Associations between income inequality at municipality level and health depend on context – A multilevel analysis on myocardial infarction in Sweden
Introduction
It has been suggested that the association between income inequality and health is weaker or non-existent in more egalitarian countries, such as the Nordic countries (Lynch et al., 2004, Subramanian and Kawachi, 2004). However, more recent Scandinavian studies challenge this suggestion. A Norwegian study concludes that there is a relation between income inequality and all cause mortality at the Norwegian regional level (Dahl, Elstad, Hofoss, & Martin-Mollard, 2006).
In a previous paper (Henriksson, Allebeck, Weitoft, & Thelle, 2007) we reported some evidence of an association between income inequality and all cause mortality, albeit a differential association, depending on social position such that unskilled manual workers seemed to have a higher risk of dying when living in municipalities with a higher degree of income inequality, while high level non-manual employees seemed to have a lower risk.
In this paper we wish to explore whether these findings may be consistent with the proposal that income inequality might be an indicator and determinant of the scale of the socioeconomic stratification (Wilkinson & Pickett, 2007), even in smaller areas in more egalitarian contexts. We do this by picking up on Wilkinson’s suggestion that what matters most is the extent of inequality between different population strata into which inequality divides the society (Wilkinson, 1997).
We want to find out whether higher levels of income inequality at the municipality level are associated with higher degrees of residential segregation measured as higher degree of homogeneity with respect to wealth within parishes, as suggested by e.g. Kawachi (2002) and Massey (1996). Kawachi, 2002, Massey, 1996).
Furthermore, we are interested in how interactions between factors at municipality, parish and individual levels impact on individuals’ health. We are especially interested in whether economic and ethnic homogeneity at the parish level, as a proxy for neighbourhood, is related to income inequality at the municipality level and if those relations are linked to health outcomes at the individual level (Fritzell, 2005).
This can be formulated as a hierarchical three level problem, with individuals living in parishes (as proxy for neighbourhoods), which in turn are organised within municipalities which are administrative and political areas with a high level of autonomy from the state. Municipalities have elected councils, they decide on taxes, they are obliged to organise and finance core welfare institutions, such as schools, care for the elderly, and social welfare services.
Residential segregation in Sweden is driven by both socioeconomic and ethnic mechanisms (Andersson, 1998). During the past decades, it has been characterised by a divide between on the one side people with moderate to high incomes and, on the other, people of non-Swedish origin together with people of Swedish origin with a very low level of material resources. This is the result of a migration process during the last two decades, where typically Swedish households with low to moderate incomes have moved out from sub-urban areas leaving behind households of non-Swedish origin but also households of Swedish origin with very low incomes, with high rates of unemployment and strong need of social support. This pattern of migration has also, unintentionally, been supported by governmental housing and economic policies, e.g. subsidising construction of owner-occupied housing, mortgages costs can be used for tax-reduction which was beneficial for middle- and high-income households (Andersson, 1998).
Most studies on segregation and health have dealt with racial and ethnic segregation. We have found a few papers that discuss the association between income segregation and health (Acevedo-Garcia and Lochner, 2003, Jargowsky, 1996, Kawachi, 2002, Lobmayer and Wilkinson, 2002). Lobmayer and Wilkinson (2002) found that segregation within urban areas in the United States was associated with additional mortality risk but that the relation between income inequality and mortality was independent of the degree of income segregation.
Since we were interested in the detrimental psychosocial effects from social stratification on individuals’ health, we chose acute myocardial infarction (AMI) as an outcome.
Ischemic heart disease has been shown to be associated with a number of psychosocial factors, acute and chronic stressors at individual and contextual levels. Evidence from different sources such as animal studies, epidemiological studies, and clinical studies imply that such factors may have significant effects on the organic manifestations of coronary artery disease (Chandola et al., 2008, Kolegard Stjarne et al., 2002, Krantz and McCeney, 2002, Marmot and Brunner, 2005). Laboratory and clinical studies have shown that detrimental psychosocial factors induce a number of physiological effects. Stressors are known to trigger hemodynamic, endocrine and immunologic mechanisms which provoke the progression of atherosclerotic processes in the arteries (Everson-Rose & Lewis, 2005).
This study aims to disentangle the effects of income inequality and residential segregation, measured as parish level homogeneity, on AMI. We hypothesise that a) increasing income inequality at the municipality level increases the risk for Acute Myocardial Infarction (AMI); b) the association between income inequality and AMI is mediated by the level of residential segregation, i.e. homogeneity at the parish level dilutes the strength of association between income inequality at the municipality level and health; c) groups with lower compared to those with higher social position are at greater risk for AMI when living in more disadvantaged areas and the risk ratio between disadvantaged and affluent groups increases with increasing affluence in the area, i.e. there is an interaction between contextual parish characteristics and individual level social position.
Section snippets
Settings
Sweden has three major metropolitan areas, Stockholm, Gothenburg and Malmö, where about 3million out of the total national population of 9 million people live; there are about 40 municipalities with more than 50 000 inhabitants each and 250 municipalities with less than 50 000 people.
Each of these municipalities is divided into parishes, the larger the city the larger the number of parishes. Parishes are old ecclesiastical divisions without any current administrative functions but they
Results
Results from the regression analyses are shown in models A–E in Table 3 (with municipality and parish level estimates) and models F–I in Table 4 (with individual level and cross level interaction variables added).
All models were adjusted for mean income at the municipality level and for age and sex. Model A is a “base model” showing relative risk (RR) for AMI with 95% confidence interval, for the average individual living in the average parish in the municipality with highest income inequality
Main findings
We found an overall association between income inequality and incidence of AMI. The association was strongest in the largest urban areas. Quite contrary to our hypothesis, the risk for AMI was lower in the municipalities with a higher degree of income inequality (Table 3, model A). The type of segregation seemed to be important. Segregation of high-income earners (model B) diluted, but did not eliminate, the association between income inequality and risk of AMI – the degree of parish affluence
Conclusions
Our findings suggest that there is an inverse association between income inequality and AMI at the municipality level in Sweden. The association was stronger in larger urban areas. We found that residential segregation of affluent households might mediate part of the association between income inequality and health at the municipality level. Those high level non-manual workers living in affluent parishes had a decreased risk compared to their colleagues living in less affluent parishes, whereas
Acknowledgements
We thank Alastair Leyland for fruitful discussions and comments on the manuscript.
References (39)
- et al.
The effects of regional characteristics on alcohol-related mortality–a register-based multilevel analysis of 1.1 million men
Social Science & Medicine
(2004) - et al.
For whom is income inequality most harmful? A multi-level analysis of income inequality and mortality in Norway
Social Science & Medicine
(2006) - et al.
Income-related inequality in life-years and quality-adjusted life-years
Journal of Health Economics
(2000) - et al.
Whose health is affected by income inequality? A multilevel interaction analysis of contemporaneous and lagged effects of state income inequality on individual self-rated health in the United States
Health Place
(2006) - et al.
Does the state you live in make a difference? Multilevel analysis of self-rated health in the US
Social Science & Medicine
(2001) - et al.
The problems of relative deprivation: why some societies do better than others
Social Science & Medicine
(2007) - et al.
Residential segregation and health
Socio-spatial dynamics: ethnic divisions of mobility and housing in post-Palme Sweden
Urban Studies
(1998)- et al.
Work stress and coronary heart disease: what are the mechanisms?
European Heart Journal
(2008) Measuring inequality
(1977)