- Split View
-
Views
-
Cite
Cite
Ioana Pop, Femke Roosma, Peter Achterberg, Detraditionalization, mental illness reports, and mental health professional care use in Europe, European Sociological Review, Volume 39, Issue 4, August 2023, Pages 532–544, https://doi.org/10.1093/esr/jcac077
- Share Icon Share
Abstract
In this study, we address the question of whether individuals that live in more detraditionalized countries have higher levels of mental illness and mental health professional care use. We argue that it is meaningful to consider the different facets of detraditionalization, that is the level of secularization, the ethos of personal autonomy, and self-realization, the erosion of traditional gender roles when understanding patterns of mental illness reports and mental health professional care use. We use data collected in 2010 in 25 European countries by Eurobarometer and find that, generally speaking, people living in more detraditionalized countries are more inclined to use mental health professional care, and that they, on average, report less mental illness than people in less detraditionalized countries. Furthermore, not all forms of adversity result in higher levels of mental health professional care use in the more detraditionalized countries. This is the case only for those experiencing financial strain while for those experiencing unemployment or divorce this was not the case. Furthermore, in more detraditionalized countries, the experience of divorce was related to fewer mental illness reports, a result that could be linked to processes such as the erosion of the traditional institution of marriage and the normalization of divorce in these societies.
Introduction
Detraditionalization encompasses broader institutional and cultural aspects taking place in late-modern western societies (Elchardus, 2009). The phenomenon of detraditionalization has emerged principally as a result of the shift from scarcity to economic security, a shift that came about due to increasing wealth and generous welfare states (Inglehart, 2018). Following up, several processes unfolded: religion lost its role as a moral compass, traditional roles, especially within the family and between sexes, eroded, and a decline in the acceptance of external authority was observed. The result of this process is the rise of the ‘autonomous, self-directing, self-realizing individual’ (Elchardus, 2009: p. 152). These are not new ideas: similar arguments are also found in the literature mapping cultural processes of individualization: late-modern societies are characterized by increased de-standardization of the life course, the expansion of life choices with unpredictable outcomes, and the increased responsibility put on individuals for their lifestyle choices (Branaman, 2007; Inglehart, 2018). All these changes, scholars argue, have profound implications for mental health, and have also resulted in the emergence of psychology as a social institution (Rieff, 1966; Elchardus, 2009).
Regarding secularization as the first facet of detraditionalization, scholars have argued that religious beliefs and ideas have been slowly eroded by a psychological view of the world, a vision in which the soul was replaced by the psyche (Lasch, 1979; Moskowitz, 2001; Rakow, 2013). As the evidence-based model of reality propagated by science gained more acceptance against Christianity’s genesis model, the possibility of salvation in the afterlife was replaced with the mundane reality of here and now, and modern man was faced with the task of finding happiness and self-realization in this world. Scholars argue that psychological disciplines have emerged to help people to cope with the task of navigating the complex and fast-changing contemporary world (Illouz, 2008; de Wachter, 2012) and therapists have become the equivalent of modern-day priests (Rieff, 1966). While scholars have traced the diffusion of psychological ideas and practices in everyday life (Berger, 1965) through the media, on the work floor, or the educational setting (Moskowitz, 2001; Illouz, 2008; Elchardus and De Keere, 2010), this is not to say that the language of psychology has replaced religion; ‘the intertwinement of therapeutic approaches and religious/spiritual concepts and practices in the New Age spirituality’ (Rakow, 2013: p. 494) suggest that this relationship is most likely bidirectional, resulting in new forms of practices at the intersection of psychology and religion.
Detraditionalized societies are also characterized by the ethos of personal autonomy and self-realization, which is closely related to the decline in the acceptance of external authority and the erosion of traditional roles (Elchardus, 2009). Facilitated by the expansion of mass media, psychological research on topics such as infant development, childrearing practices, or love and relationships has spread in society at large (Berger, 1965) and has paved the way toward a critical discussion and transformation of family and intimate relationships (Benjamin, 1988; Illouz, 1998; Perel, 2007; Ryan and Jetha, 2011). An example comes from studies that documented the role of counseling and psychology in the public recognition of family violence, which was reflected in the public attitudes regarding ‘disciplinary’ practices toward women and children (Benjamin, 1988; Wright, 2008). In turn, this could have facilitated the erosion of traditional authoritarian roles related to genders and generations within the family.
The above arguments depict an image of late-modern societies where the institution of psychology is influenced and influences back facets of detraditionalization. Furthermore, the institutional and cultural aspects encompassing processes of detraditionalization as well as the associated diffusion of the psychological model in society are argued to have had a deep impact on how we think and talk about mental illness and how we approach looking for mental health care (Horwitz, 2011). For instance, Wright (2008) points out that increased psychological knowledge and understanding have led to the public recognition of suffering. The author argues that the experience of suffering was traditionally confined to the private sphere but the emergence of a culture of disclosure has created an opportunity for expressing pain publicly and for legitimizing the quest for physical and psychical safety and well-being.
The immediate question is whether individuals living in more detraditionalized societies will differ in their accounts of mental illness1 and mental health professional care use, in comparison with individuals living in more traditional societies. The scholarly work in this field is limited, mostly theoretical, or based on qualitative methodology (Furedi, 2004; Wright, 2008; Anderson, 2009). This is precisely where this study aims to make a first contribution, that is by taking a cross-country comparative approach we acknowledge that processes of detraditionalization differ between late-modern societies, with potential implications for mental illness reporting and mental health professional care use.
Furthermore, we also examine the potential differential effects of detraditionalization for specific groups who are confronted with social or material adversity. Our choice for examining adversity rests on two reasons. First, the literature shows that adversity is related to worse mental health and higher mental health professional care use (Pearlin et al., 2005; Mandemakers, Monden and Kalmijn, 2010; Buffel, Beckfied and Bracke, 2017). Second, the pressure to do and be everything combined with the discourse of equal opportunities and meritocracy is argued to result in the centrality of individuals as the singular responsible and master of their lives (de Wachter, 2012; Lasch, 1979). Thus, adversity is more likely to be interpreted as a result of own’s fault in more detraditionalized countries and consequently, could result in different patterns of mental illness reporting and mental health professional care use.
The research questions that guide our study are the following: (i) are there differences in the level of mental illness symptoms reports and mental health professional care use between European countries with different levels of detraditionalization? and (ii) does the relationship between social and material adversity and level of mental illness symptoms reports and mental health professional care use differ between European countries with different levels of detraditionalization? We will answer them by using data from Eurobarometer 2010, including 24768 respondents from 25 European countries, which we supplement with a series of contextual measures that capture the level of detraditionalization of these societies.
Detraditionalization, mental illness reports, and mental health professional care use
Regarding the relationship between the level of detraditionalization and mental illness reports, critical theorists argue that in late-modern societies, the unpredictability of life course trajectories, the expansion of life choices, and the increased responsibility placed on individuals for these choices create perfect conditions for breeding depression and anxiety (Branaman, 2007). However, a conclusive body of empirical literature convincingly and repeatedly showed that in wealthier societies, that is more detraditionalized, the prevalence of mental illness in the population is lower than in the less wealthy counterparts (Van de Velde, 2010; Levecque et al, 2011; Van Deurzen, 2015). These findings are usually attributed to the overall better quality of the living conditions, but some authors argued that individualistic cultures breed higher happiness precisely because of the ‘cultural environment where individuals make choices to maximize their happiness rather than to meet social obligations’ (Ahuvia, 2001: p. 25). Empirical studies provide support for this claim, that is individuals living in individualistic societies are happier than those living in collectivistic ones (Veenhoven, 2015) and a possible explanation of this finding is that the positive value attached to the possibilities to make choices outweigh the costs associated with the act of making a choice.
Regarding the potential link between the level of detraditionalization and mental health professional care use, the arguments from the literature suggest a positive relationship. A consequence of the diffusion of the psychological model in society at large through mass media is a higher degree of mental health literacy in individuals (Elchardus, 2009). Individuals living in detraditionalized societies are likely more familiarized with different types of mental health conditions and their symptoms, and as a result, they could faster determine whether their emotional struggles can be described as mental illness, subject to treatment and recovery. Next, the medicalization of everyday life has resulted in the framing of an increased number of conditions as illnesses that warrant medical attention (Furedi, 2004; Horwitz, 2011), a process reflected in a higher usage of mental health professional care use. Empirical evidence both from quantitative and qualitative studies supports these ideas (Horwitz and Wakefield, 2006; Buffel, Beckfield and Bracke, 2017).
This last argument also emphasizes that an increase in mental illness reports might operate as a mediator factor between the increasing level of detraditionalization and increasing mental health professional care use. However, given the contrasting explanations on how detraditionalization relates to mental illness reports, it remains an empirical question to what extent the relationship between the level of detraditionalization and mental health professional care use still holds when the level of mental illness reports in the population is accounted for.
Based on the above we expect that the average reports of mental illness will be lower (H1) and the mental health professional care use will be higher (H2) in more detraditionalized countries compared to less detraditionalized countries and these effects are expected to be observed even when controlling for the level of country’s wealth and, for H2, also for the mental illness reports.
Material and social adversity in detraditionalized societies
The lack of resources needed to maintain a decent level of living has been linked to an increased chance of mental illness primarily through increasing exposure to stressors that range from the daily hustles of having to make ends meet due to financial strain (i.e. material adversity) to an increased chance of experiencing critical life events such as unemployment or divorce (i.e. social adversity) (Whelan, 1994; Thoits, 2010). Chronic exposure to material and/or social adversity was related to mental illness reports and mental health professional care use (Levecque et al, 2011; Buffel, Beckfield and Bracke, 2017; Mandemakers and Kalmijn, 2018). While these relationships seem to be robust in every social context, we question whether the experience of adversity has a different impact on both mental illness reports and on mental health professional care use in countries with different levels of detraditionalization.
In relation to mental illness reports, the role of different material and social adversity indicators could be different in countries with different levels of detraditionalization. First, in line with the previous literature, we expect that individuals that experience material adversity will report higher levels of mental illness compared to individuals who do not experience material adversity, and the difference between these two groups will be higher in more detraditionalized societies (H3). The argument behind this is that of increased personal responsibility that is placed on individuals for their own choices and the outcomes of these choices, which is a cornerstone of both the medicalization and individualization literature (Conrad, 1992; Branaman, 2007). In detraditionalized societies, it is more likely that individuals are exposed to, and will accept the meritocratic discourse stating that everything is achievable and linked to personal effort and valour (Rüsch et al., 2010). If this is the case, it implies that not living the best life, that is facing adversity in the form of financial strain, is a personal fault that cannot be attributed to external circumstances. As a result, the internalization of the causes of own success or failures will most likely result in higher levels of mental illness reports.
The same argument can be raised when considering unemployment: in more detraditionalized countries, the pressure of living the best life and the meritocratic discourse could result in higher reported levels of mental illness as a result of this life event (H4). Furthermore, in more detraditionalized countries gender roles are more equal (Elchardus and de Keere, 2010), divorce is more common and not seen as a ‘faux pas’ (Dronkers, Kalmijn and Wagner, 2006). However, the institutional arrangements meant to assist couples going through marital dissolution (e.g., joint custody, income support) or individuals that experience an unemployment spell (e.g., income support) are more available and more generous in more detraditionalized countries. As such, both divorce and unemployment could be less disruptive in more detraditionalized countries and could result in lower levels of mental illness reports compared to more traditional countries (H5).
In relation to mental health professional care use, we derive expectations based on medicalization theory. Research shows that especially the experience of unemployment is highly medicalized, that is when controlled for their level of mental illness, unemployed people in all European countries have higher medical care use than employed people (Buffel, Beckfield and Bracke, 2017). But the medicalization of everyday life is argued to cover more and more other areas of life, for example love relationships (Earp, 2015) as well as the experience of poverty (Hansen, Bourgois and Drucker, 2014; Mills, 2015) which have become a focus of therapeutic interventions. It is possible that especially in detraditionalized countries, individuals experiencing both material and social adversity, will more likely frame the related distress as treatable. This implies that individuals living in more detraditionalized countries will report higher levels of mental health professional care use when experiencing material and social adversity as compared to individuals in the same situation but living in less detraditionalized societies (H6).
However, a competing argument can be raised for the case of divorce and unemployment that is if indeed as argued above, in more detraditionalized countries divorce and unemployment relate to lower mental illness reports, this could also imply that the mental health professional care use will be lower than in more traditional societies. Next, there is a question of doubt whether in more traditional societies, people experiencing a divorce would make use of mental health professional care due to the stigma surrounding both mental illness and the act of divorce. As such, we leave it to the data to shed light on where the balance lies in this case.
Data and methods
To test our hypotheses, we use data from Eurobarometer (73.2)22010, a multi-countries survey that used a multi-stage, random (probability) sampling design to collect cross-country comparative data from 27,304 individuals living in 25 European countries: [Austria (AT), Bulgaria (BG), Cyprus (CY), Czech Republic (CZ), Denmark (DK), Finland (FI), France (FR), Germany (DE), Greece (GR), Hungary (HU), Ireland (IE), Italy (IT), Latvia (LV), Lithuania (LT), Luxembourg (LU), Malta (MT), the Netherlands (NL), Poland (PL), Portugal (PT), Romania (RO), Slovakia (SK), Slovenia (SI), Spain (ES), Sweden (SE), and United Kingdom (UK)]. Belgium, Estonia, and Iceland had to be dropped from the available sample because data for explanatory variables were missing. As advised by Eurobarometer, we apply population weights for the descriptive analyses. The working sample size included information from 24,768 respondents.
Dependent variables
To construct our first dependent variable, mental illness reports, we used the slightly modified version of the Mental Health Inventory (MHI-5) scale (Ware and Sherbourne, 1992), that is present in the data: ‘How often during the past 4 weeks…? Have you felt particularly tense; Have you felt so down in the dumps that nothing could cheer you up; Have you felt calm and peaceful; Have you felt downhearted and depressed; Have you felt happy’. All items were measured on a scale from 1 to 5 (All the time/ never). We recoded the items in such a way that a higher score indicates a higher mental illness. People who answered ‘don’t know’ were coded as missing (395 respondents). Based on usual practice, we first created a sum scale that was rescaled from 0 to 100 (Cronbach’s alpha reliability coefficient was.82). The average score for mental illness was 31.23 and the standard deviation was 17.58.
For our second dependent variable, mental health professional care use, we used a variable that measures whether a respondent contacted a professional for mental help. Respondents could indicate this for ten types of professional help of which we selected the four types of professionals that specialize in psychological help: ‘In the last 12 months, did you seek help from a professional because of a psychological or emotional problem? Yes, a psychiatrist; a psychologist; a psychoanalyst; a psychotherapist (not mentioned previously).’ We did not include the following answer categories of types of professionals: a general practitioner, a pharmacist, a nurse, a social worker, another health professional, or someone else. 735 respondents contacted one of these types of professionals in the last 12 months, 74 contacted two professionals, and 16 respondents contacted three professionals. We constructed a variable that measures whether or not the respondent contacted at least one of these four types of professional help and used it as a dummy variable in our analyses.
Table 1 presents the descriptive statistics for all the individual-level indicators.
. | Mean . | St.dv . | Min–Max . | N . |
---|---|---|---|---|
Mental health professional care use | 0.03 | 0.18 | 0–1 | 24768 |
Mental illness reports | 31.23 | 17.58 | 0–100 | 24373 |
Unemployment | 0.08 | 0.28 | 0–1 | 24768 |
Financial Strain | 1.49 | 0.66 | 1–3 | 24234 |
Divorce | 0.07 | 0.26 | 0–1 | 24729 |
Physical Health | 3.81 | 1.18 | 1–5 | 24546 |
Stigma | 0.27 | 0.45 | 0–1 | 22206 |
Education—Low | 0.21 | 0.41 | 0–1 | 24254 |
Education—Middle | 0.43 | 0.50 | 0–1 | 24254 |
Education—High | 0.28 | 0.45 | 0–1 | 24254 |
Education—Still Studying | 0.08 | 0.27 | 0–1 | 24254 |
Age—Young (15–31 years) | 0.21 | 0.41 | 0–1 | 24768 |
Age—Adult (32–65 years) | 0.57 | 0.50 | 0–1 | 24768 |
Age—Old (plus 65 years) | 0.22 | 0.41 | 0–1 | 24768 |
Woman | 0.54 | 0.50 | 0–1 | 24768 |
. | Mean . | St.dv . | Min–Max . | N . |
---|---|---|---|---|
Mental health professional care use | 0.03 | 0.18 | 0–1 | 24768 |
Mental illness reports | 31.23 | 17.58 | 0–100 | 24373 |
Unemployment | 0.08 | 0.28 | 0–1 | 24768 |
Financial Strain | 1.49 | 0.66 | 1–3 | 24234 |
Divorce | 0.07 | 0.26 | 0–1 | 24729 |
Physical Health | 3.81 | 1.18 | 1–5 | 24546 |
Stigma | 0.27 | 0.45 | 0–1 | 22206 |
Education—Low | 0.21 | 0.41 | 0–1 | 24254 |
Education—Middle | 0.43 | 0.50 | 0–1 | 24254 |
Education—High | 0.28 | 0.45 | 0–1 | 24254 |
Education—Still Studying | 0.08 | 0.27 | 0–1 | 24254 |
Age—Young (15–31 years) | 0.21 | 0.41 | 0–1 | 24768 |
Age—Adult (32–65 years) | 0.57 | 0.50 | 0–1 | 24768 |
Age—Old (plus 65 years) | 0.22 | 0.41 | 0–1 | 24768 |
Woman | 0.54 | 0.50 | 0–1 | 24768 |
. | Mean . | St.dv . | Min–Max . | N . |
---|---|---|---|---|
Mental health professional care use | 0.03 | 0.18 | 0–1 | 24768 |
Mental illness reports | 31.23 | 17.58 | 0–100 | 24373 |
Unemployment | 0.08 | 0.28 | 0–1 | 24768 |
Financial Strain | 1.49 | 0.66 | 1–3 | 24234 |
Divorce | 0.07 | 0.26 | 0–1 | 24729 |
Physical Health | 3.81 | 1.18 | 1–5 | 24546 |
Stigma | 0.27 | 0.45 | 0–1 | 22206 |
Education—Low | 0.21 | 0.41 | 0–1 | 24254 |
Education—Middle | 0.43 | 0.50 | 0–1 | 24254 |
Education—High | 0.28 | 0.45 | 0–1 | 24254 |
Education—Still Studying | 0.08 | 0.27 | 0–1 | 24254 |
Age—Young (15–31 years) | 0.21 | 0.41 | 0–1 | 24768 |
Age—Adult (32–65 years) | 0.57 | 0.50 | 0–1 | 24768 |
Age—Old (plus 65 years) | 0.22 | 0.41 | 0–1 | 24768 |
Woman | 0.54 | 0.50 | 0–1 | 24768 |
. | Mean . | St.dv . | Min–Max . | N . |
---|---|---|---|---|
Mental health professional care use | 0.03 | 0.18 | 0–1 | 24768 |
Mental illness reports | 31.23 | 17.58 | 0–100 | 24373 |
Unemployment | 0.08 | 0.28 | 0–1 | 24768 |
Financial Strain | 1.49 | 0.66 | 1–3 | 24234 |
Divorce | 0.07 | 0.26 | 0–1 | 24729 |
Physical Health | 3.81 | 1.18 | 1–5 | 24546 |
Stigma | 0.27 | 0.45 | 0–1 | 22206 |
Education—Low | 0.21 | 0.41 | 0–1 | 24254 |
Education—Middle | 0.43 | 0.50 | 0–1 | 24254 |
Education—High | 0.28 | 0.45 | 0–1 | 24254 |
Education—Still Studying | 0.08 | 0.27 | 0–1 | 24254 |
Age—Young (15–31 years) | 0.21 | 0.41 | 0–1 | 24768 |
Age—Adult (32–65 years) | 0.57 | 0.50 | 0–1 | 24768 |
Age—Old (plus 65 years) | 0.22 | 0.41 | 0–1 | 24768 |
Woman | 0.54 | 0.50 | 0–1 | 24768 |
Independent variables
Our main context variable is our measure of detraditionalization. We constructed this measure by using different indicators that correspond to the previously discussed aspects: the level of secularization, the ethos of personal autonomy, and the erosion of traditional gender roles associated with the decline in the acceptance of authority.
Processes of secularization and the ethos of personal autonomy were captured using data from the European Values Survey wave 4 (EVS, 2022): an aggregated measure of church attendance in which we differentiate between attendance (at least for a special occasion) and no church attendance, and an aggregated measure of autonomy that was constructed from a scale of three indicators encompassing whether the respondents agree with the statement ‘One does not have the duty to respect and love parents who have not earned it by their behavior and attitudes’ (1), and from a list of qualities which children can be encouraged to learn at home chose independence (2), and did not choose obedience (3). We constructed a sum scale and used the weighted averages as country-level scores. To capture the erosion of traditional gender roles, we used the gender inequality index from the United Nations Development Program, which covers gender inequalities in reproductive health (maternal mortality ratio and adolescent birth rates), empowerment (proportion of parliamentary seats occupied by females and the proportion of adult females and males aged 25 years and older with at least some secondary education) and economic status (expressed as labor market participation and measured by labor force participation rate of female and male populations aged 15 years and older). We took the average score of the 2005 and 2010 index values. Supplementary Table A1 gives an overview of the descriptive statistics of the contextual indicators per country.
Similar to the strategy used by (Elchardus and De Keere, 2010), we employed a principal component analysis which resulted in one factor with an eigenvalue of 2.01. All of the indicators have factor loadings higher than 0.750. Church attendance and gender inequality load positively, while autonomy loads negatively on the factor, as expected. We recoded the variable in such a way that a higher score indicates a higher level of detraditionalization. Supplementary Table A2 gives the full information about the factor model.
To test our moderation hypotheses, we used one individual-level measure for material adversity, that is financial strain, and two measures for social adversity, that is the experience of unemployment and divorce. To measure financial strain, we used an item that asked respondents how many times they had difficulties with paying the bills, which contains three categories: most of the time (3), from time to time (2), (almost) never (1). Unemployment was measured with a dummy variable indicating that a respondent was unemployed (1) or not unemployed (0) (self-employed, managers, other white collars, manual workers, house person, retired, and students). People indicating divorce or being separated on the item marital status were assigned a score of 1, others [(re)married, single living with partner, single or widow] a score of 0.
Control variables
At the contextual level, to account for possible economic factors influencing mental illness, and mental health professional care use, as well as our measure of detraditionalization, we used GDP per capita in purchasing power parity (PPP). The measure was obtained from Eurostat for the year 2009. We also controlled for the level of public stigma surrounding mental health in our models predicting mental health professional care use. The argument for including this measure is that higher levels of public stigma act like a barrier to accessing mental health professional care (Bracke and Verhaeghe, 2019). We used an aggregated measure of stigma constructed from one indicator measuring whether respondents agreed with one of the two statements: ‘You would find it difficult talking to someone with a significant mental health problem’ (0) or ‘You would have no problem talking to someone with significant mental health problem’ (1). The scores were reversed, such that a higher score indicated more stigma, and the country-level averages were multiplied by 100 to ease interpretation in the models. In the models in which the aggregated measure was included, we also controlled for composition effects by including the individual-level variable.
Finally, in the models in which we estimated the interactions with unemployment and divorce, we controlled for the unemployment rate and divorce rate respectively. In this way, we take into account the normalization of unemployment and divorce in a country, normative aspects that may confound the influence of detraditionalization on mental illness and mental health professional care use. The unemployment rate, that is the percentage of unemployed people of the total labor force for the year 2009, was obtained from the World Bank Development Indicators database. The divorce rate measured as the divorces per 100 marriages in 2009 (with exception of Malta for which the most recent data was available for 2012) was obtained from Eurostat. Since the number of observations at the country level is limited to 25, we cannot take into account all contextual factors at the same time. Supplementary Table A3 provides a correlation matrix of all the contextual-level indicators used.
At the individual level, we added control variables for physical health, education, gender, and age. For physical health we used the response to the statement: ‘During the past 4 weeks how often have you had any of the following problems in your day-to-day life whether at work, at home, or elsewhere? You have accomplished less than you would like as a result of a physical health problem’. This was measured on a five-point scale running from ‘all the time’ (1) till ‘never’ (5). For education, we used a measure of the years of education. We divided it into four categories: no full-time education, low education (less than 15 years), middle education (16–19 years), high education (+20 years), and still studying. For gender, we used man as the reference category. We used age split into three categories: young (15–30) adult (31–64), and old (65+) (adult is ref.). In the models where the dependent variable was mental health professional care use, we also included the measure of mental illness.
Methods
In order to answer our research questions and test our hypotheses, in addition to descriptive statistics, we employed linear multilevel regression analysis for the outcome variable mental illness reports and multilevel logistic regression analysis for the outcome variable mental health professional care use. We started with a model in which all individual-level control variables were added, which allowed us to control for the population composition (Model 1). In the next step, we estimated the effect of our measure of detraditionalization alone (Model 2). For our outcome variable mental health professional care use, we also controlled for public stigma and mental illness (Model 3) to which we added GDP per capita in PPP in Model 4. For the outcome variable mental illness, we only controlled for GDP per capita in PPP (Model 3). In order to test our hypotheses H3–H6, we estimated interaction terms between the detraditionalization measure with the measures of social and material adversity. We included a random slope for the individual-level variables (Heisig and Schaeffer, 2019). In these models, we also controlled for the unemployment rate (for the interaction with unemployed status) and divorce rate (for the interaction with divorce status). The number of respondents that accessed mental health professional care is relatively low in terms of percentages (3 per cent of the sample) however, it is larger than the rule of thumb of 200 cases (825 respondents), and thus, following the discussion by Allison (2012) on the matter, we do not think this poses a problem for our analyses3.
Results
Before the formal tests of our hypotheses, we present two scatterplots that illustrate the relationship between the level of detraditionalization and the average level of mental illness reports and mental health professional care use in our sample of 25 European countries. Figure 1 shows that the relationship between detraditionalization and the average level of mental illness reports is negative and strong (correlation of −0.77). As such, countries with the highest level of detraditionalization and the lowest level of mental illness reports are the Scandinavian countries and the Netherlands, while Southern and Eastern European countries show the opposite pattern. Figure 2 illustrates that in countries with higher levels of detraditionalization, a higher percentage of respondents reported having used mental health professional care (correlation of 0.56). For instance, around 7 per cent of the respondents in Denmark and the Netherlands reported having visited a professional, while in Bulgaria and Romania this percentage lies around 1 per cent.
We now turn to the results of our multivariate analyses. Table 2 summarizes the results for our dependent variable mental illness reports. Estimating a model without independent variables showed that about 6.5 per cent of the variance in mental illness reports was related to country-level differences. People that experienced adversity in terms of unemployment, divorce, or financial strain were more likely to report more symptoms of mental illness. Furthermore, individuals with worse physical health, fewer years of education, adults, and women had higher levels of mental illness reports.
. | Model 1 . | Model 2 . | Model 3 . | Model 4a . | Model 5a . | Model 6a . | ||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
Coef. . | se . | Coef. . | se . | Coef. . | se . | Coef. . | se . | Coef. . | se . | Coef. . | se . | |
Constant | 42.51*** | 0.81 | 44.51*** | 0.80 | 44.55*** | 1.46 | 43.47*** | 1.35 | 44.76*** | 0.75 | 44.94*** | 1.48 |
Individual characteristics | ||||||||||||
Unemployment | 3.43*** | 0.37 | 3.43*** | 0.37 | 3.43*** | 0.38 | 2.93*** | 0.71 | 3.37*** | 0.37 | 3.41*** | 0.37 |
Financial Strain | 6.04*** | 0.17 | 6.02*** | 0.17 | 6.02*** | 0.17 | 6.01*** | 0.17 | 5.85*** | 0.31 | 6.01*** | 0.17 |
Divorce | 2.47*** | 0.38 | 2.48*** | 0.38 | 2.48*** | 0.38 | 2.48*** | 0.38 | 2.47*** | 0.38 | 4.42*** | 0.66 |
Physical Health | −5.37*** | 0.09 | −5.37*** | 0.09 | −5.37*** | 0.09 | −5.37*** | 0.09 | −5.37*** | 0.09 | −5.37*** | 0.09 |
Education (low is ref.) | ||||||||||||
Middle | −0.90** | 0.29 | −0.91** | 0.29 | −0.91** | 0.29 | −0.92** | 0.29 | −0.91** | 0.29 | −0.92** | 0.29 |
High | −0.89** | 0.32 | −0.88** | 0.32 | −0.88** | 0.32 | −0.88** | 0.32 | −0.86** | 0.32 | −0.90** | 0.32 |
Still Studying | −0.04 | 0.52 | −0.04 | 0.52 | −0.04 | 0.52 | −0.01 | 0.52 | −0.05 | 0.52 | −0.02 | 0.52 |
Age (adult is ref.) | ||||||||||||
Young (15–31 years) | −1.46*** | 0.30 | −1.47*** | 0.30 | −1.47*** | 0.30 | −1.46*** | 0.30 | −1.48*** | 0.30 | −1.44*** | 0.30 |
Old (plus 65 years) | −2.72*** | 0.27 | −2.71*** | 0.27 | −2.72*** | 0.27 | −2.71*** | 0.27 | −2.71*** | 0.27 | −2.71*** | 0.27 |
Woman (men is ref.) | 2.33*** | 0.20 | 2.33*** | 0.20 | 2.33*** | 0.20 | 2.33*** | 0.20 | 2.33*** | 0.20 | 2.33*** | 0.20 |
Contextual characteristic | ||||||||||||
Detraditionalization | −2.11*** | 0.43 | −2.10*** | 0.52 | −2.11*** | 0.43 | −2.34*** | 0.56 | −1.91*** | 0.50 | ||
GDP per capita in ppp | −0.00 | 0.01 | ||||||||||
Unemployment rate | 0.12 | 0.12 | ||||||||||
Divorce rate | −0.01 | 0.03 | ||||||||||
Interaction terms | ||||||||||||
Detrad. * Unempl. | 0.55 | 0.53 | ||||||||||
Detrad. * Fin. Strain | 0.15 | 0.25 | ||||||||||
Detrad. * Divorce | −1.55*** | 0.43 |
. | Model 1 . | Model 2 . | Model 3 . | Model 4a . | Model 5a . | Model 6a . | ||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
Coef. . | se . | Coef. . | se . | Coef. . | se . | Coef. . | se . | Coef. . | se . | Coef. . | se . | |
Constant | 42.51*** | 0.81 | 44.51*** | 0.80 | 44.55*** | 1.46 | 43.47*** | 1.35 | 44.76*** | 0.75 | 44.94*** | 1.48 |
Individual characteristics | ||||||||||||
Unemployment | 3.43*** | 0.37 | 3.43*** | 0.37 | 3.43*** | 0.38 | 2.93*** | 0.71 | 3.37*** | 0.37 | 3.41*** | 0.37 |
Financial Strain | 6.04*** | 0.17 | 6.02*** | 0.17 | 6.02*** | 0.17 | 6.01*** | 0.17 | 5.85*** | 0.31 | 6.01*** | 0.17 |
Divorce | 2.47*** | 0.38 | 2.48*** | 0.38 | 2.48*** | 0.38 | 2.48*** | 0.38 | 2.47*** | 0.38 | 4.42*** | 0.66 |
Physical Health | −5.37*** | 0.09 | −5.37*** | 0.09 | −5.37*** | 0.09 | −5.37*** | 0.09 | −5.37*** | 0.09 | −5.37*** | 0.09 |
Education (low is ref.) | ||||||||||||
Middle | −0.90** | 0.29 | −0.91** | 0.29 | −0.91** | 0.29 | −0.92** | 0.29 | −0.91** | 0.29 | −0.92** | 0.29 |
High | −0.89** | 0.32 | −0.88** | 0.32 | −0.88** | 0.32 | −0.88** | 0.32 | −0.86** | 0.32 | −0.90** | 0.32 |
Still Studying | −0.04 | 0.52 | −0.04 | 0.52 | −0.04 | 0.52 | −0.01 | 0.52 | −0.05 | 0.52 | −0.02 | 0.52 |
Age (adult is ref.) | ||||||||||||
Young (15–31 years) | −1.46*** | 0.30 | −1.47*** | 0.30 | −1.47*** | 0.30 | −1.46*** | 0.30 | −1.48*** | 0.30 | −1.44*** | 0.30 |
Old (plus 65 years) | −2.72*** | 0.27 | −2.71*** | 0.27 | −2.72*** | 0.27 | −2.71*** | 0.27 | −2.71*** | 0.27 | −2.71*** | 0.27 |
Woman (men is ref.) | 2.33*** | 0.20 | 2.33*** | 0.20 | 2.33*** | 0.20 | 2.33*** | 0.20 | 2.33*** | 0.20 | 2.33*** | 0.20 |
Contextual characteristic | ||||||||||||
Detraditionalization | −2.11*** | 0.43 | −2.10*** | 0.52 | −2.11*** | 0.43 | −2.34*** | 0.56 | −1.91*** | 0.50 | ||
GDP per capita in ppp | −0.00 | 0.01 | ||||||||||
Unemployment rate | 0.12 | 0.12 | ||||||||||
Divorce rate | −0.01 | 0.03 | ||||||||||
Interaction terms | ||||||||||||
Detrad. * Unempl. | 0.55 | 0.53 | ||||||||||
Detrad. * Fin. Strain | 0.15 | 0.25 | ||||||||||
Detrad. * Divorce | −1.55*** | 0.43 |
*P < 0.05; **P < 0.01; ***P < 0.001
aResults for the interaction effects are ro.bust if models are controlled for GDP per capita in ppp.
. | Model 1 . | Model 2 . | Model 3 . | Model 4a . | Model 5a . | Model 6a . | ||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
Coef. . | se . | Coef. . | se . | Coef. . | se . | Coef. . | se . | Coef. . | se . | Coef. . | se . | |
Constant | 42.51*** | 0.81 | 44.51*** | 0.80 | 44.55*** | 1.46 | 43.47*** | 1.35 | 44.76*** | 0.75 | 44.94*** | 1.48 |
Individual characteristics | ||||||||||||
Unemployment | 3.43*** | 0.37 | 3.43*** | 0.37 | 3.43*** | 0.38 | 2.93*** | 0.71 | 3.37*** | 0.37 | 3.41*** | 0.37 |
Financial Strain | 6.04*** | 0.17 | 6.02*** | 0.17 | 6.02*** | 0.17 | 6.01*** | 0.17 | 5.85*** | 0.31 | 6.01*** | 0.17 |
Divorce | 2.47*** | 0.38 | 2.48*** | 0.38 | 2.48*** | 0.38 | 2.48*** | 0.38 | 2.47*** | 0.38 | 4.42*** | 0.66 |
Physical Health | −5.37*** | 0.09 | −5.37*** | 0.09 | −5.37*** | 0.09 | −5.37*** | 0.09 | −5.37*** | 0.09 | −5.37*** | 0.09 |
Education (low is ref.) | ||||||||||||
Middle | −0.90** | 0.29 | −0.91** | 0.29 | −0.91** | 0.29 | −0.92** | 0.29 | −0.91** | 0.29 | −0.92** | 0.29 |
High | −0.89** | 0.32 | −0.88** | 0.32 | −0.88** | 0.32 | −0.88** | 0.32 | −0.86** | 0.32 | −0.90** | 0.32 |
Still Studying | −0.04 | 0.52 | −0.04 | 0.52 | −0.04 | 0.52 | −0.01 | 0.52 | −0.05 | 0.52 | −0.02 | 0.52 |
Age (adult is ref.) | ||||||||||||
Young (15–31 years) | −1.46*** | 0.30 | −1.47*** | 0.30 | −1.47*** | 0.30 | −1.46*** | 0.30 | −1.48*** | 0.30 | −1.44*** | 0.30 |
Old (plus 65 years) | −2.72*** | 0.27 | −2.71*** | 0.27 | −2.72*** | 0.27 | −2.71*** | 0.27 | −2.71*** | 0.27 | −2.71*** | 0.27 |
Woman (men is ref.) | 2.33*** | 0.20 | 2.33*** | 0.20 | 2.33*** | 0.20 | 2.33*** | 0.20 | 2.33*** | 0.20 | 2.33*** | 0.20 |
Contextual characteristic | ||||||||||||
Detraditionalization | −2.11*** | 0.43 | −2.10*** | 0.52 | −2.11*** | 0.43 | −2.34*** | 0.56 | −1.91*** | 0.50 | ||
GDP per capita in ppp | −0.00 | 0.01 | ||||||||||
Unemployment rate | 0.12 | 0.12 | ||||||||||
Divorce rate | −0.01 | 0.03 | ||||||||||
Interaction terms | ||||||||||||
Detrad. * Unempl. | 0.55 | 0.53 | ||||||||||
Detrad. * Fin. Strain | 0.15 | 0.25 | ||||||||||
Detrad. * Divorce | −1.55*** | 0.43 |
. | Model 1 . | Model 2 . | Model 3 . | Model 4a . | Model 5a . | Model 6a . | ||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|
Coef. . | se . | Coef. . | se . | Coef. . | se . | Coef. . | se . | Coef. . | se . | Coef. . | se . | |
Constant | 42.51*** | 0.81 | 44.51*** | 0.80 | 44.55*** | 1.46 | 43.47*** | 1.35 | 44.76*** | 0.75 | 44.94*** | 1.48 |
Individual characteristics | ||||||||||||
Unemployment | 3.43*** | 0.37 | 3.43*** | 0.37 | 3.43*** | 0.38 | 2.93*** | 0.71 | 3.37*** | 0.37 | 3.41*** | 0.37 |
Financial Strain | 6.04*** | 0.17 | 6.02*** | 0.17 | 6.02*** | 0.17 | 6.01*** | 0.17 | 5.85*** | 0.31 | 6.01*** | 0.17 |
Divorce | 2.47*** | 0.38 | 2.48*** | 0.38 | 2.48*** | 0.38 | 2.48*** | 0.38 | 2.47*** | 0.38 | 4.42*** | 0.66 |
Physical Health | −5.37*** | 0.09 | −5.37*** | 0.09 | −5.37*** | 0.09 | −5.37*** | 0.09 | −5.37*** | 0.09 | −5.37*** | 0.09 |
Education (low is ref.) | ||||||||||||
Middle | −0.90** | 0.29 | −0.91** | 0.29 | −0.91** | 0.29 | −0.92** | 0.29 | −0.91** | 0.29 | −0.92** | 0.29 |
High | −0.89** | 0.32 | −0.88** | 0.32 | −0.88** | 0.32 | −0.88** | 0.32 | −0.86** | 0.32 | −0.90** | 0.32 |
Still Studying | −0.04 | 0.52 | −0.04 | 0.52 | −0.04 | 0.52 | −0.01 | 0.52 | −0.05 | 0.52 | −0.02 | 0.52 |
Age (adult is ref.) | ||||||||||||
Young (15–31 years) | −1.46*** | 0.30 | −1.47*** | 0.30 | −1.47*** | 0.30 | −1.46*** | 0.30 | −1.48*** | 0.30 | −1.44*** | 0.30 |
Old (plus 65 years) | −2.72*** | 0.27 | −2.71*** | 0.27 | −2.72*** | 0.27 | −2.71*** | 0.27 | −2.71*** | 0.27 | −2.71*** | 0.27 |
Woman (men is ref.) | 2.33*** | 0.20 | 2.33*** | 0.20 | 2.33*** | 0.20 | 2.33*** | 0.20 | 2.33*** | 0.20 | 2.33*** | 0.20 |
Contextual characteristic | ||||||||||||
Detraditionalization | −2.11*** | 0.43 | −2.10*** | 0.52 | −2.11*** | 0.43 | −2.34*** | 0.56 | −1.91*** | 0.50 | ||
GDP per capita in ppp | −0.00 | 0.01 | ||||||||||
Unemployment rate | 0.12 | 0.12 | ||||||||||
Divorce rate | −0.01 | 0.03 | ||||||||||
Interaction terms | ||||||||||||
Detrad. * Unempl. | 0.55 | 0.53 | ||||||||||
Detrad. * Fin. Strain | 0.15 | 0.25 | ||||||||||
Detrad. * Divorce | −1.55*** | 0.43 |
*P < 0.05; **P < 0.01; ***P < 0.001
aResults for the interaction effects are ro.bust if models are controlled for GDP per capita in ppp.
In Model 2 we estimated the effect of detraditionalization and we found it to be in line with the bivariate relationship presented in Figure 1, that is in more detraditionalized countries, the reported level of mental illness reports was lower than in less detraditionalized countries. Model 3, in which GDP per capita was added as a control variable, showed that this effect was robust thus supporting our first hypothesis.
In Models 4, 5, and 6 the interaction effects of detraditionalization and our indicators of adversity were added. We found one significant and negative cross-level interaction with divorce status (Model 6), that is the difference in levels of mental illness between divorced and not divorced individuals in detraditionalized societies was smaller than between their counterparts living in less detraditionalized societies. The predicted means of mental illness reports for this interaction is presented in Figure 3. The figure shows that in the least detraditionalized countries in our sample, the difference in the average predicted mental illness reports between divorced and not divorced individuals was around 7 points on our scale that ranged from 0 to 100, a difference that represents about 40 percent of the observed standard deviation of this measure.
In Table 3, we present the results for the dependent variable of mental health professional care use. Using the variance decomposition of a model without any independent variables added, we estimated that around 6 percent of the variance in mental health professional care use is due to countries’ differences. In Model 1, we added all the individual-level variables, excluding mental illness reports. We found that all our measures of adversity had a positive relationship with mental health professional care use, that is unemployed and divorced individuals and individuals experiencing financial strain were more likely to have sought help from a psychiatrist, psychologist, psychoanalyst, or a psychotherapist. Regarding the effect of the control variables, we found that the respondents who reported worse physical health and were highly educated had a higher likelihood to access mental health professional care. Moreover, we found that among older individuals the likelihood to access mental health professional care was lower, and higher for women compared to men. When we include our indicator of mental illness in Model 2, we see that the variable of mental illness reports mediates the effect between the adversity measures and mental health professional care use, that is the effects of unemployment and divorce rate became smaller and the effect of financial strain not significant.
. | Model 1 . | Model 2 . | Model 3 . | Model 4 . | Model 5 . | Model 6 . | Model 7a . | Model 8a . | Model 9a . | |||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Coef. . | se . | Coef. . | se . | Coef. . | se . | Coef. . | se . | Coef. . | se . | Coef. . | se . | Coef. . | se . | Coef. . | se . | Coef. . | se . | |
Constant | −2.56*** | 0.21 | −4.90*** | 0.26 | −2.89*** | 0.22 | −5.35*** | 0.27 | −4.62*** | 0.36 | −5.29*** | 0.49 | −513*** | 0.36 | −4.89*** | 0.28 | −5.27*** | 0.39 |
Individual characteristics | ||||||||||||||||||
Unemployment | 0.62*** | 0.11 | 0.43*** | 0.11 | 0.61*** | 0.11 | 0.41*** | 0.11 | 0.42*** | 0.12 | 0.43*** | 0.12 | 0.13 | 0.22 | 0.41*** | 0.12 | 0.42** | 0.11 |
Financial Strain | 0.38*** | 0.06 | 0.07 | 0.06 | 0.39*** | 0.06 | 0.08 | 0.06 | 0.10 | 0.06 | 0.11 | 0.06 | 0.08 | 0.06 | −0.21* | 0.11 | 0.08 | 0.06 |
Divorce | 0.57*** | 0.11 | 0.47*** | 0.11 | 0.56*** | 0.11 | 0.46*** | 0.11 | 0.49*** | 0.11 | 0.48*** | 0.11 | 0.44*** | 0.11 | 0.43*** | 0.11 | 0.45* | 0.19 |
Physical Health | −0.53*** | 0.03 | −0.28*** | 0.03 | −0.53*** | 0.03 | −0.27*** | 0.03 | −0.27*** | 0.03 | −0.27*** | 0.03 | −0.27*** | 0.03 | −0.27*** | 0.03 | −0.27*** | 0.03 |
Stigma | −0.06 | 0.09 | −0.06 | 0.09 | ||||||||||||||
Education (low is ref.) | ||||||||||||||||||
Middle | 0.02 | 0.11 | 0.10 | 0.12 | 0.00 | 0.11 | 0.07 | 0.12 | 0.03 | 0.12 | 0.03 | 0.12 | 0.07 | 0.12 | 0.09 | 0.12 | 0.07 | 0.12 |
High | 0.34** | 0.12 | 0.48*** | 0.12 | 0.30* | 0.12 | 0.44*** | 0.12 | 0.42** | 0.13 | 0.42** | 0.13 | 0.44*** | 0.12 | 0.47*** | 0.12 | 0.44*** | 0.12 |
Still Studying | 0.61** | 0.19 | 0.67** | 0.20 | 0.57** | 0.19 | 0.64** | 0.20 | 0.65** | 0.20 | 0.65** | 0.20 | 0.64** | 0.20 | 0.64** | 0.20 | 0.64** | 0.20 |
Age (adult is ref.) | ||||||||||||||||||
Young (15–31 years) | −0.08 | 0.12 | −0.03 | 0.12 | −0.07 | 0.12 | −0.02 | 0.12 | −0.00 | 0.12 | −0.00 | 0.12 | −0.02 | 0.12 | −0.04 | 0.12 | −0.02 | 0.12 |
Old (plus 65 years) | −1.05*** | 0.13 | −0.97*** | 0.13 | −1.06*** | 0.13 | −0.99*** | 0.13 | −0.95*** | 0.13 | −0.95*** | 0.13 | −0.98*** | 0.13 | −0.96*** | 0.13 | −0.99*** | 0.13 |
Woman (men is ref.) | 0.43*** | 0.08 | 0.32*** | 0.08 | 0.43*** | 0.08 | 0.32*** | 0.08 | 0.29*** | 0.08 | 0.30*** | 0.08 | 0.32*** | 0.08 | 0.31*** | 0.08 | 0.32*** | 0.08 |
Mental illness reports | 0.05*** | 0.00 | 0.05*** | 0.00 | 0.05*** | 0.00 | 0.05*** | 0.00 | 0.05*** | 0.00 | 0.05*** | 0.00 | 0.05*** | 0.00 | ||||
Contextual characteristic | ||||||||||||||||||
Detraditionalization | 0.35*** | 0.09 | 0.48*** | 0.10 | 0.38*** | 0.09 | 0.34** | 0.09 | 0.45*** | 0.10 | 0.09 | 0.12 | 0.49*** | 0.11 | ||||
Public Stigma | −0.02* | 0.01 | −0.01 | 0.01 | ||||||||||||||
GDP per capita in ppp | 0.05* | 0.00 | ||||||||||||||||
Unemployment rate | −0.02 | 0.03 | ||||||||||||||||
Divorce rate | −0.00 | 0.01 | ||||||||||||||||
Interaction terms | ||||||||||||||||||
Detrad. * Unempl. | 0.23 | 0.14 | ||||||||||||||||
Detrad. * Fin. Strain | 0.25*** | 0.07 | ||||||||||||||||
Detrad. * Divorce | 0.01 | 0.12 |
. | Model 1 . | Model 2 . | Model 3 . | Model 4 . | Model 5 . | Model 6 . | Model 7a . | Model 8a . | Model 9a . | |||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Coef. . | se . | Coef. . | se . | Coef. . | se . | Coef. . | se . | Coef. . | se . | Coef. . | se . | Coef. . | se . | Coef. . | se . | Coef. . | se . | |
Constant | −2.56*** | 0.21 | −4.90*** | 0.26 | −2.89*** | 0.22 | −5.35*** | 0.27 | −4.62*** | 0.36 | −5.29*** | 0.49 | −513*** | 0.36 | −4.89*** | 0.28 | −5.27*** | 0.39 |
Individual characteristics | ||||||||||||||||||
Unemployment | 0.62*** | 0.11 | 0.43*** | 0.11 | 0.61*** | 0.11 | 0.41*** | 0.11 | 0.42*** | 0.12 | 0.43*** | 0.12 | 0.13 | 0.22 | 0.41*** | 0.12 | 0.42** | 0.11 |
Financial Strain | 0.38*** | 0.06 | 0.07 | 0.06 | 0.39*** | 0.06 | 0.08 | 0.06 | 0.10 | 0.06 | 0.11 | 0.06 | 0.08 | 0.06 | −0.21* | 0.11 | 0.08 | 0.06 |
Divorce | 0.57*** | 0.11 | 0.47*** | 0.11 | 0.56*** | 0.11 | 0.46*** | 0.11 | 0.49*** | 0.11 | 0.48*** | 0.11 | 0.44*** | 0.11 | 0.43*** | 0.11 | 0.45* | 0.19 |
Physical Health | −0.53*** | 0.03 | −0.28*** | 0.03 | −0.53*** | 0.03 | −0.27*** | 0.03 | −0.27*** | 0.03 | −0.27*** | 0.03 | −0.27*** | 0.03 | −0.27*** | 0.03 | −0.27*** | 0.03 |
Stigma | −0.06 | 0.09 | −0.06 | 0.09 | ||||||||||||||
Education (low is ref.) | ||||||||||||||||||
Middle | 0.02 | 0.11 | 0.10 | 0.12 | 0.00 | 0.11 | 0.07 | 0.12 | 0.03 | 0.12 | 0.03 | 0.12 | 0.07 | 0.12 | 0.09 | 0.12 | 0.07 | 0.12 |
High | 0.34** | 0.12 | 0.48*** | 0.12 | 0.30* | 0.12 | 0.44*** | 0.12 | 0.42** | 0.13 | 0.42** | 0.13 | 0.44*** | 0.12 | 0.47*** | 0.12 | 0.44*** | 0.12 |
Still Studying | 0.61** | 0.19 | 0.67** | 0.20 | 0.57** | 0.19 | 0.64** | 0.20 | 0.65** | 0.20 | 0.65** | 0.20 | 0.64** | 0.20 | 0.64** | 0.20 | 0.64** | 0.20 |
Age (adult is ref.) | ||||||||||||||||||
Young (15–31 years) | −0.08 | 0.12 | −0.03 | 0.12 | −0.07 | 0.12 | −0.02 | 0.12 | −0.00 | 0.12 | −0.00 | 0.12 | −0.02 | 0.12 | −0.04 | 0.12 | −0.02 | 0.12 |
Old (plus 65 years) | −1.05*** | 0.13 | −0.97*** | 0.13 | −1.06*** | 0.13 | −0.99*** | 0.13 | −0.95*** | 0.13 | −0.95*** | 0.13 | −0.98*** | 0.13 | −0.96*** | 0.13 | −0.99*** | 0.13 |
Woman (men is ref.) | 0.43*** | 0.08 | 0.32*** | 0.08 | 0.43*** | 0.08 | 0.32*** | 0.08 | 0.29*** | 0.08 | 0.30*** | 0.08 | 0.32*** | 0.08 | 0.31*** | 0.08 | 0.32*** | 0.08 |
Mental illness reports | 0.05*** | 0.00 | 0.05*** | 0.00 | 0.05*** | 0.00 | 0.05*** | 0.00 | 0.05*** | 0.00 | 0.05*** | 0.00 | 0.05*** | 0.00 | ||||
Contextual characteristic | ||||||||||||||||||
Detraditionalization | 0.35*** | 0.09 | 0.48*** | 0.10 | 0.38*** | 0.09 | 0.34** | 0.09 | 0.45*** | 0.10 | 0.09 | 0.12 | 0.49*** | 0.11 | ||||
Public Stigma | −0.02* | 0.01 | −0.01 | 0.01 | ||||||||||||||
GDP per capita in ppp | 0.05* | 0.00 | ||||||||||||||||
Unemployment rate | −0.02 | 0.03 | ||||||||||||||||
Divorce rate | −0.00 | 0.01 | ||||||||||||||||
Interaction terms | ||||||||||||||||||
Detrad. * Unempl. | 0.23 | 0.14 | ||||||||||||||||
Detrad. * Fin. Strain | 0.25*** | 0.07 | ||||||||||||||||
Detrad. * Divorce | 0.01 | 0.12 |
*P < 0.05; **P < 0.01; ***P < 0.001.
aResults for the interaction effects are robust if models are controlled for GDP per capita in ppp.
. | Model 1 . | Model 2 . | Model 3 . | Model 4 . | Model 5 . | Model 6 . | Model 7a . | Model 8a . | Model 9a . | |||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Coef. . | se . | Coef. . | se . | Coef. . | se . | Coef. . | se . | Coef. . | se . | Coef. . | se . | Coef. . | se . | Coef. . | se . | Coef. . | se . | |
Constant | −2.56*** | 0.21 | −4.90*** | 0.26 | −2.89*** | 0.22 | −5.35*** | 0.27 | −4.62*** | 0.36 | −5.29*** | 0.49 | −513*** | 0.36 | −4.89*** | 0.28 | −5.27*** | 0.39 |
Individual characteristics | ||||||||||||||||||
Unemployment | 0.62*** | 0.11 | 0.43*** | 0.11 | 0.61*** | 0.11 | 0.41*** | 0.11 | 0.42*** | 0.12 | 0.43*** | 0.12 | 0.13 | 0.22 | 0.41*** | 0.12 | 0.42** | 0.11 |
Financial Strain | 0.38*** | 0.06 | 0.07 | 0.06 | 0.39*** | 0.06 | 0.08 | 0.06 | 0.10 | 0.06 | 0.11 | 0.06 | 0.08 | 0.06 | −0.21* | 0.11 | 0.08 | 0.06 |
Divorce | 0.57*** | 0.11 | 0.47*** | 0.11 | 0.56*** | 0.11 | 0.46*** | 0.11 | 0.49*** | 0.11 | 0.48*** | 0.11 | 0.44*** | 0.11 | 0.43*** | 0.11 | 0.45* | 0.19 |
Physical Health | −0.53*** | 0.03 | −0.28*** | 0.03 | −0.53*** | 0.03 | −0.27*** | 0.03 | −0.27*** | 0.03 | −0.27*** | 0.03 | −0.27*** | 0.03 | −0.27*** | 0.03 | −0.27*** | 0.03 |
Stigma | −0.06 | 0.09 | −0.06 | 0.09 | ||||||||||||||
Education (low is ref.) | ||||||||||||||||||
Middle | 0.02 | 0.11 | 0.10 | 0.12 | 0.00 | 0.11 | 0.07 | 0.12 | 0.03 | 0.12 | 0.03 | 0.12 | 0.07 | 0.12 | 0.09 | 0.12 | 0.07 | 0.12 |
High | 0.34** | 0.12 | 0.48*** | 0.12 | 0.30* | 0.12 | 0.44*** | 0.12 | 0.42** | 0.13 | 0.42** | 0.13 | 0.44*** | 0.12 | 0.47*** | 0.12 | 0.44*** | 0.12 |
Still Studying | 0.61** | 0.19 | 0.67** | 0.20 | 0.57** | 0.19 | 0.64** | 0.20 | 0.65** | 0.20 | 0.65** | 0.20 | 0.64** | 0.20 | 0.64** | 0.20 | 0.64** | 0.20 |
Age (adult is ref.) | ||||||||||||||||||
Young (15–31 years) | −0.08 | 0.12 | −0.03 | 0.12 | −0.07 | 0.12 | −0.02 | 0.12 | −0.00 | 0.12 | −0.00 | 0.12 | −0.02 | 0.12 | −0.04 | 0.12 | −0.02 | 0.12 |
Old (plus 65 years) | −1.05*** | 0.13 | −0.97*** | 0.13 | −1.06*** | 0.13 | −0.99*** | 0.13 | −0.95*** | 0.13 | −0.95*** | 0.13 | −0.98*** | 0.13 | −0.96*** | 0.13 | −0.99*** | 0.13 |
Woman (men is ref.) | 0.43*** | 0.08 | 0.32*** | 0.08 | 0.43*** | 0.08 | 0.32*** | 0.08 | 0.29*** | 0.08 | 0.30*** | 0.08 | 0.32*** | 0.08 | 0.31*** | 0.08 | 0.32*** | 0.08 |
Mental illness reports | 0.05*** | 0.00 | 0.05*** | 0.00 | 0.05*** | 0.00 | 0.05*** | 0.00 | 0.05*** | 0.00 | 0.05*** | 0.00 | 0.05*** | 0.00 | ||||
Contextual characteristic | ||||||||||||||||||
Detraditionalization | 0.35*** | 0.09 | 0.48*** | 0.10 | 0.38*** | 0.09 | 0.34** | 0.09 | 0.45*** | 0.10 | 0.09 | 0.12 | 0.49*** | 0.11 | ||||
Public Stigma | −0.02* | 0.01 | −0.01 | 0.01 | ||||||||||||||
GDP per capita in ppp | 0.05* | 0.00 | ||||||||||||||||
Unemployment rate | −0.02 | 0.03 | ||||||||||||||||
Divorce rate | −0.00 | 0.01 | ||||||||||||||||
Interaction terms | ||||||||||||||||||
Detrad. * Unempl. | 0.23 | 0.14 | ||||||||||||||||
Detrad. * Fin. Strain | 0.25*** | 0.07 | ||||||||||||||||
Detrad. * Divorce | 0.01 | 0.12 |
. | Model 1 . | Model 2 . | Model 3 . | Model 4 . | Model 5 . | Model 6 . | Model 7a . | Model 8a . | Model 9a . | |||||||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Coef. . | se . | Coef. . | se . | Coef. . | se . | Coef. . | se . | Coef. . | se . | Coef. . | se . | Coef. . | se . | Coef. . | se . | Coef. . | se . | |
Constant | −2.56*** | 0.21 | −4.90*** | 0.26 | −2.89*** | 0.22 | −5.35*** | 0.27 | −4.62*** | 0.36 | −5.29*** | 0.49 | −513*** | 0.36 | −4.89*** | 0.28 | −5.27*** | 0.39 |
Individual characteristics | ||||||||||||||||||
Unemployment | 0.62*** | 0.11 | 0.43*** | 0.11 | 0.61*** | 0.11 | 0.41*** | 0.11 | 0.42*** | 0.12 | 0.43*** | 0.12 | 0.13 | 0.22 | 0.41*** | 0.12 | 0.42** | 0.11 |
Financial Strain | 0.38*** | 0.06 | 0.07 | 0.06 | 0.39*** | 0.06 | 0.08 | 0.06 | 0.10 | 0.06 | 0.11 | 0.06 | 0.08 | 0.06 | −0.21* | 0.11 | 0.08 | 0.06 |
Divorce | 0.57*** | 0.11 | 0.47*** | 0.11 | 0.56*** | 0.11 | 0.46*** | 0.11 | 0.49*** | 0.11 | 0.48*** | 0.11 | 0.44*** | 0.11 | 0.43*** | 0.11 | 0.45* | 0.19 |
Physical Health | −0.53*** | 0.03 | −0.28*** | 0.03 | −0.53*** | 0.03 | −0.27*** | 0.03 | −0.27*** | 0.03 | −0.27*** | 0.03 | −0.27*** | 0.03 | −0.27*** | 0.03 | −0.27*** | 0.03 |
Stigma | −0.06 | 0.09 | −0.06 | 0.09 | ||||||||||||||
Education (low is ref.) | ||||||||||||||||||
Middle | 0.02 | 0.11 | 0.10 | 0.12 | 0.00 | 0.11 | 0.07 | 0.12 | 0.03 | 0.12 | 0.03 | 0.12 | 0.07 | 0.12 | 0.09 | 0.12 | 0.07 | 0.12 |
High | 0.34** | 0.12 | 0.48*** | 0.12 | 0.30* | 0.12 | 0.44*** | 0.12 | 0.42** | 0.13 | 0.42** | 0.13 | 0.44*** | 0.12 | 0.47*** | 0.12 | 0.44*** | 0.12 |
Still Studying | 0.61** | 0.19 | 0.67** | 0.20 | 0.57** | 0.19 | 0.64** | 0.20 | 0.65** | 0.20 | 0.65** | 0.20 | 0.64** | 0.20 | 0.64** | 0.20 | 0.64** | 0.20 |
Age (adult is ref.) | ||||||||||||||||||
Young (15–31 years) | −0.08 | 0.12 | −0.03 | 0.12 | −0.07 | 0.12 | −0.02 | 0.12 | −0.00 | 0.12 | −0.00 | 0.12 | −0.02 | 0.12 | −0.04 | 0.12 | −0.02 | 0.12 |
Old (plus 65 years) | −1.05*** | 0.13 | −0.97*** | 0.13 | −1.06*** | 0.13 | −0.99*** | 0.13 | −0.95*** | 0.13 | −0.95*** | 0.13 | −0.98*** | 0.13 | −0.96*** | 0.13 | −0.99*** | 0.13 |
Woman (men is ref.) | 0.43*** | 0.08 | 0.32*** | 0.08 | 0.43*** | 0.08 | 0.32*** | 0.08 | 0.29*** | 0.08 | 0.30*** | 0.08 | 0.32*** | 0.08 | 0.31*** | 0.08 | 0.32*** | 0.08 |
Mental illness reports | 0.05*** | 0.00 | 0.05*** | 0.00 | 0.05*** | 0.00 | 0.05*** | 0.00 | 0.05*** | 0.00 | 0.05*** | 0.00 | 0.05*** | 0.00 | ||||
Contextual characteristic | ||||||||||||||||||
Detraditionalization | 0.35*** | 0.09 | 0.48*** | 0.10 | 0.38*** | 0.09 | 0.34** | 0.09 | 0.45*** | 0.10 | 0.09 | 0.12 | 0.49*** | 0.11 | ||||
Public Stigma | −0.02* | 0.01 | −0.01 | 0.01 | ||||||||||||||
GDP per capita in ppp | 0.05* | 0.00 | ||||||||||||||||
Unemployment rate | −0.02 | 0.03 | ||||||||||||||||
Divorce rate | −0.00 | 0.01 | ||||||||||||||||
Interaction terms | ||||||||||||||||||
Detrad. * Unempl. | 0.23 | 0.14 | ||||||||||||||||
Detrad. * Fin. Strain | 0.25*** | 0.07 | ||||||||||||||||
Detrad. * Divorce | 0.01 | 0.12 |
*P < 0.05; **P < 0.01; ***P < 0.001.
aResults for the interaction effects are robust if models are controlled for GDP per capita in ppp.
Our main variable of interest, detraditionalization had a significant positive effect on mental health professional care use, supporting the bivariate results and our expectations (Model 2) and this effect was substantial and significant irrespective of respondents’ mental illness reports (Model 4). When we controlled for the level of public stigma, the effect of our measure of detraditionalization remains strong and significant (Model 5). Furthermore, in line with evidence from the literature that a higher level of public stigma acts as a barrier to accessing mental health professional care use, we also report a small negative effect for this contextual variable. When GDP per capita was added (in Model 6), the effect of detraditionalization remains robust. Overall, our results provide evidence for the idea that individuals living in more detraditionalized countries seek mental health professional care more often than individuals living in less detraditionalized countries.
In Models 7, 8, and 9 we estimated the interactions between the level of detraditionalization and our measures of social and material adversity. We found only one significant interaction supporting our expectation, with financial strain, that is, in more detraditionalized societies, people who experienced more financial strain were more likely to access mental health professional care. Figure 4 shows the predicted means of mental health professional care use for individuals experiencing high, medium, and low levels of financial strain for different levels of detraditionalization. The figure shows that in the most detraditionalized countries in our sample the difference in the predicted means of mental health professional care use between individuals experiencing low and high financial strain standard deviation.
Additional analyses
In order to test the robustness of our results, we estimated a series of alternative models (Supplementary Material). In Supplementary Tables A4 and A5, we present estimates from our main models separately for each of the contextual indicators of detraditionalization as well as GDP per capita. Each component of the detraditionalization index had a significant effect on the two dependent variables while GDP per capita only had a significant effect on mental health professional care use. We also re-calculated the dependent variable mental health professional care use to include ‘social workers’ (Supplementary Table A6) in order to capture different institutional arrangements regarding mental health professional care that could influence its use. In addition, we show the results for the models for mental health professional care use excluding from the sample the respondents that used anti-depressants, to account for differences across contexts based on the availability of, and the culture surrounding the use of such medication (Supplementary Table A7). The effect of the detraditionalization index remained the same in both cases, while the significant interaction with financial adversity was not significant anymore in the restricted samples. Next, we found that the effect of detraditionalization on mental illness reports controlled for individual and public stigma was robust (Supplementary Table A8). Furthermore, the interaction between detraditionalization and mental illness reports on mental health professional care use was not significant (Supplementary Table A9). This suggests that individuals with similar levels of mental illness do not make more use of mental health professional care in countries that are more detraditionalized.
Conclusion and discussion
In this paper, we studied differences in mental illness reports and mental health professional care use in Europe. We hypothesized that the level of detraditionalization in a country is a relevant factor for these individual-level outcomes, by itself and in combination with the experience of social and material adversity.
Our results show that, generally speaking, people living in detraditionalized countries are more inclined to use mental health professional care, and that they, on average, report mental illness less than people in less detraditionalized European countries. We also hypothesized that material and social adversity effects are different in more or less detraditionalized countries. Our results do not provide a conclusive picture of all indicators of material and social adversity that we examined. We proposed that in more detraditionalized countries the combination of the meritocratic discourse with the internalization of the individual responsibility for personal biographies results in higher mental health professional care use when individuals experience material and social adversity (Branaman, 2007). However, this argumentation only received support for the case of financial strain. Regarding mental illness, we found that in more detraditionalized countries, individuals who were divorced reported fewer mental illness symptoms compared to people who did not experience this kind of adverse event. Taken together, these results provide some support for arguments derived from theories on medicalization, individualization, and the emergence of psychology as an institution in late-modern societies.
Another general conclusion that emerges from our study refers to the relationship between the prevalence of mental illness reports and mental health professional care use, that is in those countries where mental illness reports were higher, mental health professional care use was lower. Our findings suggest systematic patterns of overuse and underuse of mental health professional care between countries. Other explanations could be linked to the measurement of the two variables. First, the mental illness reports items refer to the previous weeks, while the measurement of mental health professional care seeking refers to the previous year. Presumably, seeking care positively impacted mental illness reporting.
Second, we used a measure of mental illness reports that could suffer from desirability problems, that is individuals would be less compelled to report mental illness due to stigma but due to the availability and extension of mental health professional care in more detraditionalized countries, they would more easily access such services. We were able to control for the level of stigma both at the individual and contextual level and the conclusion that in more detraditionalized countries individuals report less mental illness but seek more mental health professional care still holds. Furthermore, higher mental health professional care use among more detraditionalized countries could be explained by the culture of disclosure and the legitimization of seeking to achieve well-being as maintained by scholars that focused on the emergence of psychology as an institution (Wright, 2008). Still, the inverse relationship with mental illness reports is not explained by such arguments and remains open for investigation.
Two sets of findings evoke some discussion and can lead to more research. First, we elaborated in the theory section that the generally more favorable living circumstances in detraditionalized countries are expected to relate to lower levels of mental illness reports. This expectation contradicted scholars who warned that levels of anxiety and depressive symptoms could be higher in late-modern societies due to processes of individualization and medicalization (Horwitz, 2011; Furedi, 2004; Branaman, 2007). We found that in detraditionalized societies the level of mental illness is lower, even when we control for the country’s wealth. We did not find a statistically significant and strong association between GDP levels and mental illness reports in models where the two contextual measures are simultaneously accounted for. That wealth has ceased to play an important role in health in rich societies is not a new result, that is previous research has shown that other contextual characteristics are more relevant in these societies, for example income inequality (Van Deurzen, 2015), or trust and social cohesion (Jen et al., 2010; Layte 2012). Our findings add to this body of literature and show that the cultural aspects, as captured by our measure of detraditionalization, are more important for the level of mental illness in late-modern societies than the overall level of societal wealth.
Second, we turn to the remarkable differences between the effects of material and social adversity in detraditionalized societies. While we found that those experiencing financial strain use mental health professional care more in detraditionalized countries, they do not report more mental illness symptoms. Since the measure of mental illness reports refers to the last 2 weeks before the interview, while the questions regarding mental health professional care use refer to the previous year, these findings could be explained by the potential effectiveness of therapy. It is also possible that in more detraditionalized countries healthcare is more affordable and therefore the difference in use is particularly noticeable for this income bracket when compared to more traditional countries where health care access is stronger linked to material resources. We also note that these results are derived from models where mental illness reports are accounted for, a modeling strategy that reflects previous operationalization of medicalization (Buffel, Beckfield and Bracke, 2017). Subsequently, our results also suggest the medicalization of poverty in more detraditionalized societies. Future research should be able to shed a light on the tenability of these explanations.
When we turn to the differential effect of divorce, one of our measures of social adversity, we found that the difference in mental illness reports between those who experienced divorce and those who did not is smaller in more detraditionalized countries. We note that this finding is robust when we control for the societal divorce rates, as an indicator of the normalization of divorce. We explain this finding by referencing the ethos of finding happiness in the here and now, the ethos that is characteristic of detraditionalized societies. Even though leaving a marriage that is not satisfying anymore is a stressful life event, it can be interpreted as a step toward finding this happiness and less as a sign of personal failure. In less detraditionalized countries, however, the sanctity of marriage is valued more strongly and not being able to live up to this norm seems to cause more anxiety for divorced individuals and be reflected in their levels of mental illness.
An alternative explanation for this result points toward the different institutional arrangements surrounding the dissolution of marriage in late-modern societies in comparison to the more traditional ones. We refer here to the types and levels of income support available for the couple’s members or the types of arrangements linked to child support after marital dissolution. In more detraditionalized and wealthier countries, such more advantageous and supportive institutional arrangements could buffer the negative impact of divorce on mental illness reporting, which could explain our findings. Since we did not directly measure such institutional arrangements (even though to some extent they are captured by the measure of the wealth of a country), future research is needed for disentangling cultural and institutional mechanisms at work.
This said it appears that the relationship between adversity and mental illness reports and mental health professional care use in countries with various levels of detraditionalization has to be evaluated on a case-by-case basis.
Two issues have limited our study and call for more research. First, the measure of mental health professional care covers the previous year while the measure of mental illness reports covers the weeks before the interview. As such, we do not know whether the levels of mental illness reporting have to do with the treatment received. However, if this is the case, actual levels of mental illness reports in all countries in our sample would be higher. Most likely our effects would be stronger but the conclusions would still hold. Future research should be able to shed light on this matter. Second, our data does not include direct measures of religiosity, autonomy, or gender values at the individual level, so we cannot exclude compositional effects that could weaken the contextual ones. However, we were able to include important individual-level variables, that is education and age, that were shown to be good predictors of the values held by individuals (Halman and Draulans, 2006; Sieben and Halman, 2014). In addition, we also included additional contextual measures that accounted for the different cultural influences that could confound our results, that is the stigma surrounding mental illness, normalization of divorce, and unemployment. Future research endeavors using a richer data with respect to religiosity and values reflecting autonomy and self-direction can confirm that our results are robust more directly than we did.
To sum up, we found evidence that in more detraditionalized societies individuals report lower levels of mental illness but higher mental health professional care use. Furthermore, not all forms of adversity result in higher levels of mental health professional care use in the more detraditionalized countries, as scholars that examine processes of the medicalization of society caution. This seems to be the case only for financial strain, while for unemployment and divorce this was not the case.
References
Footnotes
Throughout the paper, when we use the terms “accounts of mental illness” or “mental illness reports”, we refer to “self-rated mental illness”. See “Data and methods” section for an explanations on measurement.
Further information about Eurobarometer can be found at https://europa.eu/eurobarometer (accessed 14 June 2021).
We performed a robustness check, using the Penalized Maximum Likelihood Estimation available in STATA (known as Joseph Coveney’s firthlogit) to assess if the low number of cases of our dependent variable mental health professional care use leads to biased results. We follow the instructions from this handout about analyzing rare events with logistic regression analysis: https://www3.nd.edu/~rwilliam/stats3/RareEvents.pdf (accessed 8 November 2022). The results of the fixed effects model show that this made little to no difference to the effects compared to a ‘normal’ fixed effects logistic regression.