Introduction

There is a gender gap in mental health service use. Men make fewer visits to mental health professionals than women [1, 2]. These gender differences are particularly pronounced with regard to general care, but decrease when it comes to specialist or residential care [3]. This gender gap in mental health service use cannot be explained by need [4]. There are no gender differences in the overall rates of psychopathology [5], yet men and women differ in types of symptoms. Women are more likely to internalize [6]; men are more likely to report externalizing disorders [7]. This study therefore will examine why men engage less in the use of mental health services. Explanations for gender differences have often been posited, but seldom tested empirically [2, 4, 8].

Following the doing gender perspective [9], we will look for an explanation by studying how gender is shaped in dyadic interactions. This social constructionist perspective emphasizes that gender differences do not reside in the individual, but are actively (re)produced in social transactions [10]. Men and women think and act in the ways they do not because of role identities or psychological traits but because of the adopted cultural conceptions about femininity and masculinity. People share common values about what behavior is considered appropriate for men and women [11]. During social interactions, people constantly engage in (re)creating a sense of gender difference. People know they will be judged in terms of their success or failure in meeting gendered societal expectations [9]; therefore they try to match those expectations. In other words, gender is negotiated [12].

In the context of help seeking for mental illness, the normative expectations guiding female and male behavior are of crucial importance [13]. Doing health is a form of doing gender [14]. Emotional expressiveness, caring for one’s health, and asking for help are constructed as forms of idealized femininity [15]. Men are expected to be in emotional control and to appear strong, independent, and self-reliant [4]. Consequently, to conform to the socially prescribed male role, men are encouraged to define themselves in opposition to women by suppressing their own health needs and by not seeking help. Otherwise, a gendered role conflict might arise if men behave in a way that is not in accordance with the (hegemonic) idea of masculinity [15].

Because of this gendered role conflict, men may be reluctant to seek professional care and may fear greater stigma if they do seek help. With regard to treatment, men are more likely than women to choose to deal with mental illness on their own and to rely on self-care options [16], in part because they do not want to end up in a subordinate relationship to a health care provider [15]. If they do seek professional treatment, men seem to prefer a quick solution, one that is not time consuming [17]. Consequently, we suggest that men more often opt for medical treatment instead of psychotherapy than women do. Men also report greater stigmatizing attitudes because of the role conflict between the male gender role and the patient role [18]. Stigma can be related to viewing personal weakness as the cause of mental illness [19]. Hence, when men disclose mental health problems, they may feel that they have failed and feel ashamed of their so-called weakness [20]. We suggest that men therefore attribute more blame and shame to mental illness.

We control for a range of other socio-demographic variables. Older people are more inclined to use general medical care [21] and to report more stigmatizing attitudes [22]. People with a higher level of education are more likely to contact mental health specialists [23], and are less likely to stigmatize people with mental illness [24]. Further, we control for marital status and employment status. In addition, it is important to consider whether respondents are familiar with mental health services, as this might influence treatment recommendations [25] and stigmatizing attitudes [26]. Last, but most important, we control for labeling, since women are known to have higher levels of mental health literacy and are also more likely to be labeled as deviant [27].

In sum, we will examine why men engage less in mental health service use by studying how gender is performed in interactions, following the doing gender perspective. We hypothesize that seeking help for mental illness might constitute a role conflict among men. We expect that men will recommend reliance on self-care to other males. And when they do prefer professional treatment, we assume that they will opt for medication to deal with it rather than discussing their problems in psychotherapy. Moreover, we hypothesize that men will report greater stigmatizing attitudes toward people in general and males in particular who suffer from mental illness.

Methods

Sample and data

Our data are derived from the survey Stigma in a Global Context-Mental Health Study (SGC-MHS), the first cross-national survey designed to examine attitudes toward mental health services and people with mental illness. Fully structured, face-to-face computer-assisted personal interviews were implemented among representative samples of non-institutionalized adult populations. We obtained approval for the ethical aspects related to our research from the Privacy Commission.

Part of the interview consisted of questions referring to a vignette about a hypothetical person. The vignettes specifically selected for this study consist of unlabeled psychiatric case histories with symptoms that fulfill the criteria of the DSM-IV (Diagnostic and Statistical Manual of Mental Disorders of the American Psychiatric Association) regarding major depression or schizophrenia. In addition, socio-demographic variables are assessed by means of the ISSP indicators (International Social Survey Program).

To define the target population, we used a multistage cluster sampling design based on the data from the Belgian National Register. In stage 1, municipalities were weighted according to their number of inhabitants and 140 of them were selected randomly (with the possibility of being selected more than once), using the statistical program SPSS 19. In stage 2, the Belgian National Register data provided us with a random sample of 15 respondents within each of the 140 selected municipalities. This resulted in a target sample of 2100 people. After complete description of the study to the participants, written informed consent was obtained from 1166 respondents in total. We followed the guidelines of the American Association of Public Opinion Research [28]. The response rate was 55.5 % (AAPOR Response Rate 1) and the cooperation rate was 66.9 % (AAPOR Cooperation Rate 3). Additionally, a post-stratification weight procedure was applied in SPSS 19 to approximate the cross-classification of the census population count within gender, age, and education. As a result, our weighted sample consisted of 743 respondents.

Variables

Independent variables

The gender of the respondent and the vignette person are considered in interaction, since we made the distinction between a male respondent who judges a male vignette, a female respondent who judges a female vignette (same gender), and a male respondent who judges a female vignette and a female respondent who judges a male vignette (cross-gender).

Dependent variables

Recommending self-care was assessed by asking for spontaneous treatment suggestions. This question was asked immediately after the description of the vignette had been read out by means of the following: ‘What should the vignette person do if he or she needs to do something?’ The respondents could offer three suggestions at the most. The responses were assigned to a range of predefined categories by the interviewers and if a response did not correspond with one of the predefined categories, it was coded verbatim. The verbatim responses were subsequently recoded independently by two raters through a content analysis procedure (if the raters disagreed, they were reconsidered). One of the resulting categories considered self-care options. In the end, a dichotomous variable referring to ‘recommending self-care’ was created.

Attitudes toward psychotherapy and tranquilizers were assessed by asking the respondent whether they thought that certain treatment options were helpful. The answering categories were situated on a 7-point Likert scale, ranging from very harmful to very helpful.

Stigmatizing attitudes were assessed by means of the stigma dimensions ‘blame’ and ‘shame’. Blame refers to a person attributing the situation of the vignette subject to a weak character or lack of willpower. The original 4-point Likert scale was dichotomized. Shame refers to the opinion that the vignette person should be embarrassed about his/her situation. This original 4-point Likert scale was also dichotomized.

Control variables

‘Age’ is measured as a continuous variable. ‘Education’ is measured as a categorical variable, referring to either primary, lower secondary, higher secondary, or tertiary education. ‘Marital status’ refers to being married or cohabiting, separated, divorced, widowed, or single. ‘Employment status’ compares the employed population with those who are unemployed, retired, or who are students or homemakers. The ‘level of familiarity’ with mental health services is categorized as follows: (1) personal experience (respondents have had psychiatric treatment themselves), (2) interpersonal contact (respondents know a family member, friend, or acquaintance who has received psychiatric treatment), and (3) no (inter) personal contact. Furthermore, to assess labeling, the following question is used: ‘How likely is it that the vignette person is experiencing a mental illness?’ This item is scored on a 4-point Likert scale. A higher score expresses that people are more likely to agree that the situation refers to mental illness. In addition, we also control for the in-group/out-group status of the vignette. For example, in Belgium and Germany, the out-group status was given the Turkish nationality, as they represent one of the largest ethnic minority groups. If the vignette subject suffered from major depression, he/she receives the score 1; symptoms of schizophrenia constitute the reference category.

Analysis

First, the descriptives of the whole study population and the dependent variables are illustrated in Table 1. Second, in Table 2, the gendered interactions between the respondent and the vignette person are linked to treatment recommendations. Finally, the gendered interactions are associated with stigmatizing attitudes in Table 3. With regard to recommending self-care and the stigma dimensions blame and shame, logistic regression models are estimated. The odds ratios are shown, accompanied by their 95 % confidence interval and p values. Concerning the helpfulness of psychotherapy and tranquilizers, linear regression models (Ordinary Least Squares) are estimated and the unstandardized regression coefficients are reported, accompanied by their standard errors and p values. In case of all models, whether logistic or linear, all independent core variables and all control variables were entered as a block. All models mentioned above are estimated using the statistical program IBM SPSS Statistics 19.

Table 1 Descriptives of the study population and the independent variables (N = 743, weighted data, SGC-BMHS, 2009)
Full size table
Table 2 Gendered interactions related to treatment recommendations (weighted data, SGC-BMHS, 2009)
Full size table
Table 3 Gendered interactions related to stigmatizing attitudes (N = 738, weighted data, SGC-BMHS, 2009)
Full size table

Results

The descriptives of the study population and independent variables are presented in Table 1.

A first potential explanation of why men engage less in mental health service use is that they prefer self-care options. The results in Table 2 regarding treatment recommendations show that the subjects of the male vignettes are more likely to be advised to rely on self-care options, both by male and female respondents. A second hypothesis regarding help-seeking attitudes was that if professional treatment would be preferred, men would opt for medication rather than psychotherapy. Our findings indicate that male respondents are less likely to acknowledge the helpfulness of psychotherapy, both for male and female vignettes. Moreover, female respondents are less inclined to rate psychotherapy as a helpful treatment option when considering a male vignette versus a female vignette. Additionally, male respondents are more likely to rate tranquilizers as helpful when making recommendations for a male vignette, compared to female respondents making a similar recommendation for a female vignette.

Furthermore, we examined men’s stigmatizing attitudes. Table 3 shows that male respondents seem to consider the situation of male vignettes more shameful than female respondents considering the situation of female vignettes. Also, male respondents seem to attribute more blame to female vignettes than female respondents attribute to female vignettes.

In addition, the control variables reveal that older people are more inclined to rate tranquilizers as helpful and that higher educated people are less predisposed to recommend self-care options. People who have been in psychiatric treatment perceive tranquilizers to be more helpful and report less stigmatizing attitudes than those who have not. Next to this, people who label the situation of the vignette subject as mental illness are less likely to recommend self-care, while they are more likely to acknowledge the helpfulness of psychotherapy. With regard to stigmatizing attitudes, people who label the situation as mental illness are more likely to agree that the vignette person should be embarrassed, but there was no significant association between labeling and blame. Finally, when people have to judge someone with an out-group instead of an in-group status, they seem to be less inclined to recommend self-care.

Discussion

Research has consistently found that men report less mental health service use [1, 2]. However, the reasons for this gender gap remain unclear. This study makes a contribution to the research field by empirically testing a range of potential explanations. We start out from the doing gender perspective [9] and suggest that gender differences are due to the role conflict that men may experience when seeking help for mental illness, resulting in differences in treatment recommendations and stigmatizing attitudes. This study is carried out using data from the SGC-MHS survey, which questioned a representative sample of the general population in Belgium.

The first strength of this study is that a representative sample of the Belgian sample is examined; other research examining the theme of gender differences in mental health service use often relies on selective samples, such as a student sample [29]. Second, although the interaction between the respondent and the vignette subject is not a real social interaction, the advantage of this approach is that many cases can be questioned through randomly varying the gender of the vignette subject. Third, in particular, gender differences in stigmatizing attitudes have received relatively little attention. When they were studied, stigma dimensions such as social distance tended to be the focus, which measures the perception that men exhibit more dangerous behavior and therefore explains why men experience stronger rejection [8], but does not capture the gendered role conflict.

The main finding of the current study is that not only men ‘perform’ gender in the sense that they report more negative attitudes toward professional help seeking, but that also women (re)construct masculinity norms by giving different treatment advice to a man compared to another woman. Men report less positive attitudes toward the helpfulness of psychotherapy and advise male vignettes to rely on self-care options. Yet women also contribute to the maintenance of the role conflict between the male role and the patient role by rating psychotherapy as less helpful for men and by recommending self-care to men. It is striking that although women are more psychologically minded [30], their acknowledgment of the helpfulness of psychotherapy depends on the gender of the person they are giving advice to.

Furthermore, men seem more likely to rate tranquilizers as a helpful option for other men in comparison to women who gave advice to other women. This finding is in line with the study by Ettorre et al. [31] that showed that the lay referral system was an important channel of introduction to psychotropic drug use, especially for men. One potential explanation is that men just want to be able to continue in their work role, so drugs may provide a quick solution [14, 17].

Additionally, our results concerning stigma indicate that men seem to attribute more shame and blame to mental illness than women do. Men consider suffering from mental illness as more shameful than women do, particularly when a male vignette subject is dealing with mental illness. This is consistent with the finding of an older study by Phillips [32] that indicated that people more strongly rejected cases that had to do with men rather than women. Shame can also be perceived as a proxy for self-stigma [33]; thus this finding is consistent with the study by Pederson and Vogel [18]. Additionally, male respondents seem to ascribe more blame to female vignette subjects, compared to women judging other female vignettes.

This has implications for both informal and formal care. Women in general pay attention to not only their own health, but also to the health of their partner and children [34]. They are crucial source of lay referral. Yet our findings show that masculinity norms also seem to influence women’s mental health treatment recommendations in a detrimental way. Regarding formal care, empirical research has shown that the gender of the patient influences the amount and type of treatment provided by physicians, independent of objective symptoms [35]. Clinicians were less likely to identify the presence of depression in men [36], resulting in fewer prescriptions for psychotropic drugs for men compared to women [37].

Finally, we address a limitation of the study and give a suggestion for future research regarding this theme. We acknowledge that we did not explicitly consider people’s adherence to traditional masculinity norms. Multiple beliefs about masculinity exist that depend, among other things, on socio-economic status and ethnicity [38]. Previous research pointed out that those who score higher on measures of traditional masculinity ideology seem to have more negative attitudes toward help seeking for mental illness [39]. We argue that future research should link this research question to substance abuse, such as alcohol disorder or drug use, since Rosenfield [40] suggested that men experience less rejection when suffering from externalizing disorders that are more in line with stereotypical masculine behavior.