiety disorder (43–45%) was almost identical across individuals diagnosed with bipolar disorder, schizophrenia, and schizoaffective disorder. Few of this cohort had received specific treatment for their anxiety disorder, despite the evidence that cognitive therapy is effective in the treatment of social anxiety in general [4] and social anxiety among individuals diagnosed with schizophrenia [5].
There are a number of views which may account for the relationship between social anxiety and psychosis [1]. First, social anxiety and poor premorbid adjustment may be an early precursor or risk factor for the development of psychosis. Social anxiety and poor early adjustment has also been shown to be elevated among individuals who go on to develop psychosis [6–10]. In a cohort study of 50 054 Swedish army conscripts, Malmberg et al. [7] found that a combination of interpersonal problems including difficulty mixing in groups, having few friends, feeling more sensitive than others and not having a girlfriend were associated with an increased risk of developing schizophrenia. Jones et al. [6] in a birth cohort study of 4746 individuals born in one week in 1946 found that self-reported social anxiety at age 13 years and teacher-rated social anxiety at age 15 years was associated with an increased risk of developing schizophrenia in adulthood.
Second, social anxiety may share common psychological processes with psychosis. For example, both persecutory delusions and social phobia are associated with similar cognitive processes insofar as they are characterized by abnormal selective attention to threat-related material [11], [12], and both represent a response to interpersonal threat [13]. In addition there is evidence that auditory hallucinations, paranoia and social anxiety are all characterized by heightened self-consciousness [14]. Furthermore social anxiety, paranoia and auditory hallucinations are associated with the use of safety-seeking behaviours aimed at preventing or minimizing threat [15].
Third, social anxiety may occur as a consequence of self-stigmatizing negative beliefs about the experience of psychosis [16] and thus represents a marker for poor social and emotional adjustment [17]. Negative appraisals of self and psychosis have also been found to be associated with the development of post-psychotic depression (PPD) [18], [19]. Birchwood et al. [18] found that in a sample of 105 individuals with schizophrenia, 36% (n = 28) developed PPD without concomitant changes in positive and negative symptoms. Iqbal et al. [19] have proposed that certain life situations are likely to be depressogenic, particularly if they encapsulate feelings of loss, humiliation and entrapment. In line with their proposal, psychosis is seen as a life event whose appraisal may involve these elements. Participants who developed PPD were more likely than their non-PPD counterparts to attribute the cause of psychosis to themselves, to perceive greater loss of autonomy and a valued role, and perceive themselves to be entrapped and humiliated by their illness.
This study sets out to test the hypothesis that individuals with schizophrenia who also have a concurrent social anxiety disorder will feel more loss, entrapment, shame and humiliation, and will blame themselves more for their psychosis, in comparison to individuals with schizophrenia who have no concurrent psychological disorder. In addition, it is hypothesized that those with concurrent social anxiety will have lower self-esteem than those without social anxiety.
Method
Participants
A sample of 38 individuals participating in a randomised controlled trial of cognitive behavioural therapy for relapse [20] was selected for this analysis. Entry to this larger study required that patients fulfilled DSM-IV [21] criteria for schizophrenia or a related disorder confirmed by the Structured Clinical Interview for DSM-IV [3], were aged 18–65, were receiving antipsychotic medication and were considered relapseprone. Patients were considered relapse-prone if they had one or more of the following characteristics: a history of relapse in the last two years; if their keyworker viewed them as living in a stressful environment (e.g. a home environment characterized by high levels of expressed emotion); living alone or socially isolated; nonadherence with antipsychotic medication (where this was viewed as problematic by the participant's keyworker and/or prescribing psychiatrist); and being on a neuroleptic dosage reduction program. Participants were also included in the study on the basis of their informed consent to participate and the consent of their Responsible Medical Officer. Patients were excluded if they were a non-English speaker, had organic brain disorder, there was presence of significant learning disability, severe positive psychotic symptoms (rating of 5 or more on the positive scale of the Positive and Negative Syndrome Scale, [PANSS]) [22], a primary drug or alcohol dependence disorder (based on the opinion of the key worker), or being in receipt of a concurrent psychotherapy outside the study.
Of the 144 participants in the larger study, 19 who met criteria for concurrent social anxiety disorder [1], [21] were included in the analysis. Nineteen participants with no concurrent psychological disorder, taken from the same larger sample, matched for gender and primary diagnosis, acted as a control group.
Measures
All measures were taken at entry to the study prior to randomization. Positive and negative symptoms were measured using the Positive and Negative Syndrome Scale [22]. The PANSS is a 30-item observer-rated scale. Each item is rated on a severity scale ranging from 1 (absence of psychopathology) to 7 (extremely severe). The raters were trained using interviews with patients, either face-to-face or on video. Interclass correlations for the positive, negative and global subscales were 0.91, 0.87, 0.72, respectively, and for the total score 0.80.
Negative beliefs about illness were measured using the Personal Beliefs about Illness Questionnaire (PBIQ) [23]. This 16-item questionnaire has five subscales; ‘self versus illness’; ‘entrapment’; ‘shame’; ‘humiliation’; and ‘loss’. In addition, the questionnaire has been shown to have adequate internal consistency and test-retest reliability [23].
Negative beliefs about self were assessed using the Rosenberg Self Esteem Scale (RSES) [24]. Items are in statement form, and respondents are asked to rate their degree of agreement on a four-point scale (strongly agree to strongly disagree).
Psychological distress was assessed using the Brief Symptom Inventory (BSI) [25]. The BSI is a 53-item self-report scale, which provides a measure of psychological symptomatology. Participants are asked to rate how much they have been distressed by specific symptoms in the last week by rating on a 5-point scale anchored by ‘not at all’ and ‘extremely’. Adequate test-retest and internal consistency reliabilities, and convergent, construct and criterion-referenced validities have been demonstrated for this scale [25].
Data analysis
The data were checked to ensure that the assumptions for parametric statistical analysis were met. Parametric statistics were used for all the analyses. Given the exploratory nature of this study Bonferonni corrections were not utilized.
Results
Demographic and clinical data
Table 1 shows characteristics of a total of 38 participants (social anxiety disorder, n = 19; no comorbid disorder controls, n = 19) matched for primary diagnosis and gender. No significant differences were detected for age, positive symptoms, negative symptoms and global psychopathology. Table 1 also shows the univariate comparisons between participants with social anxiety disorder and controls on the BSI subscales. As can be seen, those with social anxiety disorder had significantly higher scores for phobic anxiety (F = 4.433, p < 0.05) and depression (F = 4.910, p < 0.05).
Clinical and demographic characteristics of the study sample
Negative beliefs about self and Illness
In order to investigate whether participants with social anxiety disorder had more negative beliefs about self and illness than controls, a multivariate analysis of variance (MANOVA) was conducted with the subscales of the PBIQ and RSES total score entered as dependent variables. This revealed a significant difference between the two groups (F6,31 = 3.3, p < 0.05). Univariate contrasts for this analysis are shown in Table 2. Participants with social anxiety disorder had significantly higher scores for self versus illness (F1,36 = 5.0, p < 0.05), entrapment (F1,36 = 12.7, p < 0.01), shame (F1,36 = 10.6, p < 0.01) and RSES (F1,36 = 10.2, p < 0.01).
Univariate comparisons for negative beliefs about self and illness
A second MANOVA was conducted, this time entering BSI depression as a covariate. Following this the overall model was no longer significant (F6,30 = 2.2, p = 0.07). BSI depression was a significant covariate (F6,30 = 6.3, p < 0.01). Univariate contrast revealed that having controlled for BSI depression, entrapment (F1,35 = 7.4, p < 0.01), shame (F1,35 = 5.0, p < 0.05) and RSES (F1,35 = 4.6, p < 0.05) remained significant.
Post hoc analysis
Partial correlations between negative beliefs about illness, selfesteem and BSI phobic avoidance were conducted, controlling for depression as a covariate. Self-blame (r = 0.35, p < 0.05), entrapment (r = 0.50, p < 0.01), shame (r = 0.45, p < 0.01) and loss (r = 0.42, p < 0.01), correlated with BSI phobic avoidance.
Discussion
The results suggest that individuals with psychosis who have a concurrent social anxiety disorder, exhibit significantly higher negative beliefs about psychosis and have lower levels of self-esteem than controls without social anxiety disorder, who were matched for gender and diagnosis. They scored significantly higher on selfblame, entrapment, shame and self-esteem measures. Given that there were no differences between the two groups in terms of their positive and negative symptoms, these differences cannot be attributed to differences in psychotic symptomatology. Neither can these differences be attributed wholly to depressed affect. After controlling for depression, participants with social anxiety disorder continued to exhibit significantly higher scores on entrapment, shame and self-esteem. The results are consistent with Birchwood's [17] proposal that in psychosis, social anxiety is associated with the perception of being shamed and socially subordinated by others because of their psychosis and patient status and that shame, entrapment and lower self-esteem underpin social anxiety and avoidance in schizophrenia. Our results also confirm the association between negative beliefs concerning shame, entrapment, self-blame and loss arising from psychotic illness and phobic avoidance among individuals with psychosis.
There are a number of important limitations, which should be borne in mind in interpreting and generalizing the results of this study. First, this study comprised of a relatively small sample size of convenience. Second it was not possible to investigate whether participants' social anxiety disorder predated the onset of their psychosis. There is considerable evidence that social anxiety and interpersonal problems predate the onset of psychosis. Third was the use of the Rosenberg Self Esteem Scale as a measure of self-esteem. Barrowclough et al. [26] have raised the importance of methodological limitations of studies investigating self-esteem and schizophrenia. One major problem is the use of questionnaire measures of self-esteem as they only measure global self-esteem and are highly mood-dependent. However, having controlled for depression as a covariate, socially anxious participants' sense of shame, entrapment and low self-esteem remained significantly higher than the non-socially anxious controls. Fourth, the study was cross-sectional in nature and it is not possible to conclude that negative beliefs about psychosis are causal to the development of concurrent social anxiety in schizophrenia. However, the results are consistent with a growing body of evidence that emotional adjustment following the development of schizophrenia is linked to the appraisal of psychosis and patient status as entrapping, shameful, humiliating and involving significant loss [19], [23]. Further research would benefit from the use of prospective methodology to investigate the role of negative appraisals of psychosis and selfesteem in the development and evolution of social anxiety and other psychological disorders, including post-psychotic depression following the first episode psychosis.
Halperin et al. [5] found that group CBT for individuals with social anxiety disorder with a concurrent diagnosis of schizophrenia was successful in reducing symptoms of social anxiety, avoidance and depression compared to controls. Group CBT in this study comprised of a combination of exposure and cognitive restructuring of cognitions related to social anxiety and avoidance. Our results suggest that psychological treatments, such as CBT, for social anxiety disorder could also incorporate strategies aimed at helping individuals explore how their experience of psychosis is associated with social anxiety via their negative appraisals of themselves and their illness. Indeed, strategies aimed at modifying appraisals of fear, entrapment and shame arising from psychosis may be helpful in alleviating the emotional distress, which underpins social anxiety and avoidance in schizophrenia.
Footnotes
Acknowledgements
The authors thank the Chief Scientist Office, Scottish Executive, Grant Number K/RED/18/13 and the Chief Scientist Support Fund, administered by Ayrshire and Arran Primary Care Trust, for financial support.