Shame amplifies the association between stressful life events and paranoia amongst young adults using mental health services: Implications for understanding risk and psychological resilience

Highlights

• Shame is associated with increased risk of paranoia.

• Shame amplifies the likelihood that life stressors will lead to paranoia.

• Shame may be an important factor amongst young adults using mental health services.

• Low levels of shame may confer resilience against risk of paranoia.

Abstract

Shame is associated with a range of psychological disorders, and is a trans-diagnostic moderator of the association between stressors and symptoms of disorder. However, research has yet to investigate shame in relation to specific psychotic symptoms in clinical groups. In order to address this, the present study investigated shame in young adults with mental health problems, to test whether shame was i) directly associated with paranoia, a prevalent psychotic symptom, and ii) a moderator of the association between stress and paranoia. Sixty participants completed measures of stressful events, paranoia, shame, depression and anxiety. Results from a cross-sectional regression analysis suggested that shame was associated with paranoia after the stressful life event measure was entered into the model, and shame moderated the association between stress and paranoia. For individuals scoring high on shame, shame amplified the association between stress and paranoia, but for low-shame individuals, the association between stress and paranoia was non-significant. These findings suggest that high levels of shame could confer vulnerability for paranoia amongst clinical groups, and that resistance to experiencing shame could be a marker of resilience.

Introduction

Shame is associated with a range of psychological disorders including depression (Kim et al., 2011), anxiety (Fergus et al., 2010), PTSD (Andrews et al., 2000), eating disorders (Troop et al., 2008) and schizophrenia (Suslow et al., 2003). Importantly, research suggests that shame is a trans-diagnostic moderator of the association between stressors and psychological difficulties (Harper and Arias, 2004, Beck et al., 2011, Shorey et al., 2011). Despite this, research has yet to investigate shame in relation to specific psychotic symptoms amongst clinical groups. The present study addressed this by investigating shame in relation to paranoia, a prevalent psychotic symptom in young people. We conducted this investigation in a sample of adolescents and young adults with mental health difficulties, who did not meet criteria for a schizophrenia-spectrum disorder. This group is at higher risk of developing severe mental health disorders due to their elevated rates of depression and sub-threshold psychotic symptoms (Yung et al., 2003, McGorry, 2013). As research suggests that experiences of paranoia predict subsequent increases in general psychotic symptoms (Kramer et al., 2014), these findings may have implications for concepts of risk and resilience.

Shame is thought to arise when one perceives oneself as defective, experiences a sense of threat to the social self and a feels a need to hide (Feiring et al., 2002, Tangney and Dearing, 2003). It has been proposed that when experienced in response to specific situations and decisions, feeling shame can promote pro-social behaviours (de Hooge et al., 2008). However, when it is experienced more generally – and not in relation to a particular decision – shame no longer promotes pro-social behaviour and can become harmful (de Hooge et al., 2008). The vast majority of research has focused on this more general sense of shame, and theoretical perspectives have long supported its importance in the aetiology of psychotic symptoms, particularly paranoia. In an early model, Colby and colleagues (Colby, 1975, Faught et al., 1977) suggested that paranoia was a way of avoiding underlying general shame by attributing perceived inadequacy to external sources. A similar view has been outlined by Harper and Cromby (2013), who propose that symptoms of paranoia need to be contextualised. They suggest that paranoia tends to evolve in the presence of difficult life situations and may offer a compensatory function to the aversive emotion states this causes, particularly feelings of shame (Harper and Cromby, 2013).

In studies of non-clinical participants, shame has been found to be associated with higher levels of paranoid anxiety (Matos et al., 2013) and paranoid ideation (Tangney and Dearing, 2003, Matos et al., 2012). In clinical samples, higher levels of shame have been found in individuals with anhedonic schizophrenia (Suslow et al., 2003) and first-episode psychosis, than controls (Michail and Birchwood, 2012). In family members of individuals with psychosis, shame has been associated with greater presence of expressed emotion (EE; Wasserman et al., 2012). More broadly, in populations with mental health disorders, higher levels of shame have been associated with greater levels of hopelessness, stress, and lower self-esteem (Rüsch et al., 2009, Rüsch et al., 2014). Shame has also been found to mediate the association between insight and self-stigma, which may be detrimental for levels of self-esteem, hope, and inter-personal relationships (Hasson-Ohayon et al., 2012).

However, only a limited amount of research has investigated shame in relation to psychotic symptoms, and there are four areas which need addressing. First, it is necessary to investigate whether shame is associated with specific psychotic symptoms in clinical groups. Previous research in clinical groups has focused on overall ratings of psychotic symptoms or diagnostic categories (Suslow et al., 2003, Michail and Birchwood, 2012), which offers general indications of the relevance of shame in psychosis, but is unable to suggest whether shame may lead to specific symptoms or experiences of psychosis. In particular, there is a need to investigate whether shame is associated with paranoia. Paranoia is a common feature of psychosis (Wigman et al., 2011, Stompe et al., 2005), which predicts subsequent psychotic symptoms (Kramer et al., 2014). In order to address this, the current study investigated the association between shame and paranoia in a clinical population.

Second, it is necessary to investigate whether shame moderates the association between stressful life events and paranoia. Several studies have implicated a role for stressful events in the aetiology of general psychotic symptoms (Bebbington et al., 1993, Freeman et al., 2011) but results have been inconsistent (Chung et al., 1986, Van Os et al., 1994). It is possible that these varying results have been caused by the presence of unrecognised moderating factors, which may “amplify” or “buffer” the association between an input and an outcome variable. Given the theoretical and empirical associations between stress, shame and paranoia in particular, it is important to investigate whether shame may act as a moderator of the stress–paranoia relationship. If shame was found to “amplify” the association between stressful life events and paranoia, this would suggest that a greater association between stressful events and paranoia would be found amongst populations with higher levels of shame than those with low levels of shame. Alternatively, lower levels of shame would confer resilience. As “amplifying” and “buffering” factors can be viewed as representing two ends of a vulnerability/resilience spectrum (Johnson et al., 2011), these terms can be considered interchangeable. In order to address this, the current research investigated whether shame acted as a moderator of the association between stressful events and paranoia.

Third, there is a need to investigate the association between shame and paranoia in a population at higher risk of experiencing severe mental health problems. Previous research has focused upon non-clinical groups and groups diagnosed with psychotic disorders. However, there is a growing emphasis upon preventative approaches in high-risk sub-groups (McGorry, 2013). Research has found that sub-threshold psychotic symptoms such as paranoia and depressed mood can predict subsequent increases in psychotic symptoms (Yung et al., 2007a, Kramer et al., 2014). Treating individuals reporting symptoms can prevent transition to diagnosable disorder (Preti and Cella, 2010), which may represent the most economically and clinically effective intervention approach (Yung et al., 2007b, McGorry, 2013). An understanding of psychological risk factors for psychotic symptoms could improve identification of individuals at highest risk. Participants in the present study were young adults attending a mental health care team who did not meet criteria for a schizophrenia-spectrum disorder. This study did not seek to sample using traditional ultra-high-risk for psychosis criteria proposed by Yung et al. (2003), but instead utilised the Clinical Staging Approach. This approach suggests that severe mental health disorders develop from undifferentiated symptoms in adolescence (McGorry et al., 2006). That is, adults who are diagnosed with severe mental health disorders are likely to have demonstrated evidence of a range of possible mental health symptoms in earlier years. According to this approach, young adults presenting to mental health services can be viewed as potentially being in the earlier stages of more severe mental health disorders. They may go on to have a range of possible symptom and functioning trajectories, and are deemed as high risk for all severe mental health disorders, including psychotic disorder. Consistent with this, the young adults attending the secondary mental health team from which participants were recruited were presenting with elevated levels of various symptoms of psychological distress. These symptoms were too severe to be suitable for treatment by primary care services. Thus, using the Clinical Staging Approach, the current population can be considered as higher risk than the general population for a range of severe psychopathology, including psychosis.

Fourth, it is necessary to explore whether there are particular sub-types of shame which are relevant for this group. Research suggests that shame may have different aspects, which could have divergent origins and require varying forms of therapeutic management (Gilbert, 1998, Andrews et al., 2002). However, much empirical research has been overly focused on measuring general shame (Leeming and Boyle, 2004). Andrews et al. (2002) have suggested that shame can be divided into three subcategories, behaviour shame, characterological shame and body shame. Behaviour shame refers to shame regarding actions and perceived failures, characterological shame refers to shame regarding habits, manners and more general sense of self, and body shame refers to shame regarding the body. Studies using this conceptualisation have found that these different shame types have varying associations with symptoms of psychopathology, for example, whilst characterological shame and body shame have been found to be significantly associated with eating disorder symptoms (Swan and Andrews, 2003, Doran and Lewis, 2012), and deliberate self-harm (Flett et al., 2012) behaviour shame has been found to prospectively predict symptoms of depression (Andrews et al., 2002). This suggests that this conceptualisation of shame may have both clinical and predictive utility. Despite this, research has yet to explore these types of shame in relation to psychotic symptoms, or in groups at higher risk of experiencing psychotic symptoms. The current study investigated total shame and these shame sub-types in relation to symptoms of paranoia in a high risk group.

In summary, the current study aimed to investigate whether total shame and three shame subscales were i) positively associated with symptoms of paranoia and ii) moderated the association between stressful events and paranoia amongst young adults attending mental health services. It was predicted that there would be a positive association between total shame and paranoia, and that total shame would amplify the association between stressful events and paranoia. As the shame subscales have not yet been investigated in relation to paranoia, these tests were exploratory. The study also investigated whether significant findings were maintained while controlling for the variance explained by depression and anxiety, as these have been found to be associated with shame (Swan and Andrews, 2003, Rusch et al., 2006, Fergus et al., 2010).