Income-related health inequalities in the Nordic countries: Examining the role of education, occupational class, and age

Abstract

Numerous studies have concluded that people’s socioeconomic position is related to mortality and morbidity, but that the strength of this association varies considerably both within and between European regions. This has spurred several researchers to more closely examine educational and occupational gradients in health in the Nordic countries to clarify the causes of cross-national differences. However, comparable studies using income as an indicator of socioeconomic position are still lacking. This study uses recent and highly comparable data to fill this gap. The aim of this study is threefold. First, we ask to what extent there is an income gradient in health in the Nordic countries, and to what extent the association differs between these countries. Second and third, we examine whether differences in the attenuation of the income gradient by education and occupational class, and age-specific differences between countries, may act as explanations for differences in the income gradient between the Nordic countries. The data source are three waves of the European Social Survey (ESS, 2002/2004/2006), which included 17,801 people aged 25 and over from Denmark, Finland, Norway, and Sweden. Two subjective health measures (physical/mental self reported health and limiting longstanding illness) were analysed by means of logistic regression. The results show that, in all countries, people reported significantly better health and were less likely to suffer from longstanding illness as they had a higher income. This association is strongest in Norway and Finland and weakest in Denmark. The income gradient in health, but not country differences in this gradient, is partly explained by education and occupational class. Additionally, the strength of the income gradient in health varies between age groups. The relatively high health inequalities between income groups in Norway and Finland are already visible in the youngest age groups. The results imply that the socioeconomic gradient in health will arguably not be strongly reduced in the near future as a result of cohort replacement, as has been suggested in previous studies. Health policy interventions may be particularly important five to ten years prior to retirement and in early adulthood.

Introduction

During the last few decades, numerous studies have examined socioeconomic inequalities in mortality and morbidity in Europe (Avendano et al., 2006, Cavelaars et al., 1998, Kunst et al., 2005, Kunst et al., 1999, Mackenbach et al., 2003, Mackenbach et al., 2008, Mackenbach et al., 1997, Mackenbach et al., 1999, Mackenbach et al., 2005, Mackenbach et al., 2008). Generally, this has lead to the conclusion that people’s socioeconomic position is consistently related to their health status, but that the strength of this association varies considerably both within and between European regions. Probably the most puzzling finding from these studies is that socioeconomic inequalities in health are not among the weakest in the Nordic welfare states (i.e., Denmark, Finland, Norway, and Sweden) as compared to other European regions (for instance, socioeconomic inequalities in health are particularly low in Southern Europe, and even virtually absent in the Basque region) (Aberg Yngwe et al., 2001, Cavelaars et al., 1998, Dahl et al., 2006, Eikemo, 2009, EUROTHINE, 2007, Lahelma et al., 2000, Lundberg et al., 2008, Rahkonen et al., 2000, Rahkonen et al., 2002). This may come as a surprise, as one could expect that socioeconomic inequalities in health would be less dominant in the Nordic countries, since strong emphasis is put on equality of results, such as economic resources, as opposed to equality of opportunity in these nations (Esping-Andersen, 1990, Fritzell, 2001). Combined with the fact that substantial differences within the Nordic welfare regime type have been found in the strength of the socioeconomic gradient in health, this has prompted several researchers to more closely examine the differences between these countries, since by doing so more insight might be obtained about the reasons why some countries have smaller socioeconomic gradients in health than others.

Up until now, several articles have focused on mapping differences between and within the Nordic countries in the strength of the socioeconomic gradient in health and examining the underlying causes (Lahelma et al., 2002, Lahelma et al., 2001, Silventoinen and Lahelma, 2002). Generally, socioeconomic inequalities in health were found to be strongest in Finland and weakest in Denmark, with Norway and Sweden taking in intermediate positions. Most notably, these articles have demonstrated that these differences are largely attributable to the fact that the older cohorts have experienced different circumstances during the formative years than the youngest age groups. First, this would explain why the socioeconomic gradient in health is particularly steep in Finland: the post-war years were especially tough in this country. Second, these studies have ascribed the relatively weak socioeconomic gradient in health in Denmark to health-damaging behaviour (e.g., smoking and alcohol use) being more equally prevalent across social groups than in the other Nordic countries. These findings suggest that socioeconomic gradients in health may become more similar across the Nordic countries in the near future: the post-war cohorts in Finland will be replaced by cohorts that grew up in more favourable circumstances, and health-damaging behaviour will probably become more socially selective in Denmark as well. All in all, this would imply a pattern of convergence between the Nordic countries in the next few decades.

However, studies examining the socioeconomic gradient in health in multiple Nordic countries have mostly reported on the causes of health inequalities by education and occupational class. The third prominent indicator of people’s socioeconomic position (income) has been largely disregarded, probably due to the lack of sufficiently comparable income data for all the Nordic countries. Although education, class, and income are correlated, they are also strongly associated with health independently as they measure different phenomena and tap into different causal mechanisms (Davey Smith et al., 1998, Geyer et al., 2006, Rognerud and Zahl, 2006). Education is often regarded as a non-material resource that promotes a healthy life-style, and occupational class is usually considered to reflect mostly material conditions which are related to paid work, working conditions, and social status. Income, on the other hand, indicates the availability of material resources and purchasing power, which determine to what extent people are able to improve and maintain their health. Additionally, income may influence health through non-material mechanisms, by inducing feelings of relative deprivation and frustration in those who are at the bottom of the income scale. Previous research has demonstrated that income is related to health in several ways (Benzeval et al., 2001, Dahl, 1994a, Fritzell et al., 2004, Martikainen et al., 2001, Wagstaff and Van Doorslaer, 2000). Studies focusing on individual health differences within wealthy nations have shown that there is a strong and consistent gradient along the whole income hierarchy (Adler et al., 1994).

The fact that education, class, and income influence health through different pathways implies that differences between the Nordic countries in income inequalities in health may not be equal to differences in educational and class inequalities in health found in earlier studies on these countries. It is therefore an open question whether the pattern of convergence between the Nordic countries observed in earlier work is also found when income gradients in health are considered. Moreover, because of these differential underlying mechanisms, differences between the Nordic countries in income inequalities may have different policy implications than differences in educational and class inequalities in health. Although previous work on education and class may be used to draw conclusions on the utility of modifying the educational system or working conditions, these studies cannot be used to assess whether changes in the distribution of financial resources and goods may reduce health inequalities. Therefore, an examination of income inequalities in health in the Nordic countries in addition to the education and class gradients investigated in earlier work is necessary to obtain a comprehensive overview of socioeconomic inequalities in health in the Nordic countries and the underlying causes. Eikemo (2009) examined income inequalities in health in several European welfare regimes, including the Nordic regime, but did not differentiate between the four Nordic countries. In this study, we contribute to existing knowledge in this field by examining the income gradient in health in all four Nordic countries. In order to do so, we use data from the European Social Surveys (2002, 2004, and 2006). The high comparability of these data in terms of sampling design, survey questionnaires, and measurements of income and health allow us to be the first study to rigorously compare income inequalities in health between all Nordic countries. Additionally, these data were recently collected as compared to data sources used in earlier studies examining income inequalities in health. In sum, the first research question is to what extent there is an income gradient in health in the Nordic countries, and to what extent the Nordic countries differ in this respect.

In addition to mapping differences in income inequalities in health between the Nordic countries, this study aims to examine the underlying causes of these differences. A Finnish study has examined the causal interdependencies between education, occupational class, and income (Lahelma, Martikainen, Laaksonen, & Aittomaki, 2004). They found that two thirds of income inequalities in self-rated health were explained by education and occupational class. With regard to income-related inequalities in longstanding illness on the other hand, only a small part could be explained by education and occupational class according to this study. In the present study, we examine to what extent this Finnish finding is also present in the other Nordic countries. Differences between the Nordic countries in the extent to which the income gradient in health is explained by education and occupational class may have important policy implications for the approach that may best reduce socioeconomic health inequalities. After all, if patterns of income inequality in health are mostly due to the material and non-material factors associated with income itself this calls for a different approach than if these differences are in fact caused by educational or class inequality underlying this income gradient. Therefore, our second research question is to what extent the income gradient in health is differentially attributable to educational and occupational health inequalities in these four countries. It should be noted that income may also causally precede education and occupational class, instead of being a consequence of educational and occupational success. Although research empirically confirming the causal order of these socioeconomic characteristics is still largely lacking, Lahelma et al. (2004) argue that income is a consequence rather than a cause of education and occupational class. Given that education and occupational class are mostly acquired early in the life course, and that parental income rather than current own income determines access to education, we will assume in this study that income is mostly a consequence of education and occupational class. Finally, it should be kept in mind that access to all levels of education is affordable in the meritocratic Nordic countries.

Furthermore, studies investigating educational and occupational gradients in health in the Nordic countries have found that socioeconomic health inequalities vary between age groups (Lahelma et al., 2002). In our study, we examine to what extent this also holds for income inequalities in health. Country differences in educational inequalities in health appear to converge in the youngest cohorts (Silventoinen & Lahelma, 2002). However, this may not necessarily be true for country differences in the income gradient in health, given that education and income tap into different causal mechanisms (Geyer et al., 2006). Consequently, our third research question extends earlier work on educational and occupational gradients in health by asking whether income-based health inequalities vary between different age groups, and whether country differences in the income gradient in health are weaker for the youngest cohorts. This may clarify whether specific age- and cohort-related circumstances are important in explaining income inequalities in health. Additionally, comparing age groups may help to predict whether income inequalities and cross-national differences in the income gradient will persist in the coming decades.