CAUSINEQ. Causes of health and mortality inequalities in Belgium: multiple dimensions, multiple causes,

Damiens, Joan Jany [UCL] Eggerickx, Thierry [UCL] Gourbin, Catherine [UCL] Majérus, Paul [UCL] Masquelier, Bruno [UCL] Sanderson, Jean-Paul [UCL] Vandeschrick, Christophe [UCL] Van Cleemput, Océane [UCL] et al.

In Belgium, life expectancy is now twice what it was 170 years ago due to improvements in public and private hygiene, nutrition, medical procedures and the organization of health services. The average lifespan of a Belgian resident is now over 80 years, but inequalities persist and sometimes even intensify. In Belgium, like in other European countries, a negative relationship has been demonstrated between socio-economic position on the one hand and health and mortality on the other hand. This project aimed to investigate whether or not this social gradient in health and mortality is associated with the growing de-standardisation of employment arrangements and family situations. The goal was to obtain information about the mechanisms by which social differences in mortality and health are generated in Belgium, in order to provide insight in policy measures that could prove effective in countering these inequalities. In a first part of the project, the evolution of mortality inequalities in Belgium between 1991 and 2016 was studied. It has been shown, based on a multidimensional indicator combining educational level, employment category and housing characteristics, that social inequalities in death are significant in Belgium and that they have been increasing for both women and men since at least the 1990s. In relative terms, gaps between social groups are especially wide between 25 and 50 years but have little impact on differences in life expectancy because the risk of dying is inherently low at these ages. The social disparities in mortality for older people (65+) – and especially the delay of the underprivileged in the health transition and the fight against cardiovascular diseases – have a far greater effect on differences in life expectancy. In general, social inequalities in mortality and their variation along the social continuum are observed for each of the major causes of death, but they are particularly pronounced for diseases of the respiratory and circulatory systems. The results also point out that spatial disparities in mortality linger and have been worsening for at least a quarter century. The spatial pattern of mortality for the underprivileged social group resembles the spatial patterns of the privileged and intermediate social groups. Within the same social group, spatial disparities endure. This indicates that beyond the socioeconomic composition of regions and districts, other factors – environmental, cultural and behavioural – are involved and affect mortality across social groups in the same way. The second part of the project focused on the relationship between employment situation and (cause-specific) mortality. Mortality inequalities by employment status were considered, controlled for other dimensions of socio-economic position. The association between mortality and unemployment in Belgium was investigated in detail at the individual level. Results showed that the unemployed have a two times higher mortality risk than the employed. Even when they have a high educational level, excellent housing conditions and healthy living arrangements, unemployed men still have a higher mortality risk than their employed counterparts. Conclusions for women were comparable, but inequalities were smaller. A slight protective effect of education against the detrimental health impact of unemployment was demonstrated. Analyses clearly showed that the mortality excess of the unemployed results from an excess mortality in practically all main cause groups. Focusing on specific causes within each main group, we observed large inequalities for alcohol-related mortality. Furthermore, our results illustrated the importance of municipality of district characteristics, such as urbanicity, the aggregated unemployment rate, etcetera. The mortality excess for unemployed men and women was for example smaller in regions with higher aggregate unemployment levels. The higher mortality of the unemployed suggests a positive effect of having a job, which is a reasonable assumption since employment is a key factor determining the financial and psychological wellbeing of individuals and their families. Getting people into work should therefore be an important goal of labour market policies and is of critical importance for reducing health inequalities. However, this is not the whole picture. Analyses including non-standard types of contract show that jobs need to be sustainable as well. Overall and cause-specific mortality differentials by marital status and type of family situation were the focus of a third part of the CAUSINEQ project, in order to gain insight into the effect of the de-standardisation of family formation processes. With regard to marital status, it does appear that matrimonial behaviour varies from one Belgian administrative region to another. Flanders is characterised by a greater share of married couples and a lower proportion of divorced and single people. In contrast, isolated individuals are overrepresented in Brussels, compared to Flanders in particular, but also compared to Wallonia. At the same time, married individuals and those who are not married but live in a couple do seem to be subject to lower levels of mortality in each of the three regions of Belgium. Results from Poisson regression models point out the interactions between the undermortality of (married) couples and the differences in matrimonial behaviour between the three regions are relatively small and do not provide a satisfying explanation for the differences in mortality observed between the regions. The main factor explaining regional differences in mortality is the socio-economic situation of the residents. With regard to type of family situation, excess mortality was revealed for children under 5 years of age living in single-parent families, even after controlling for the main socio-demographic variables. This excess mortality is particularly pronounced for violent deaths. This observation raises questions. Is this a manifestation of less parental supervision or is it a reflection of less safe housing or living conditions? In order to characterise violent deaths in single-parent families and to distinguish them from those in two-parent families, Chi-square tests were used with the place of death (home, school, etc.), the type of death (traffic accident, other accidents, etc.) and the day of the week in which the death occurred. However, none of these analyses was statistically significant. Finally, analyses were performed to show how adverse health – as an important precursor of mortality – varies jointly by employment and living arrangements. A typology of labour market positions was constructed and enriched with survey-data information about respondents’ employment conditions (for those in waged employment). This endeavour shows Belgium as a country with a stable, low prevalence of precarious employment when compared to other EU-countries. However, the prevalence of one particular type of precarious employment is considerable and growing in Belgium, namely ‘precarious unsustainable employment’, characterised mainly by low working hours and low monthly wages and often held by female workers. The labour market typology shows clear relationships with the health of individuals. Three labour market positions are most likely to be associated with general and mental health problems: unemployment, precarious employment and instrumental employment. Controlling for social precarity indicators causes a (sometimes spectacular) reduction in the odds for poor health, which clearly shows how the broader social situation “interacts” with labour market positions in creating social health inequalities. It is therefore important to keep in mind that the accumulation of health-damaging positions in different life spheres is in fact a worrisome reality for a part of the Belgian population. These results indicate that precarious labour market situations are to be taken serious as a public health risk and suggest that stable and secure employment of good quality is the “healthiest form of employment”. From a policy perspective, our findings make clear that the potential health impact of labour market policies should be considered whenever labour market reforms are planned. Policy makers should be aware of the fact that flexible labour market policies may stand at odds with policies aiming for longer and sustainable working careers.