Introduction

Unemployment and mental illness often co-occur [1,2,3] and may interact in a vicious circle, with mental health problems being cause or consequence of unemployment [4]. On one hand, unemployment can result in psychological distress [5,6,7]; on the other hand, there is evidence from longitudinal studies that poor mental health causes unemployment [8] and premature retirement [4, 9]. However, unemployed people with mental health problems often choose not to use mental health services or job-seeking support [10] and therefore do not benefit from available psychosocial therapies [11] or supported employment [12]. Besides other factors such as mental health literacy, the stigma associated with unemployment and with mental illness can affect help- and job-seeking behaviors [10].

Mental illness stigma is common and describes a process that involves labeling, stereotypes, separation, status loss, and discrimination [13, 14]. There are three forms of stigma that can be barriers to help-seeking: public stigma, self-stigma and structural stigma [15]. The latter involves macro level units rather than individuals, for example increased waiting times as a consequence of limited funding for mental health services. Public stigma (i.e. stereotypes, prejudice and discrimination among members of the general public) and self-stigma (if people with mental illness agree with negative stereotypes and turn them against themselves) additionally contribute to low help-seeking behavior [16]: People with mental illness may avoid treatment in order not to be labeled “mentally ill” by others; and self-stigma or shame can undermine the motivation to seek help [17].

Unemployed people with mental health problems may additionally experience everyday discrimination associated with their unemployment status. The general public often associates unemployment with incompetence [18], and especially young people tend to blame unemployed individuals, unless own experience of unemployment weakens these prejudices [19]. Unemployment can result in loss of self-esteem and in social withdrawal [10]. This can lead to reduced job search self-efficacy [20, 21], the belief that one can successfully seek and find a job [22].

Nevertheless most previous studies among people with mental health problems have looked at mental illness stigma without considering other social conditions that may be stigmatized such as unemployment [23]. Multiple stigma might increase distress among affected individuals and limit the effectiveness of existing help-seeking interventions. One concept reflecting the implications of multiple stigma is intersectionality. The idea was initially developed within feminist psychology to describe how overlapping stigmatized social identities affect the level and quality of oppression and disadvantage experienced by African American women [24]. Accordingly, intersectionality highlights the need for considering intersections of social identities (e.g. gender, age, sexual orientation, obesity) as well as related systems of discrimination to comprehensively understand stigma [25, 26]. According to the double disadvantage hypothesis [27] individuals with more than one disadvantaged status may experience worse outcomes than their singly disadvantaged counterparts. However, findings have been inconsistent. In a recent population survey, von dem Knesebeck and colleagues [28] found no evidence for increased stigma towards people with depression who were of low socio-economic status or immigrants as compared to high socio-economic status persons and non-migrants. On the other hand, Grollman et al. [27] observed negative cumulative health effects among persons experiencing multiple forms of discrimination.

In terms of multiple stigma among unemployed people with mental health problems, the consequences for job search and help-seeking behavior are unclear. In a large English study among people using health services, employed participants reported significantly lower levels of experienced mental health-related discrimination in different areas of life compared to unemployed individuals [29]. O´Donnell et al. [30] tested the impact of anticipated social discrimination on psychological distress and somatic symptoms in a sample of unemployed individuals and found that participants with higher anticipated unemployment-related discrimination reported greater distress as well as more physical health problems.

Because individuals who experience multiple discrimination may encounter greater barriers to job- and help-seeking, associations between multifactorial discrimination and mental health inequities [31] as well as discrimination as a barrier to health utilization matter [32]. To develop adequate services and to increase help-seeking among unemployed individuals with mental health problems, a better understanding of characteristics and consequences of multiple discrimination due to unemployment and mental health problems is needed. In this study, we therefore focused on two attributions: (1) the attribution of experienced discrimination to unemployment; and (2) the attribution of experienced discrimination to mental health problems. We expected that individuals who attribute discrimination to mental health problems as well as to unemployment would experience worse job- and help-seeking outcomes compared to those participants who do not attribute discrimination to both characteristics.

Methods

Study design and participants

During the AloHA project on unemployment, mental health problems and help-seeking [33, 34], 301 participants were recruited outside healthcare settings and mainly from unemployment agencies in Southern Germany. Inclusion criteria were age between 18 and 64 years, current unemployment and sufficient German language skills. Another inclusion criterion was a score ≥ 13 on the K6 Psychological Distress Scale [35] or a score ≥ 1 on items 2–4 of the CAGE-AID screening tool for current alcohol and substance-use disorders [36]. For the sake of specificity, we excluded item 1 (cut down) from calculating CAGE-AID scores because in previous studies nearly half of healthy controls endorsed that item [37]. In addition to fulfilling either the K6 or CAGE-AID criterion, a score of ≥ 17 (range of possible scores: 12–60) on the 12-item WHO-Disability Assessment Schedule 2.0 [38] was needed, corresponding approximately to the 85th general population percentile and to the average disability level of persons with one mental disorder [39]. We therefore only included participants with significant illness-related disability. Persons were excluded if they worked more than 14 h per week, earned more than €450 per month (above this limit, social insurance contributions must be paid in Germany), or received full disability pension. Participants were on average 44 years old (M = 43.7, SD = 11.2) and about half were female. The average length of lifetime unemployment was about 5 years (M = 63.2 months, SD = 56.4).

Measures

Experienced discrimination was assessed by the 9-item Everyday Discrimination (EDD) Scale [40]. Participants rated the extent to which they experienced various forms of discrimination in their everyday lives on a scale from 1/never to 6/almost every day, e.g. “being treated with less courtesy than others or being threatened or harassed”. An EDD sum score was calculated as the sum of all items (range of possible sum scores 9–54; Cronbach's alpha in this sample α = 0.88). Due to participants´ experiences of being unemployed and mentally distressed, individuals were asked to rate their agreement with two statements on 7-point Likert scales (1/strongly disagree to 7/strongly agree): (EDDa) “I am treated like this because I am unemployed” (M = 4.0, SD = 2.2) and (EDDb) “I am treated like this because I am mentally ill/I am mentally distressed” (M = 3.4, SD = 2.0). Based on these two attribution scores, we built subgroups of discrimination attribution to mental health problems and/or to unemployment. For both statements, we classified attribution of discrimination for anyone who scored above the midpoint (> 4) of the attribution scale. Excluding those respondents who did not report any form of experienced discrimination in their everyday lives (EDD sum score = 9, n = 12), this resulted in four subgroups: group i (n = 71) attributed experienced discrimination to both unemployment (EDDa: M = 6.0, SD = 0.8) and mental health problems (EDDb: M = 6.0, SD = 0.8), group ii (n = 128) neither to unemployment (EDDa: M = 2.1, SD = 1.2) nor to mental health problems (EDDb: M = 2.1, SD = 1.1), group iii (n = 70) mainly to unemployment (EDDa: M = 5.9, SD = 0.9; EDDb: M = 2.5, SD = 1.1), and group iv (n = 20) mainly to mental health problems (EDDa: M = 2.6, SD = 1.3; EDDb: M = 5.7, SD = 0.7).

Job search self-efficacy was assessed with an established 6-item measure of job search self-efficacy [41, 42] and respondents rated their confidence to engage in several job search activities, e.g. “making the best impression in an interview”, from 1/not at all confident to 5/a great deal confident, yielding a job search self-efficacy mean score (Cronbach's alpha in this sample α = 0.86). Perceived barriers to seeking help for mental health problems were assessed by the 30-item Barriers to Access to Care Evaluation Scale (BACE [43]). Participants were asked to rate potential barriers on a 4-point Likert scale (1/not at all – 4/a lot). Two mean scores for perceived barriers to help-seeking for mental health problems were calculated; (1) stigma-related barriers (12 items, e.g. “Concern that I might be seen as weak for having a mental health problem”, Cronbach's alpha in this sample α = 0.94) and (2) not stigma-related barriers (18 items, e.g. “Wanting to solve the problem on my own”, Cronbach's alpha in this sample α = 0.91). The Self-Appraisal of Illness Questionnaire (SAIQ [44]) was used to assess perceived need for treatment (6 items, e.g. “I think my condition requires psychiatric treatment”, Cronbach's alpha in this sample α = 0.86) and higher perceived presence of illness (4 items, e.g. “How ill do you think you are?”, Cronbach´s alpha in this sample α = 0.71).

Statistical analyses

We excluded respondents who did not report any form of experienced discrimination in their everyday lives (EDD score = 9, n = 12), so that only 289 participants were included in the final analyses. First, multiple linear regressions on job search self-efficacy, BACE stigma- and not stigma-related barriers, and on need for treatment as well as presence of illness (SAIQ) were calculated (Table 1). Independent variables included level of everyday discrimination, attribution of discrimination to unemployment (EDDa, yes vs. no, as defined above), attribution of discrimination to mental health problems (EDDb, yes vs. no, as above). All regressions were adjusted for age, gender and length of lifetime unemployment. In an additional step and to examine possible interaction effects between attribution of discrimination to unemployment and to mental health problems, we repeated all linear regressions with the interaction term of both attributions (EDDa * EDDb) as additional independent variable. Second and to test for differences in job- and help-seeking behaviors associated with attributions, we compared four subgroups based on the attribution of experienced discrimination to unemployment and to mental health problems (group i), neither to unemployment nor to mental health problems (group ii), mainly to unemployment (group iii) and mainly to mental health problems (group iv). The four subgroups were compared regarding job search self-efficacy, perceived barriers to accessing mental health care, need for treatment and presence of illness using analyses of variance (ANOVA; Table 2) and Scheffé tests for post hoc comparisons. Because both attribution variables (EDDa, EDDb) showed skewed distributions in the subgroups, we repeated the subgroup comparisons with the Kruskal–Wallis test [45]. The significance level was set to p < 0.05 and SPSS 21 was used for all analyses.

Table 1 Five multiple linear regressions on job search and help-seeking outcomes
Full size table
Table 2 Analyses of Variance for four EDD attribution subgroups for job- and help-seeking outcomes
Full size table

Results

In multiple linear regressions, adjusted for age, gender and length of lifetime unemployment, better job search self-efficacy was associated with lower levels of everyday discrimination, less attribution of discrimination experiences to mental health problems and shorter unemployment, explaining about one-ninth of job search self-efficacy variance (Table 1). More perceived stigma-related barriers to care were related to more everyday discrimination and a greater attribution of discrimination to both unemployment and mental health problems; the regression model explained about a third of variance in stigma-related barriers. A higher level of not stigma-related barriers was predicted by more discrimination and a stronger attribution of discrimination to unemployment, explaining about one-sixth of not stigma-related barrier variance. More perceived need for treatment and higher presence of illness were associated with less attribution of discrimination to unemployment and more attribution to mental health problems. When repeating the regressions with the interaction term of both attributions as additional predictor variable, the interaction term was not significant throughout (all p values > 0.12, data not shown).

In ANOVAs and post hoc comparisons using the Scheffé test (Table 2), people who attributed discrimination to both unemployment and mental health problems (group i) showed significantly lower job search self-efficacy than participants who attributed discrimination mainly to unemployment (group iii). Subgroups differed with respect to perceived stigma-related barriers to help-seeking, i.e. participants who attributed discrimination to both unemployment and mental health problems (group i) perceived significantly more stigma-related barriers to help-seeking than those who attributed discrimination mainly to unemployment (group iii) and more than those who attributed discrimination to mental health problems (group iv). No significant differences in not stigma-related barriers between multiple (group i) and single discrimination attribution groups (groups iii and iv) were found. People who attributed discrimination to both unemployment and mental health problems (group i) perceived significantly more need for treatment and higher presence of illness than participants who attributed discrimination mainly to unemployment (group iii). When repeating the subgroup comparisons with the Kruskal–Wallis test and Bonferroni correction for multiple tests, all post-hoc comparisons that were significant in the ANOVAs remained significant.

Discussion

While there is evidence that multiply disadvantaged individuals are more likely to report mental health problems than their singly disadvantaged counterparts [46], most empirical studies have looked at members of ethnic or sexual minorities or obese individuals [7] without considering stigma against unemployed people with mental health problems. Supporting our hypothesis, individuals who attributed experienced discrimination to mental health problems as well as to unemployment experienced worse job- and help-seeking outcomes.

Our finding that the experience of multiple discrimination is associated with reduced job search self-efficacy is consistent with the notion that the stigma associated with unemployment and mental health problems contributes to long-term unemployment. In a Romanian study job search self-efficacy was negatively associated with anxiety, possibly due to the threatening nature of unemployment and of unsuccessful job search [47]. In our study, unemployment stigma has likely been relevant as it was conducted in Southern Germany, a region currently characterized by a low unemployment rate of about 3% [48]. Under these conditions, the general public may show less understanding for the unemployed, and we could speculate that stigma and discrimination against unemployed individuals might be increased, undermining their confidence to find a job [22].

Stigma-related barriers to treatment were predicted by attributions of discrimination to both unemployment and to mental health problems. This is consistent with the double disadvantage hypothesis that multiple stigma results in additive or cumulative effects [27]. On the one hand people with mental health problems may avoid treatment in order not to be labelled “mentally ill” [17]. On the other hand O´Donnell et al. [30] showed that unemployed participants who experienced higher anticipated stigma because of their unemployment reported increased psychological distress. A new unemployment-related stigmatized identity could lead to withdrawal from support systems [10] that would otherwise help to cope with unemployment and mental health problems.

Participants who attributed discrimination to unemployment and to mental health problems reported more need for treatment and presence of illness than people who attributed discrimination to unemployment. This finding is in line with the double disadvantage hypothesis and results of Grollman [27] that respondents who held more than one disadvantaged status were more likely to experience distress compared to their singly disadvantaged counterparts. Stigma-related stress may be an important mediator of the relationship between discrimination and health among members of multiple minority groups [31, 49]. Vice versa, participants with higher symptom levels may have self-identified as having a mental health problem and therefore may have perceived more need for treatment as well as more illness-related discrimination.

Limitations and future research

Limitations of our study need to be considered. Rather than measuring specific examples of discrimination associated with unemployment or mental health problems, participants rated the extent to which they attributed general everyday discrimination to their unemployment and mental health problems. Due to our sample characteristics we could not examine the role of ethnic minority status. Our cross-sectional data preclude conclusions on causality. The lack of evidence for an interaction of attributions might be related to the fact that not all participants self-identified as having a mental health problem (especially those not receiving mental health care at the time of the study) and not all may have considered unemployment as an important element of their identity (especially those who had been employed until very recently and expected fast reemployment, or those with part-time employment who earned 450 Euros/week or less).

Despite these limitations, our findings highlight the need to consider multiple stigma among unemployed people with mental health problems. Future studies could focus on diverse social structures and conditions to recognize the complex characteristics of stigma, including ethnic minority or migrant status, education and gender. Considering the potential interaction between mental health problems and unemployment, future studies of multiple discrimination should collect qualitative data, e.g. by individual or focus groups interviews.

Implications to reduce multiple discrimination

Whether existing interventions to reduce mental illness stigma can effectively reduce multiple forms of discrimination is a topic for future research. Programs to reduce unemployment-related stigma seem to be lacking. Our results call for interventions to reduce multiple stigma and discrimination against unemployed people with mental health problems on an individual, structural and public level. Because unemployed people with mental health problems can suffer from self-stigma and shame, undermining help-seeking and job search motivation, they should be supported in their coping with mental illness-related as well as unemployment-related stigma. One approach could be Job Club group interventions that increase personal control, self-esteem and job search self-efficacy [50] and offer the opportunity of peer support to cope with stigma.

Because intersectionality refers to the interdependence between social identities and structural inequities [46], there is a need to offer support for people who suffer from multiple stigma such as supported employment which includes employment activities based on individual preferences and needs, integration of employment services into mental health services and personalized benefits planning [12]. Due to the fact that the concept of supported employment is not routinely implemented in the German health care or employment agency systems, policy changes should be considered. Additionally, contact-based interventions [26, 51] could reduce multiple stigma and attitudes and behaviors among members of the general public.