Self-esteem is associated with premorbid adjustment and positive psychotic symptoms in early psychosis

Self-esteem, a global and complex concept, is comprised of both appraisal of self-worth based on personal achievements and anticipation of evaluation by others [1, 2]. Although not uniformly low, self-esteem is often found to be compromised among persons with mental illnesses [3]. Low self-esteem is therefore of considerable interest as it is both a possible consequence and a possible cause of psychiatric symptoms [4‐6].

Regarding self-esteem as a consequence of mental illness, studies predictably show that stigmatization and self-stigmatization may lower self-esteem in persons with mental illness [7]. Low self-esteem also appears to increase the risk of psychiatric disorders such as depression, eating disorders and substance abuse [8]. In psychotic disorders, low self-esteem has been implicated in both the development of delusions [9, 10] and the maintenance of psychotic symptoms [11].

Recent models of global self-esteem suggest that it is both a trait and a state measure [12]. People have a typical, average or trait level of self-esteem, while their momentary, or 'state', judgments of self-esteem can fluctuate around this level dependent on social feed-back and self-judgment. Furthermore, it is the person's interpretation of the event or circumstance, and its relevance to his or her contingencies of self-worth, that determines both if and how strongly it will affect state self-esteem [12, 13]. It appears that treatment failures, functional loss, demoralization and stigmatization may lower self-esteem in patients with severe mental illnesses. To what extent low levels of self-esteem in severe mental disorders are based on underlying, or trait levels of self-esteem, and how this in turn may increase vulnerability to more severe symptoms has not been thoroughly explored. This is of importance both for the understanding of the mechanisms behind the development of psychotic symptoms and also for improving treatment as self-esteem can be influenced by therapeutic interventions [14, 15].

Studies have suggested that difficult childhood experiences such as childhood loss and social marginalization contribute to a cognitive vulnerability accompanied by a negative view both towards the person himself and towards others [4, 11, 16]. It can be hypothesized that individuals with a history of poor premorbid adjustment, both social and academic, are more prone to negative self-evaluation and reduced global self-esteem. MacBeth and Gumley have shown in their review of premorbid adjustment and early symptom development that premorbid problems in psychosocial functioning are associated with a greater severity of illness course and, in particular, more negative symptoms [17]. They also found that reduced quality of life (QoL) was reported by individuals with poorer premorbid functioning. Interestingly, the course of premorbid social adjustment has been found to exert a greater effect on QoL than premorbid academic adjustment [18], and may also be more influential on trait self-esteem. This underlines the importance of separating the social and academic domains of premorbid adjustment.

To our knowledge only one study has tried to examine the relationship between premorbid adjustment and self-esteem in patients with schizophrenia spectrum disorder [19]. They found no relationships between self-esteem and premorbid adjustment in recovered psychotic patients. However, premorbid adjustment was not assessed with a specific instrument which may account for the negative results.

A relatively rich literature exists on the relationship between low self-esteem and symptom formation in severe mental disorders including psychotic disorders. One study showed that the contents of patient's delusions were consistent with patient's global self-esteem and suggested that low self-esteem accounted for the persistence of delusions [20]. Other studies found significant correlations between negative self-evaluation and a wider variety of positive symptoms i.e hallucinations and delusions, in schizophrenia [10]. It has also been found that patients with a low level of self-esteem and more depressive symptoms had more intense auditory hallucinations with a more negative content [21]. In addition, it has been found that patients who had both high levels of suspiciousness and low self-esteem made more misattributions of anger which may also fuel delusional ideation [22]. This is in line with findings from the general population, where delusion prone individuals show lower self-esteem [23]. Finally, it has been found that several delusional themes including persecution, thought disturbances/thought broadcasting, catastrophic ideation, and negative self beliefs were related to low self-esteem [24].

Other studies have shown higher levels of self-esteem in patients with delusional disorder compared to depressed patients [25]. However, the authors found that the group without depressive symptoms had significantly higher levels of grandiose ideation than the other groups which may have accounted for the elevated levels of self-esteem. The authors concluded that persecutory delusions may reflect an attributional style protecting the individual from low self-esteem. The same has been hypothesized for grandiose delusions, but the few studies in this area do not clearly support this hypothesis [26]. Other studies have found equal levels of self-esteem in patients with delusions and matched healthy controls with both groups demonstrating higher levels than depressed patients [27].

Self-esteem has been found to fluctuate over the short-term. It has been demonstrated that paranoid individuals display more fluctuations in their self-esteem, and that the fluctuation predicts the degree of subsequent increase in paranoid thinking [28]. However, other studies indicate that changes in both positive and negative beliefs about the self are related more to changes in negative symptoms than changes in paranoid symptoms [29]. In summary the relationship between premorbid function, self-esteem and the formation of psychotic symptoms remains unclear.

To date the relationship between self-esteem, psychotic symptoms and premorbid adjustment in the early stages of psychosis has not been thoroughly explored. Previous studies of that nature have all been conducted with patients with chronic psychotic disorders where the effects of a long-term severe illness and secondary processes may significantly confound relationships. More studies are thus needed to explore the relationship between self-esteem and psychotic symptoms during the early phases of psychotic disorder. This is of importance as patients coming to their first treatment for a psychotic disorder are less influenced by stigmatization, treatment failures, and subsequent disappointments which may contribute to lowered self-esteem.

The aims of the current study are thus to investigate the following questions in a large and well characterized group of patients with first episode psychosis:

Correlations between demographic/clinical characteristics and self-esteem were calculated using Pearson's product moment co-efficients. To estimate how much of the variance in self-esteem was explained independently by premorbid functioning we performed a block-wise hierarchical multiple regression analysis with age and gender entered in the first block and premorbid adjustment in the second block. As academic adjustment in childhood versus academic adjustment in early adolescence, and social adjustment in childhood versus social adjustment in early adolescence were strongly inter-correlated (with r = 0.66 and r = 0.77 respectively), only results for early adolescence were entered to represent PAS and avoid co-linearity problems. The associations between global self-esteem and hallucinations and delusions, both general and persecutory, were analyzed similarly using Pearson's correlations and followed up with three block-wise hierarchical multiple regression analysis with hallucinations, delusions and persecutory delusions as the dependent variables. Demographic information was placed in the first block, premorbid adjustment in the second block, depression i.e whether the patient was in a major depressive episode or not, in the third block and self-esteem in the fourth. By entering self-esteem in the fourth we adjusted for the amount of variance explained by the variables in the first three blocks. Finally, we conducted various interactional analyses to explore whether self-esteem acted as a mediator or moderator of the relationship between premorbid adjustment and symptoms.

Table 1 shows the patient characteristics of the 113 included patients.

As shown in table 2, self-esteem was significantly correlated with several demographic and clinical characteristics, including the four sub-scale measures of premorbid adjustment and with current levels of symptoms (depression, persecutory delusions and hallucinations, poor rapport and stereotyped thinking). Furthermore, females reported lower self-esteem than men.

In the first hierarchical multiple regression analysis, with self-esteem as the dependent variable, the included variables explained 25% of the variance in self-esteem (Table 3). Only gender and social adjustment in early adolescence contributed significantly to the level of global self-esteem. Gender explained 16% of the variance while premorbid social adjustment explained an additional 9%.

In the second hierarchical multiple regression analysis performed, with positive psychotic symptoms as the dependent variable, self-esteem explained a significant amount of the variance in both hallucinations and persecutory delusions, even after adjusting for age, gender, premorbid adjustment and depression. In general, levels of self esteem did not explain a significant amount of the variance in occurrence of delusions (P1) (Table 4).

Finally, various interactional analyses revealed no significant interaction between premorbid adjustment and symptoms mediated or moderated by self-esteem.

This study demonstrates both a statistically significant relationship between poor premorbid social adjustment and low levels of global self-esteem and between self-esteem and positive psychotic symptoms i.e hallucinations and persecutory delusions. The relationship between self-esteem and positive psychotic symptoms remained significant even after adjusting for the presence of a major depressive episode indicating that this effect is not mediated by the presence of depressive symptoms.

The current study is one of the first to show a relationship between poor premorbid social adjustment and level of global self-esteem in psychotic disorders. The only other study exploring this relationship [19] did not apply a specific validated measure of premorbid adjustment such as the PAS, but instead divided subjects by use of data collected by means of the Diagnostic interview for Psychosis (DIP) into 'yes' or 'no' for poor premorbid adjustment. This implies both a less validated measure of premorbid adjustment and a subsequent loss of variance in statistical analysis. In addition their sample consisted of older participants with a longer duration of illness, and thus a higher risk for recall bias.

Premorbid social adjustment as a concept incorporates issues including such factors as how you interact with your schoolmates, adjust to groups and friends and the presence of age-relevant sexual interest. Previous studies have shown that general cognitive abilities, exposure to bullying, social marginalization, abuse, neglect or the presence of neurodevelopmental factors are the strongest predictors of social adjustment [35‐37]. Many patients with psychotic disorder may have a premorbid vulnerability that reduces their ability to achieve and maintain social competence and thus affect their premorbid social adjustment [9, 11, 38]. This will in turn affect the individuals schematic beliefs about themselves and others. These beliefs influence how we experience ourselves in relation to the world. In these patients the effect on these schematic beliefs may lead to social adversity and a feeling of low self-esteem [16].

Furthermore, studies of persons with auditory hallucinations have shown that voice hearers experience a subordinate relationship to their voices mirroring other social relationships. This suggests the existence of maladaptive inter-personal schemata serving both [39]. These schemata are not necessarily a result of the psychotic illness, but may be a result of poor premorbid social adjustment and are in line with theories of how long-term experience of social defeat can be a risk factor for psychosis [35].

We also found self-esteem to be a predictor of both hallucinations and persecutory delusions in early psychosis, even though the explained variance was rather moderate. This is in line with previous studies [10, 21, 28, 40]. Garety's cognitive model of psychosis [11] suggests that the experience of social adversity and lowered self-esteem can lead to the development of psychotic symptoms through an increased vulnerability to psychotic disorders.

We further argue that poor premorbid social adjustment with social withdrawal and subsequent marginalization provides content for psychotic attribution due to a lack of correcting social feedback. This is supported by findings in studies of patients at high risk of developing psychosis [41]. In line with this, cognitive behavioral therapy aiming to improve self-esteem by correcting misattributing tendencies has shown clinical benefits in terms of both increased self-esteem, reduced positive symptoms and improved social functioning [14].

Our findings are also supported by findings in the general population. Negative ideas about oneself and others have been found to be predictors of paranoid thinking in the general population [42]. In addition, premorbid neuroticism and low self-esteem were associated with subsequent development of psychosis or psychosis-like symptoms at 3-year follow-up in a Dutch population sample [43]. If we take the continuum hypothesis of psychosis into consideration [44], it is not surprising to find the same pattern in a first episode patient (FEP) sample.

However, there are studies showing that patient's self-stigma tends to be most affected during the early course of the disease, and that self-stigma and self-esteem are closely related [45]. It may be that there are sub-groups within the psychosis spectrum that differ with regard to stability in self-esteem, and also differ in which factors, either long or short term, have the most impact on levels of self-esteem. These complex mechanisms need to be explored further in longitudinal studies.

Gender was a significant predictor of self-esteem in this FEP sample, with women having significantly lower levels of self-esteem than men even after correction for differences in levels of depression. A vast body of literature from the general population indicates a small but significant gender difference in the same direction [46]. There is surprisingly little research on gender differences regarding self-esteem in psychosis but the present study is supported by findings from the Danish Opus trial [47] and suggests that gender difference is a factor which warrants further investigation.

The present study has some limitations. This is a cross-sectional study where conclusions about directions of relationships cannot be ascertained, and where data on premorbid adjustment is necessarily gathered retrospectively. There may also be a recall bias regarding the scores for premorbid adjustment. To what degree self-esteem is affected before the development of psychosis is thus not possible to test directly using the current design. Furthermore, the present study, due to the limitations of the study protocol, did not allow for more sophisticated measures of self-esteem, such as the measurement of fluctuations in self-esteem, which would have been of interest in this group.

The current study revealed both a significant association between premorbid adjustment and self-esteem and between self-esteem and positive psychotic symptoms. Future studies of self-esteem should consider examining how self-esteem changes over the course of illness from the prodromal stage. It would also be of interest to explore factors such as stigma and metacognition and their relation to self-esteem. There are several factors that in further studies could be explored as possible mediators in this context. Women with psychotic disorders report a high prevalence of sexual trauma [48] which is a known risk factor for low self-esteem. In addition, women's self-esteem may be more affected by medication induced weight-gain than men.

It is our opinion that this study may have clinical implications. It is possible that psychotheraputic interventions, such as cognitive behavioral therapy, may increase self-esteem and lessen the likelihood of development of positive psychotic symptoms or decrease their severity. Furthermore, psychotherapeutic intervention may help patients to acquire a broader personal narrative which would benefit their self-esteem.