Abstract
Recent advances in the management of psychosis offer a promise of better outcome for affected persons [1]. Individuals with better symptomatic outcome will expect to have commensurate enhancement of quality of life. However, low self-esteem might make an enhanced quality of life more difficult to achieve. Recent quantitative [2] and qualitative [3] research suggests that the experience of mental illness has an adverse impact on self-esteem and that improving self-esteem may be integral to recovery [4]. In this report, we examine the profile and correlates of self-esteem among a group of community-dwelling persons who have achieved symptomatic recovery from psychosis and explore its relationship with quality of life.
Method
The present study was an extension of the study of low prevalence disorders (LPD), a component of the National Survey of Mental Health and Wellbeing which was commissioned and funded by the Commonwealth Government of Australia [5]. The aims of the LPD were to determine the prevalence of psychotic disorders in a representative sample of the urban and suburban Australian population, to assess the level of disability associated with psychosis and to evaluate the use of health services by individuals with psychotic disorders.
The LPD was conducted by four research teams based in Brisbane, Canberra, Melbourne, and Perth. However, the add-on component reported here was conducted only in Melbourne. Details of the method of the LPD are provided elsewhere [5], [6]. In brief, during a one-month period in 1997, persons consulting health facilities in the designated catchment areas at each of the sites were screened for the presence of a lifetime history of psychosis. The health facilities included in the survey were public psychiatric services, both inpatient and communitybased, and facilities providing treatment and services for special groups such as individuals with drug or alcohol problems. Also included were the clinics of general practitioners (GPs) and private psychiatrists who had previously indicated that they provided care for persons with psychotic disorders and were willing to participate in the survey. For the present study, participants were drawn from those using public community mental health services and from those recruited from the clinics of GPs and private psychiatrists.
Assessment tools
Screening was conducted with the psychosis screen [5], which is largely a modification of the Psychosis Screening Questionnaire (PSQ) [7]. The psychosis screen is a brief 6-item questionnaire that elicits information about the lifetime occurrence of common symptoms of psychosis. It also inquires whether respondents have ever received a diagnosis of psychotic disorder from a clinician. The screening was conducted by the clinicians primarily responsible for the care of the respondents and who had knowledge of patients' clinical history. Included in the screening questionnaire was a question asking the clinician to indicate whether, in their own judgement and considering their knowledge of the person's clinical history, they thought the person had been free of any psychotic symptom for at least the previous 6 months.
The second phase assessment was conducted by trained research interviewers using a specially designed semistructured interview schedule, the Diagnostic Interview for Psychosis (DIP) [5], [6], [8]. A computer-generated diagnosis based on the ICD-10 [9] was produced for the present report. The DIP provides for the semistructured collection of information about patients' premorbid social adjustment and the history of illness. Using this information, interviewers made ratings of evidence of poor premorbid social adjustment and course of illness. For the purpose of this paper, we dichotomized the former as yes versus no and the latter as good (single or multiple episodes with good recovery) versus poor (incomplete recovery from single or multiple episodes or continuous chronic illness). The DIP included questions on current medication, its perceived helpfulness and perceived impairment due to medication [10].
As part of the second phase assessment of this sample, and for the purpose of this extension of the LPD study, respondents were given a set of three self-report questionnaires. (i) The Self-Esteem Rating Scale (SERS) is a 40-item scale that assesses not only problematic levels of self-esteem, but also positive levels of self-esteem [11]. Scores range from −120 to ++120, spanning from low or problematic self-esteem to high or good levels of self-esteem. A score of 43 or less is often indicative of a problematic level of self-esteem [12]. (ii) The WHOQOL-Bref is a 26-item questionnaire designed by the World Health Organization as a culturally sensitive brief tool for measuring quality of life [13]. Derived from the longer WHOQOL-100, the WHOQOL-Bref evaluates perceived quality of life in four main domains: physical health, psychological health, social relationships, and environment. (iii) The 28-item General Health Questionnaire (GHQ-28) is a widely used screening tool for psychological distress [14]. A score of 5 or more is usually indicative of the possibility of diagnosable psychological disorders, commonly depressive or anxiety disorders. Since low self-esteem is often related to depression, we also derived a depression subscore from the GHQ-28 using the seven items related to such experience.
Participants
The plan was to assess 100 consecutive individuals who were judged by their clinicians to have been free of psychotic symptoms for 6 months. At the end of the survey, a total of 99 individuals, who met the other selection criteria of being aged 18–64 years and residing in the catchment area, were judged by their clinicians to have been free of psychotic symptoms. Of these 99 individuals, 38 had evidence of hitherto clinically unrecognized symptoms in the previous month on detailed DIP assessment and were therefore excluded from the study. Among the 61 confirmed to have been free of psychotic symptoms, 43 had been symptom-free for at least 1 year. These 61 individuals who met this definition of being ‘asymptomatic’ form the basis of this report. They represented a fair mix in terms of service use: 7 (12%) were not in receipt of regular after-care service, 12 (20%) each were receiving such services from their GPs or private psychiatrists, respectively, while the rest were in outpatient follow-up in specialist psychiatric clinics. All of these respondents were asked to complete the DIP interview, the SERS, the GHQ-28, and the WHOQOL-Bref. There were slightly more males than females (53% vs. 47%). The median age of the group was 41 years (range 22–63, with interquartile values of 32, 41, and 53 years). More than three quarters of the sample were unmarried, never married, widowed, or separated/divorced. There was a high level of unemployment in this group – approximately 56%. Only about 12% were engaged in full-time jobs outside the home, while approximately 26% were in part-time employment. Based on the DIP interview, 25 of the participants were assigned an ICD-10 lifetime diagnosis of schizophrenia or schizoaffective disorder by the computerized algorithm, 23 received a lifetime diagnosis of affective psychosis, while 13 had other forms of psychosis. One male did not complete the SERS and was therefore excluded from the subsequent analyses.
Analysis
We compared the distribution of discrete variables between groups with the χ2 test and differences between continuous variables with t-test statistic. We explored factors that may be associated with level of selfesteem using a multiple regression modelling in which the total score on the SERS was the dependent variable and putative correlates and covariates were the independent variables. Using multiple regression, we determined the contribution of self-esteem to level of quality of life after controlling for the effects of sex, age, level of depression (using the score on the GHQ-28 depression subscore) and lifetime diagnosis (schizophrenia/schizoaffective disorder or affective psychoses). Reported probabilities are two-tailed (except for the multiple regression) and level of statistical significance was set at 0.05. Analyses were performed with the SPSS software (SPSS Inc, Chicago, IL).
Results
Self-esteem profile
The median score on the SERS was 52. As many as 26 (43%) of the participants scored less than 43, a level often regarded as indicative of a problematic degree of self-esteem [12]. Men were similar to women in the proportion scoring less than 43: 41.9% versus 44.8% (χ2 = 0.05, df = 1, p = 0.82). Single, separated/divorced or widowed participants did not significantly differ from married participants in the proportions with less than 43 on the SERS: 43% versus 46% (χ2 = 0.05, df = 1, p = 0.53). There was also no significant difference in the proportions scoring less than 43 among those employed (41%) and those who were currently not (46%) (χ2 = 0.13, df = 1, p = 0.46).
Self-esteem, psychological morbidity and socio-demographic attributes
We next explored the possibility that the level of self-esteem may bear some relationship to psychological morbidity, or socio-demographic features of the participants. We first compared individuals with a lifetime diagnosis of schizophrenia or schizoaffective psychosis (51%) with those with a lifetime diagnosis of affective psychosis (31%) in terms of their scores on the SERS. Even though the former group scored lower than the latter on SERS (40.4 ± 39.0 vs. 55.7 ± 32.5), this was not a statistically significant difference. In terms of current psychopathology, there was a negative correlation between the SERS and the GHQ, indicating that low self-esteem was associated with elevated level of psychological distress (r = −0.46, p < 0.001). More specifically, SERS score was significantly negatively correlated with the depression subscore on GHQ-28 (r = −0.65, p < 001). (For the entire sample, the mean score on the GHQ-28 was 2.2; 28 of the 61 (46%) scored 0). In order to identify factors associated with the level of self-esteem in this group, we next fitted a multiple regression equation in which the total score on the SERS was the dependent variable. The independent variables entered were: current age, sex, highest educational qualification, the GHQ-28 depression subscore, diagnostic grouping, course of illness (single or multiple episodes with good recovery vs. single or multiple episodes with partial recovery or continuous chronic illness), and premorbid social adjustment (yes vs. no). The result showed that, after controlling for every other variable, only the depression subscore of the GHQ-28 was associated with level of self-esteem (B = −0.532, t = − 3.48, p < 0.002).
Relationship of self-esteem to current medication experience
Most (59) of our subjects were on antipsychotic medication. The type of medication was varied, from typical to atypical antipsychotics. Most reported some form of medication side-effect. Subjects were asked to rate any perceived impairment in daily life due to medication on a 4-point scale: not at all, mild, moderate, severe, or impossible to determine. Forty-three (71%) endorsed ‘not at all’ or ‘mild’, while 11 (28%) endorsed ‘moderate’ or ‘severe’. When the two groups were compared in terms of their median scores on the SERS, there was no significant difference observed. In this sample, only 12% thought their medication was ‘not helpful’. Over 80% considered their medication to be ‘helpful’ or ‘very helpful’.
Self-esteem and quality of life
In a preliminary correlational analysis, SERS total score was significantly related to all the domains of the WHOQOL-Bref (physical health domain, r = 0.50, p < 001; psychological domain, r = 0.69, p < 0.001; social relationships domain, r = 0.44, p = 0.001; and environment domain, r = 0.55, p < 0.001). Thus, across the spectrum of scores, the lower the self-esteem, the lower the quality of life. We next examined the independent impact of self-esteem on quality of life by fitting multiple regression equations with the four domains of the WHOQOL-Bref as the dependent variables, the total score on the SERS as the independent variable and age, sex, depression subscore on the GHQ-28, and diagnostic grouping as covariates. The results are as shown in Table 1. Low self-esteem was significantly associated with an impaired quality of life in both the ‘psychological’ and ‘environment’ domains of the WHOQOL-Bref after controlling for the possible effects of gender, age, depression, and lifetime (psychosis) diagnostic status.
Discussion
Before discussing the findings of this study, certain relevant limitations need to be highlighted. First, the sample on which the findings were made was small. Even though we have examined the profile of selfesteem in a sample of community-dwelling people who have recovered from psychosis and we have selected this sample in the course of a catchment-based epidemiological survey of psychosis, we have limited selection to individuals identified by their primary clinicians to have been free of psychotic symptoms for 6 months. The resulting sample is small and may not be representative of broad populations who have recovered from psychosis. However, our sample consists of a rigorously defined group of individuals who have recovered symptomatically from a previous episode of psychosis and who, in terms of service use, represented a broad mix that may not be too different from similar groups of recovered patients. Second, even though we have assessed the sample after symptomatic recovery from psychosis, we have no way of determining the relationship of the level of selfesteem to the experience of psychosis. Thus, the level identified in a particular individual may antedate the psychotic experience and be unrelated to it.
Association of self-esteem with quality of life (n = 61)
We found that a substantial proportion of this group of individuals had levels of self-esteem that may be described as problematic [12]. As earlier indicated, we are unable to determine the relationship of the levels of self-esteem we observed to the experience of psychosis. Indeed, persons with low self-esteem might have been in that way before they experienced a psychotic breakdown. However, there is both quantitative [2] and qualitative [3] research evidence that the experience of mental illness has an adverse impact on self-esteem.
We were unable to find a relationship between employment status or medication side-effects and level of self-esteem. In regard to employment, the evidence for an association with self-esteem is inconsistent. A previous study found lower self-esteem among unemployed persons with severe mental illness, compared to employed persons [15]. Two other studies found that work status was not related to self-esteem [16], [17]. It appears that the relationship between work and selfesteem may be complex, and may require further investigation in outcome studies, taking into account factors such as nature and amount of work undertaken. In an earlier report, based on data derived from the entire LPD sample, people who described themselves as being ‘mostly dissatisfied’ with their life in the previous 12 months were significantly more likely to report more medication side-effects than those who said they were ‘mostly satisfied’ with their life [10]. However, in this sample, experience with antipsychotic medication was unrelated to level of self-esteem. Even though most of our respondents had experienced side-effects, most also had a positive view about the benefit of medication. We cannot exclude other treatment-related factors as risk factors for low self-esteem.
Multivariate analysis in which potential predictors of self-esteem were entered as independent variables revealed that only depression had an independent and significant relationship to the level of self-esteem. It is of course a common finding that low self-esteem is both a vulnerability factor for depression and also a feature of the disorder [18], [19]. While the direction of causality is therefore open to debate [20], a co-occurring depression may presumably provide the full explanation for low self-esteem. Indeed, self-esteem and affective symptoms were strongly correlated in a recent study of adults with severe mental illness who were receiving vocational rehabilitation [17]. It is noteworthy that, once level of depression was accounted for, gender was unrelated to level of self-esteem. Along with depression, gender is another variable with which self-esteem, and specifically scores on the SERS, may interact [11], [21], [22]. Level of self-esteem was also unrelated to premorbid social functioning or with the course of the psychotic illness. Thus, there was no indication to suggest that low self-esteem in this group of recovered patients was a continuation of a pattern of premorbid social difficulties or was explainable on the basis of a poorer course of illness.
A number of features relevant to the experience of psychosis could be speculated as potentially detrimental to self-esteem. Probably the most ego-threatening is the stigma that accompanies the experience of a psychotic illness [23]. Beck has suggested that an individual's level of self-esteem is determined, in part, by judgements made by others [24]. Ingham and colleagues have also shown that life circumstances may influence feelings of selfesteem independent of psychiatric illness [25]. Features such as the relative poverty experienced by many people who are recovering from psychotic illness may be relevant here. As an evaluative aspect of self-concept, with both cognitive and affective components [11], [26], self-esteem might indeed be affected by any of these features. Our study suggests that if these features were at play in our sample, they probably acted through the medium of depressive symptoms and feelings.
Whatever the factors influencing it, self-esteem was related to quality of life. For the entire group, the lower the self-esteem the worse the subjective quality of life. This is consistent with a reported correlation between lower self-esteem and lower life satisfaction as assessed by Lehman's Quality of Life Interview [17], [24], [27]. Specifically and after controlling for confounding factors, low self-esteem was an independent risk factor for impairment in the domains of psychological health and the environment on WHOQOL-BREF. Our observation of an association between low self-esteem and quality of life did not permit us to infer a causal relationship, or even to rule out another intervening variable. For example, the often-reported association between poor social relations and social life on the one hand and quality of life on the other among those with severe mental illness [28], [29], may indeed be a proxy for the consequence of low self-esteem. Further, Torrey et al. [17] interpreted their findings of stability of self-esteem scores over time to indicate that global self-esteem is a relatively stable personality trait. This interpretation cannot be ruled out within our study. What the present findings do suggest is that while symptomatic relief might have been achieved, low self-esteem might compromise expected gains in quality of life with individuals having a negative ‘perception of their position in life in the context of the culture and value systems in which they live’ [13]. This suggests that assessment of self-esteem be conducted as part of outcome evaluations [30], [31]. To the extent that low selfesteem is remediable [32], this observation challenges professionals who care for persons with psychoses. Remediation interventions, such as cognitive-behaviour therapy are effective in improving psychological health and success at job-seeking among the long-term unemployed [33], by improving self-esteem and may provide similar benefits to those with previous psychotic illness in whom cognitive behaviour techniques are used to modify dysfunctional assumptions about the self [34].
Footnotes
Acknowledgements
The study was funded by the Commonwealth Department of Health and Aged Care (Brisbane, Melbourne and Perth) and by local sources (Canberra).