Social Inequalities in Mortality in Belgium: "Causineq" Project

1. 1. Social inequalities in mortality in Belgium from 1991 to 2016 "Causineq" project: "Causes of inequalities of health and mortality in Belgium (multiple dimensions, multiple causes)". Thierry Eggerickx, Jean-Paul Sanderson, Christophe Vandeschrick , Luc Dal, Joan Damiens

2. Introduction • An observation: while life expectancy is increasing in Western countries, the differences between 'social groups', 'regions' and 'places of residence' remain, or even increase. • A main question: how have social inequalities in mortality evolved in Belgium in recent decades? • Twosecondaryquestions: • Do social inequalities in mortality also affect older people ... and how have they evolved? 2. What is the role played by health status on social differences in mortality at older ages?

3. 1. Data and methods 1.1. The data Linkage the individual data of the National Register (1991-2016) with those of the censuses of 1991, 2001 and 2011 and those of the death certificates of the civil registration(Thank’s to Statbel) 1.2. The methods • Calculation of life tables by sex, region, district and social group. • Use of logistic regression models on the probability of death within 5 years of the census.

4. Social groups • Why ? Different dimensions are used to determine the social positioning of individuals: educational level, socio-professional status, housing conditions and income. But : • they can have a different role on health status and mortality : EducationallevelAttitudes of prevention, recourse and access to health care Income, socio-professionalstatus, housing conditions materialresources • the correlation between these different dimensions is not optimal: • How ? • Three dimensions: educational level, socio-professional category and housing characteristics. • Scoring method. Each individual is assigned a score according to their position on each dimension. The score varies from 1 to 10. • Assignment of the parents' score to the children. • Each individual is positioned on a social continuum, but to synthesize information, 4 groups were identified by quartile score.

5. 2. The main results: • Important inequalities in mortality by social group in 2011-2015 Privileged SG(25%) versus Underprivileged SG (25%): E0 : 9.1 years difference for men (73,9 / 83,0)and 6.6 years for women.(80,4 / 87,0 ) Between GS Favored (5%) and GS Disadvantaged (5%): E0 : 13 years difference for men (70,2 / 83,0) and 10 years for women (76,6 / 87,0). . • Life expectancy at birth, at 25 years, at 65 years or at 80 years varies according to the social gradient • The gap between female and male life expectancy is greater for the underprivileged SG (6.5 years) than for the privileged SG (4 years)

6. Life expectancy at 25 years (men) in Belgium according to different population categories • A significant delay of underprivileged social groups on the "road of health transition". • Multidimensional indicator maximizes inequalities in mortality

7. Mortality quotient ratios (‰) by age, 2011-15 (total population = 100) • According to the ‘mortality quotient ratio’, the excess mortality of the underprivileged SG is observed at all ages, but especially between 25 and 50 years old and for young children Men Women

8. Hommes Femmes Ages Absolu Relatif Absolu Relatif 0-19 0,42 4,6% 0,31 4,6% 20-39 1,13 12,4% 0,55 8,4% 40-59 3,21 35,3% 1,93 29,3% 60-79 3,54 38,8% 2,30 34,9% 80+ 0,81 8,8% 1,50 22,8% 9,1 1 6,6 1 According to the contribution of age groups to differences in life expectancy at birth between social groups(Arriaga method) Contribution of age groups to differences in life expectancy at birth between privilege and underprivileged social groups (2011-2015) • Among men: significant contribution of people aged 40-59 (35%) and 60-79 years (39%) • For women, significant contribution of people aged 60-79 (35%) and over 80 (23%)

9. Life expectancy gains are greater for the privileged social group ... and social inequalities in the face of death increased between 1992/96 and 2011/15 The life expectancy of women over 50 years of age in the most underprivileged social group (5%) decreased between 1992/96 and 2011/15. The increase in social inequalities in mortality affects all ages and varies across the social continuum For each major cause of death, the risk of dying is higher among the disadvantaged than among the most advantaged. This observation applies to both men and women and to all age groups. The greatest differences are for diseases of the circulatory and respiratory systems

10. 2. Are there any social differences in mortality beyond age 65? • Do social inequalities in mortality also affect the elderly... and how have they evolved? • What is the role of health status? For an equivalent state of health, do differences in mortality between social groups persist? 2 assumptions: - they persist in a process of accumulation of beneficial or harmful effects in the life course - they disappear following a selection process that gradually "eliminates" the weakest.

11. Social inequalities in the face of death beyond age 65 or age 80 are significant and have increased over the last 25 years for both women and men. • Nevertheless, the intensity of the differentials is lower at 80 than between 65 and 79 years of age, for both men and women. In other words, social inequalities in mortality persist at oldest ages, but it is likely that a selection effect contributes to the reduction of the social differential beyond 80 years of age

12. Health status is an important determinant of risk of death, especially for older people • Underprivileged social groups are characterised by a much worse state of health than the most privileged It could therefore legitimately be assumed that social inequalities in mortality at older ages are essentially the consequence of differential health status across social groups. If this is the case, for the same state of health, there should be no difference in mortality according to the social group. • Health status captures some of the effect of social inequalities on mortality, but does not eliminate it, both for people aged 65 to 79 and for those aged 80 and over. • The poorer the health status, the lower the social inequalities in mortality. • The part of the differential mortality of older people that is not explained by health status is therefore due to social differences (resources, behaviour) to which are added other factors, such as differences in medical consumption (level of expenditure, nature and quality of care), access to health care or the social and family environment. Relative risk factors of dying within 5 years of the census (2002-2006). Stratified models by health status

13. 2. Spatial differences in mortality by social group "Causineq" project: "Causes of inequalities of health and mortality in Belgium (multiple dimensions, multiple causes)". Thierry Eggerickx, Jean-Paul Sanderson, Christophe Vandeschrick , Luc Dal, Joan Damiens

14. Introduction Yesterday, as today, there are differences in mortality at the level of regions, districts, municipalities... and these differences do not disappear. In 2011-2015, the difference in life expectancy at birth between the districts is 6 years for men and 4 years for women. Male life expectancy at birth by district 1992-1996 2011-2015

15. For the same social group, are there still differences in mortality at the regional and sub-regional level? Differences in life expectancy at birthbetweenFlanders and Wallonia by social group and sex (2011-2015) 1. At the regionallevel • For all social groups, there are regionaldifferences in mortality • Thesedifferences are more important for the underprivilegedsocial group than for the privilegedsocial group

16. 2. At the district level ? • There is a relationship between the level of mortality and the socio-economic characteristics of the population of each 'arrondissement' The relationship between male life expectancy at birth and median income at the district level

17. But for the same social group, life expectancies still have important differences Male life expectancy at birth by ‘arrondissement’ and social group (2012-2016) Privileged SG Underprivileged SG • The social environment has a positive effect on health and mortality where the privileged groups are better represented and a negative effect where they are in the minority • Other factors come into play: the physical, social and institutional environment, historical and 'cultural' factors that go beyond social belonging, supply and quality of health services that are spatially differentiated

18. Conclusions One of the main contributions of this research is the creation of the database, its reliability and optimization for mortality studies. This database allows multiple questions, which go well beyond Causineq and must generate new research dynamics (... and therefore new funding). This database will continue to be populated with recent data until 2025 and matching of 1981 census data is underway.

19. But two important limits = leads for future work • The district scale, which is very large and includes ‘environments’ that are highly disparate, is not the most appropriate for measuring the respective effects of these factors. • It would be important to take into consideration the selective effect of migration and residence times