You are here

  • Home
  • Archive
  • Volume 66, Issue 7
  • Socioeconomic inequalities in circulatory and all-cause mortality after retirement: the impact of mid-life income and old-age pension. Evidence from the Uppsala Birth Cohort Study

Article Text

Article menu

Download PDFPDF

Research report
Socioeconomic inequalities in circulatory and all-cause mortality after retirement: the impact of mid-life income and old-age pension. Evidence from the Uppsala Birth Cohort Study
  1. Stefan Fors1,2,
  2. Bitte Modin3,
  3. Ilona Koupil3,
  4. Denny Vågerö3
  1. 1Department of Social Work, Stockholm University, Stockholm, Sweden
  2. 2Aging Research Center, Karolinska Institutet/Stockholm University, Stockholm, Sweden
  3. 3Center for Health Equity Studies, Stockholm University/Karolinska Institutet, Stockholm, Sweden
  1. Correspondence to Stefan Fors, Aging Research Center, Karolinska Institutet/Stockholm University, Gävlegatan 16, SE-113 30 Stockholm, Sweden; stefan.fors{at}ki.se

Abstract

Background The aim of this study was to explore the impact of mid-life income and old-age pensions on the risk of mortality in later life. Furthermore, the study explored whether income inequalities in old-age mortality can be explained by differences in early childhood development, social class during childhood, education or marital status.

Methods The study sample comprises all individuals born at Uppsala Academic Hospital during the period 1915–1924 who had retired but not died or emigrated by 1991 (n=4156). Information on social and biological conditions was retrieved from national registries.

Results The results show that income during mid-life and income during retirement were associated with old-age mortality. However, mutually adjusted models showed that income in mid-life was more important for women's late-life mortality and that income during retirement was more important for men's late-life mortality. Furthermore, differences in education and marital status seemed to explain a substantial part of income inequalities in late-life mortality.

Conclusions It is unlikely that egalitarian social policies aimed at older populations can eradicate health inequalities accumulated over the life course. However, retirement income appears to have an effect on late-life mortality that is independent of the effect of income in mid-life, suggesting that egalitarian pension schemes could affect health inequalities in later life or, at the very least, slow down further accumulation of inequalities.

  • Circulatory
  • mortality
  • income
  • pensions
  • retirement
  • life course
  • Sweden
  • mortality SI
  • social differences
  • social inequalities

http://dx.doi.org/10.1136/jech.2010.131177

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    Introduction

    Money matters for health. Epidemiological studies have consistently shown an association between lower incomes and higher rates of poor health and mortality.1–5 These associations are repeatedly found across time, countries and gender. However, the magnitude of these associations varies between populations.

    Recently, as a result of the population ageing occurring throughout the developed world, interest in health inequalities in later life has increased. A growing body of research has shown that health inequalities persist into old age. Older individuals with lower socioeconomic positions run a higher risk of poor health and mortality than do those with higher socioeconomic positions.6–11 However, very little is known about the root causes of these inequalities. Insights into the mechanisms generating health inequalities in later life could serve as important stepping stones towards a better, and more equitable, health in later life.

    Assessing socioeconomic inequalities in health in old age is associated with certain caveats. Because conventional measures of socioeconomic position (education, occupation and, to some extent, income) draw upon experiences from earlier stages of the life course, it is difficult to disentangle the impact of the contemporary socioeconomic position from the impact of earlier socioeconomic conditions. This distinction is crucial when it comes to evaluating and understanding the impact of social policies designed to level out inequalities.12 If the health inequalities observed in old age are influenced by income inequalities during retirement, and not only the results of earlier experiences, then an egalitarian pension scheme would be expected to increase health equity in later life.

    In addition, research in life-course epidemiology has suggested that differential exposures throughout the life course, including childhood, may accumulate to cause socioeconomic inequalities in health.13 Employment status has been shown to be of great significance in the study of income differences in health.14 15 Two indicators of early-life circumstances that have proven to be particularly important to later circulatory mortality risk are socioeconomic conditions at birth16–19 and growth rate in utero.20–22 Finally, two predictors of adult circumstances have been shown to be of particular importance: achieved education23 24 and marital history.25–28

    We wished to disentangle the effects of earlier socioeconomic conditions and contemporary socioeconomic conditions among the elderly by exploring the association between earlier income, retirement income and health in later life. Thus, the first aim of this study was to explore the impact of retirement income on the risk of mortality in later life, net of the effect of mid-life income. Second, we wished to estimate any net effect of mid-life income on old-age mortality. Third, the study explored whether the effect of income on old-age mortality can be explained or modified by (1) early-life conditions such as growth rate in utero and social class of origin and (2) adult conditions such as achieved education and marital status. Finally, all analyses are performed separately for men and women in order to explore whether the patterning of income and mortality differs by gender.

    Materials and methods

    The Uppsala Birth Cohort Study is a life-long follow-up of the 14 193 boys and girls who were born alive at Uppsala Academic Hospital during the period 1915–1929.29 The present study is restricted to all 5771 Uppsala children who were born at Uppsala Academic Hospital during the period 1915–1924 and who had not died or emigrated by 1 January 1991 when they were in the age range 67–76 years and eligible for old-age pension. Although everyone in Sweden is entitled to old-age pension from 65 years of age, a substantial proportion of the study subjects were registered as working part time or full time in the 1990 census. Once these individuals had been removed, 1954 male and 2512 female full-time pensioners remained in the data set. Complete information on all the variables used in the mortality analyses is available for 1823 men and 2333 women, corresponding to 93% of those identified as being full-time pensioners in 1990.

    Cox's proportional hazards regression was applied in the mortality analyses.30 Subjects entered the mortality follow-up on 1 January 1991 and were censored on the year of their first emigration, upon death or at the end of follow-up (31 December 2002), whichever came first. Three types of mortality are defined according to the International Statistical Classification of Diseases and Related Health Problems (ICD): all-cause, circulatory disease (CD; ICD 9: 390–459, ICD 10: I00–I99; ie, all diseases of the circulatory system) and ischaemic heart disease (IHD; ICD 9: 410–414, ICD 10: I20–I25). Mortality differences according to the quartiles of disposable household income in 1970 and 1990 are studied. Statistics Sweden's Equivalence Scale was used to transform disposable household income into the individual's share of this amount. According to the scale, a disposable household income of 10 000 in a family of two adults and two children (15 years or younger) corresponds to a disposable household equivalent income of: 10 000/[1study subject + (0.55partner × 1) + (0.47child × 2)] ≈ 4016study subject.

    Birth weight by gestational age was used as an indicator of growth rate in utero. Individual sex-specific z-scores for birth weight for each week of gestation were calculated for all births with gestations of 30 weeks or more. These scores were then divided into tertiles. This measure has been shown to predict IHD mortality better than birth weight alone.21

    Control variables were added stepwise into new models to determine whether the initial effect of income is altered. All analyses were carried out separately for men and women with adjustments for age as a standard procedure. Table 1 presents the gender-specific distributions and RRs of the selected background variables according to all-cause, CD and IHD mortality in 1991–2002.

    View this table:
    Table 1

    Associations of the social characteristics with mortality from all-cause, circulatory disease and IHD in 1991–2002 for men and women born at Uppsala Academic Hospital in 1915–1924

    The study was approved by the regional ethics committee in Stockholm.

    Results

    The correlation between household income in 1970 and 1990 was moderate, corresponding to r=0.44 for men and women. Similarly, the patterns of income distribution, as measured by Gini-coefficients, did not differ substantially by gender. The Gini-coefficient is an indicator of income inequality that may vary from 0 (total income equality) to 1 (maximal income inequality). In 1970, the Gini-coefficient was 0.24 for men and women. In 1990, it was 0.23 for men and 0.27 for women.

    Based on table 1, it is evident that the control variables selected for our study are of varying importance for different mortality outcomes among men and women. Contrary to women, men outside the workforce have a considerably higher mortality risk from all three studied causes compared with those in paid work. As expected, the influence of growth rate in utero only manifests itself for the two circulatory mortality outcomes, with the lowest tertile distinguishing itself as the most detrimental among men and women. Social class at birth and achieved education are of greater importance for women's overall mortality risk, while for men, achieved education appears to be more crucial for the two circulatory outcomes. Moreover, when it comes to pre- and post-retirement marital status, some interesting gender differences may be noted. Thus, for all studied mortality outcomes, unmarried and separated women seem to fare much worse in relation to their married counterparts than the corresponding groups of men do, while among men, it is the divorced and widowed individuals who appear to be at the highest risk. However, among those widowed in 1990, men and women have a similar heightened risk of mortality when compared with their married peers.

    Mortality risks by quartiles of net household equivalent income during mid-life and retirement are presented in table 2 for men and in table 3 for women. Perhaps the most striking finding is that, in general, the association between income and CD mortality is substantial among women and men for mid-life as well as for retirement income. Model I reveals significant associations between income in mid-life, during retirement and risk of mortality from CD and IHD for men and women, as well as for all-cause mortality for women.

    View this table:
    Table 2

    RRs (HRs) of all-cause, circulatory disease and IHD mortality in 1991–2002 among men born in 1915–1924 (full-time pensioners in 1990) by quartiles of net household equivalent income in 1970 and net household equivalent income in 1990

    View this table:
    Table 3

    RRs (HRs) of all-cause, circulatory disease and IHD mortality in 1991–2002 among women born in 1915–1924(full-time pensioners in 1990) by quartiles of net household equivalent income in 1970 and net household equivalent income in 1990

    Model II shows that these associations appear to be robust to adjustments for early-life conditions (ie, growth rate in utero and social class at birth), although all associations are somewhat attenuated, suggesting that at least a fraction of the association between income and mortality in later life can be explained by social class of origin.

    Additional adjustment for achieved education (Model III) led to a notable attenuation of the associations between income and mortality. For women, most associations were still statistically significant, albeit attenuated. Among men, in contrast, the point estimates for IHD mortality by income class (in 1970 and 1990), as well as for all-cause mortality by income class in 1990, were rendered statistically nonsignificant when adjusting for education. However, the trends for the associations between income at both time points and IHD mortality still remain significant.

    When marital status was added in Model IV, almost all point estimates were rendered statistically insignificant for both genders. For men, significant trends remained for the associations between income in mid-life and mortality from CD and IHD. Moreover, having belonged to the lowest income quartile during mid-life was associated with an excess risk of CD mortality when compared with those from the highest income quartile. For women, on the other hand, statistically significant trends remained for the associations between income (at both time points) and all-cause mortality as well as for the association between retirement income and mortality from CD. Additionally, the second lowest income quartile in mid-life was still significantly associated with an excess risk of CD mortality in relation to the highest one.

    In Model V, household income in 1990 and 1970 are mutually adjusted. The most important finding here is that the net effect of retirement income on circulatory mortality during retirement was significant for men and women. Among women, there was also a significant association between retirement income and all-cause mortality. Moreover, among women, income during mid-life was associated with the risk of CD and all-cause mortality.

    Finally, in Model VI, all variables are included in the model. In this model, none of the associations observed in Models IV or V remain statistically significant, and the effect estimates are substantially lower, although the general patterns of health inequalities remain similar (albeit somewhat attenuated). This suggests that some of the associations between income and mortality might be confounded by earlier social factors or mediated by differences in marital status, but it is also possible that our inability to detect significant associations in the fully adjusted model is due to low statistical power. Thus, whereas it appears as if some of the effect of income, during mid-life and during retirement, on mortality in later life is mediated and/or confounded by the other variables in the model, the fully adjusted estimates suggest that there may be independent associations between income, at both time points, and late-life mortality.

    We also found a significant interaction effect between income in mid-life and retirement income on the risk of all-cause mortality among men. Separate analyses by income at mid-life (results not shown) revealed few statistically significant associations, but the patterns suggested a stronger association between retirement income and mortality among those who had higher incomes in mid-life. Among the men who were in the lowest income quartile in mid-life, there was no discernible association between retirement income and mortality risk at all.

    Finally, two significant interaction terms are illustrated in figures 1 and 2. The first shows that even though there was no overall association between mid-life income and all-cause mortality among men, a strong mortality gradient by income still existed in the small group of never married men. The second interaction shows that the association between retirement income and circulatory mortality among women was much stronger among those never married than among those ever married.

    Figure 1
    Figure 1

    RR of all-cause mortality in 1991–2002 for never and ever married men born in 1915–1924 by quartiles of net household equivalent income in 1970. (n.s.= p>0.10).

    Figure 2
    Figure 2

    RR of circulatory mortality in 1991–2002 for never and ever married women born in 1915–1924 by quartiles of net household equivalent income in 1990. (n.s.= p>0.10).

    Discussion

    The results of the present study showed that age-adjusted mortality from circulatory disease and especially IHD, during the period 1991–2002 (at age 67–87), was negatively associated with disposable income during mid-life (46–55 years) and retirement (66–75 years). These associations were found for men and women but were more pronounced among females for income during mid-life and retirement. Adjustments for early-life conditions (growth rate in utero and social class at birth) did little to help explain these associations. Differences in education and marital status, on the other hand, seem to explain a substantial part of the income inequalities in late-life mortality. In particular, we found interaction effects between income and marital status on the risk of mortality in later life. For men, there was a steep gradient in all-cause mortality risk by income during mid-life among the never married but not among those who married at some point in life. A similar pattern was found among women during retirement, where a gradient in CD mortality by income was found among the never married, but not among the ever married.

    In sum, the main findings from this study are threefold. First, there were substantial income inequalities by income class in the risk of dying, among retired men and women. Second, these inequalities could not be explained by early childhood conditions but appeared to be to a certain extent mediated by differences in achieved education and marital status. Finally, retirement income was associated with all-cause mortality among women as well as with circulatory mortality among men and women, even when differences in mid-life income were accounted for. Similar patterns, albeit somewhat attenuated, were found in the complete models (including all covariates). However, in these models, all estimates where rendered statistically insignificant.

    Nevertheless, these results should be interpreted with caution. In the present study, income does not encompass wealth and property proportionally. It includes income from capital and other assets but not the capital and assets themselves. This means that relatively low-income inequalities in old age, as caused by the pension system, could co-exist with relatively large differences in accumulated assets. Such circumstances might attenuate the associations, but as the patterns found in this study resemble those found in other studies (ie, inhabitants of lower income groups have an increased risk of mortality), it is unlikely that it substantially changes the patterns of association.

    Furthermore, one of the overarching aims of this study was to explore whether old-age pensions had an effect on old-age mortality, independent of mid-life income. In order to investigate this, those who were not retired at age 66–75 years were excluded from the study. Thus, the sample is no longer representative of the population at large, but only for those who were retired at age 66–75 years. However, exploratory analyses (not shown) were performed, including those who were not retired at that age. The results showed that, whereas the effect sizes varied when those working during late life were included, the very patterns of association remained unchanged. The only clear pattern that was discerned was that the associations between income in mid-life and late-life mortality were attenuated among the men when the full population was included in the analyses.

    Overall, women exhibited a stronger association between mid-life income and mortality risk during old age than men did. Moreover, when income during mid-life and retirement were mutually adjusted for, mid-life income turned out to be more important to mortality among women, whereas retirement income seemed to be more important for mortality among men. This could be caused by the higher rates of premature mortality among men.31 It is possible that this leads to male cohort inversion. Cohort inversion occurs when selective mortality removes those with poor health first, making cohorts that were initially disadvantaged appear compositionally advantaged and, thus, inverting previous associations.32–35 In other words, the excess mortality associated with low income in mid-life might already have removed the frailest men from the lower income groups as the cohort reaches retirement age, leaving only the effects of contemporary income to affect mortality risk. Among women, on the other hand, premature mortality is less common and, as an effect, the impact of socioeconomic conditions in mid-life might be postponed until retirement ages when female mortality rates increase. This hypothesis is supported by explorative analyses (results not shown) showing that premature mortality (ie, mortality before age 66) was, indeed, substantially more common among men than among women. In addition, among the men, premature mortality was clearly most common in the lowest income quartile and then increasingly less common in the higher income quartiles. Furthermore, the interactive effect of income at mid-life and income during retirement on the risk of all-cause mortality among men suggests that the impact of retirement income on mortality risk is dependent on income in mid-life. Among those who had a low income in mid-life, there was no income gradient in all-cause mortality in later life, suggesting that, in the lower income groups, selective mortality might have decreased the heterogeneity in susceptibility to mortality, leaving a robust cohort.

    These results have several implications and possible interpretations. The differences in income distribution in mid-life and during retirement are a crucial aspect to consider. Retirement income in Sweden stems from several sources. The main source of income among Swedes aged 65 years and older is public pensions, which account for two thirds of their aggregate income. Together with occupational and private pension schemes, it accounts for around three-quarters of their total income. Other main sources of income among older Swedes are means tested income support and income from capital. Even though the public pensions are, to a certain extent, earnings-related, the distribution is still limited by a floor and a ceiling that compress the income dispersion, leading to lower levels of income inequality among old-age pensioners36 37 than among those of working age.

    The effect of income in mid-life on mortality risk in later life could be interpreted in different ways. It could suggest that mid-life income is a powerful indicator of accumulated material working and living conditions during the adult life course, which could cause inequalities in health and mortality that persists into old age.38 In contrast, another potential interpretation draws on psychosocial theories of health inequalities.39 According to these theories, relative socioeconomic positions are critical in the production of health inequalities. A relatively deprived position is assumed to cause chronic stress and anxiety, which, in turn, cause poor health.40–42 As the income distribution is wider during mid-life than during retirement, the relative (as well as the absolute) differences are greater. Thus, if absolute and relative income inequalities drive the production of health inequalities, income in mid-life is likely to be a more severe, and potentially lasting, exposure than the less unequal retirement incomes.

    On the other hand, the net effect of retirement income may suggest that socioeconomic inequalities in old-age mortality are also caused by contemporary socioeconomic conditions and, thus, susceptible to modification by contemporary egalitarian social policies, in particular the design of pension systems.

    These results are quite remarkable; they suggest that more generous and egalitarian pension schemes may indeed result in an overall lowered mortality risk among older adults. This is compatible with a recent finding showing that generous pension systems affect old-age mortality.43 44 Furthermore, the results indicate that the income inequalities in mortality risk during later life is, to some extent, dependent on marital status. Thus, it is possible that interventions targeting social isolation could offer an effective path to the levelling of social inequalities in late-life mortality.

    In conclusion, it is unlikely that egalitarian social policies aimed at older populations could totally eradicate inequalities accumulated over the life course. However, it seems as if retirement income does have an effect on mortality independent of mid-life income, suggesting that egalitarian pension schemes could reduce health inequalities in later life or, at the very least, slow down further accumulation of inequalities.

    What is already known on this subject

    • Socioeconomic inequalities in mortality risk prevail into old age.

    • Individuals holding higher socioeconomic positions have, on average, a lower mortality risk during old age than those holding lower socioeconomic positions.

    What this study adds

    • The results from this study show that whereas some of the income inequalities in mortality during later life can be explained by income differences during mid-life, there seem to be an independent effect of income during retirement on risk of mortality even when adjusting for previous income.

    • These results suggest that income inequalities in late-life mortality may be susceptible to moderation by egalitarian pension schemes.

    Acknowledgments

    We thank Rawya Mohsen for help with data management.

    References

    1. ↵
      1. Fritzell J,
      2. Nermo M,
      3. Lundberg O
      . The impact of income: assessing the relationship between income and health in Sweden. Scand J Public Health 2004;32:6–16.
      OpenUrlAbstract/FREE Full Text
    2. ↵
      1. Geyer S,
      2. Hemstrom O,
      3. Peter R,
      4. et al
      . Education, income, and occupational class cannot be used interchangeably in social epidemiology. Empirical evidence against a common practice. J Epidemiol Community Health 2006;60:804–10.
      OpenUrlAbstract/FREE Full Text
    3. ↵
      1. Eikemo TA,
      2. Bambra C,
      3. Joyce K,
      4. et al
      . Welfare state regimes and income-related health inequalities: a comparison of 23 European countries. Eur J Public Health 2008;18:593–9.
      OpenUrlAbstract/FREE Full Text
    4. ↵
      1. Elstad JI,
      2. Dahl E,
      3. Hofoss D
      . Associations between relative income and mortality in Norway: a register-based study. Eur J Public Health 2006;16:640–4.
      OpenUrlAbstract/FREE Full Text
    5. ↵
      1. Kunst AE,
      2. Bos V,
      3. Lahelma E,
      4. et al
      . Trends in socioeconomic inequalities in self-assessed health in 10 European countries. Int J Epidemiol 2005;34:295–305.
      OpenUrlAbstract/FREE Full Text
    6. ↵
      1. Fors S,
      2. Lennartsson C,
      3. Lundberg O
      . Health inequalities among older adults in Sweden 1991–2002. Eur J Public Health 2008;18:138–43.
      OpenUrlAbstract/FREE Full Text
    7. ↵
      1. Fors S,
      2. Lennartsson C,
      3. Lundberg O
      . Childhood living conditions, socioeconomic position in adulthood, and cognition in later life: exploring the associations. J Gerontol B Psychol Sci Soc Sci 2009;64:750–7.
      OpenUrlPubMedWeb of Science
    8. ↵
      1. Huisman M,
      2. Kunst AE,
      3. Mackenbach JP
      . Socioeconomic inequalities in morbidity among the elderly; a European overview. Soc Sci Med 2003;57:861–73.
      OpenUrlCrossRefPubMedWeb of Science
    9. ↵
      1. Thorslund M,
      2. Lundberg O
      . Health and inequalities among the oldest old. J Aging Health 1994;6:51–69.
      OpenUrlAbstract/FREE Full Text
    10. ↵
      1. Olausson PO
      . Mortality among the elderly in Sweden by social class. Soc Sci Med 1991;32:437–40.
      OpenUrlCrossRefPubMedWeb of Science
    11. ↵
      1. Grundy E,
      2. Holt G
      . The socioeconomic status of older adults: how should we measure it in studies of health inequalities? J Epidemiol Community Health 2001;55:895–904.
      OpenUrlAbstract/FREE Full Text
    12. ↵
      1. Dahl E,
      2. Birkelund GE
      . Health inequalities in later life in a social democratic welfare state. Soc Sci Med 1997;44:871–81.
      OpenUrlCrossRefPubMedWeb of Science
    13. ↵
      1. Kuh D,
      2. Power C,
      3. Blane D,
      4. et al
      . Socioeconomic pathways between childhood and adult health. In: Kuh D, Ben-Shlomo Y, eds. A Life Course Approach to Chronic Disease Epidemiology. 2nd edn. Oxford: Oxford University Press, 2004:371–95.
    14. ↵
      1. Rahkonen O,
      2. Arber S,
      3. Lahelma E,
      4. et al
      . Understanding income inequalities in health among men and women in Britain and Finland. Int J Health Serv 2000;30:27–47.
      OpenUrlAbstract/FREE Full Text
    15. ↵
      1. Stronks K,
      2. van de Mheen H,
      3. van den Bos J,
      4. et al
      . The interrelationship between income, health and employment status. Int J Epidemiol 1997;26:592–600.
      OpenUrlAbstract/FREE Full Text
    16. ↵
      1. Galobardes B,
      2. Lynch JW,
      3. Davey Smith G
      . Childhood socioeconomic circumstances and cause-specific mortality in adulthood: systematic review and interpretation. Epidemiol Rev 2004;26:7–21.
      OpenUrlFREE Full Text
    17. ↵
      1. Beebe-Dimmer J,
      2. Lynch JW,
      3. Turrell G,
      4. et al
      . Childhood and adult socioeconomic conditions and 31-year mortality risk in women. Am J Epidemiol 2004;159:481–90.
      OpenUrlAbstract/FREE Full Text
    18. ↵
      1. Notkola V,
      2. Punsar S,
      3. Karvonen MJ,
      4. et al
      . Socio-economic conditions in childhood and mortality and morbidity caused by coronary heart disease in adulthood in rural Finland. Soc Sci Med 1985;21:517–23.
      OpenUrlCrossRefPubMedWeb of Science
    19. ↵
      1. Vågerö D,
      2. Leon D
      . Effect of social class in childhood and adulthood on adult mortality. Lancet 1994;343:1224–5.
      OpenUrlPubMed
    20. ↵
      1. Barker DJP
      . Mothers, Babies, and Disease in Later Life. London: BMJ Publishing Group, 1994.
    21. ↵
      1. Leon DA,
      2. Lithell HO,
      3. Vågerö D,
      4. et al
      . Reduced fetal growth rate and increased risk of death from ischaemic heart disease: cohort study of 15 000 Swedish men and women born 1915–29. BMJ 1998;317:241–5.
      OpenUrlAbstract/FREE Full Text
    22. ↵
      1. Rajaleid K,
      2. Manor O,
      3. Koupil I
      . Does the strength of the association between foetal growth rate and ischaemic heart disease mortality differ by social circumstances in early or later life? J Epidemiol Community Health 2008;62:e6.
      OpenUrlAbstract/FREE Full Text
    23. ↵
      1. Erikson R
      . Why do graduates live longer? Education, occupation, family and mortality during the 1990s. In: Jonsson JO, Mills C, eds. Cradle to Grave. Durham: Sociology press, 2001.
    24. ↵
      1. Erikson R,
      2. Torssander J
      . Clerics die, doctors survive: a note on death risks among highly educated professionals. Scand J Public Health 2009;37:227–31.
      OpenUrlAbstract/FREE Full Text
    25. ↵
      1. Koskenvuo M,
      2. Kaprio J,
      3. Kesaniemi A,
      4. et al
      . Differences in mortality from ischemic heart disease by marital status and social class. J Chronic Dis 1980;33:95–106.
      OpenUrlCrossRefPubMedWeb of Science
    26. ↵
      1. Valkonen T
      . Psychosocial stress and sociodemographic differentials in mortality from ischaemic heart disease in Finland. Acta Med Scand Suppl 1982;660:152–64.
      OpenUrlPubMed
    27. ↵
      1. Hu YR,
      2. Goldman N
      . Mortality differentials by marital status: an international comparison. Demography 1990;27:233–50.
      OpenUrlCrossRefPubMedWeb of Science
    28. ↵
      1. Modin B
      . Born out of wedlock and never married—it breaks a man's heart. Soc Sci Med 2003;57:487–501.
      OpenUrlCrossRefPubMedWeb of Science
    29. ↵
      1. Koupil I
      . The Uppsala studies on developmental origins of health and disease. J Intern Med 2007;261:426–36.
      OpenUrlCrossRefPubMedWeb of Science
    30. ↵
      1. Clayton D,
      2. Hills M
      . Statistical Models in Epidemiology. Oxford: Oxford University Press, 1996.
    31. ↵
      1. Case A,
      2. Paxson C
      . Sex differences in morbidity and mortality. Demography 2005;42:189–214.
      OpenUrlCrossRefPubMedWeb of Science
    32. ↵
      1. Lynch SM
      . Cohort and life-course patterns in the relationship between education and health: a hierarchical approach. Demography 2003;40:309–31.
      OpenUrlCrossRefPubMedWeb of Science
    33. ↵
      1. Noymer A
      . Mortality selection and sample selection: a comment on Beckett. J Health Soc Behav 2001;42:326–7.
      OpenUrlCrossRefPubMedWeb of Science
    34. ↵
      1. Ferraro KF,
      2. Shippee TP
      . Aging and cumulative inequality: how does inequality get under the skin? Gerontologist 2009;49:333–43.
      OpenUrlAbstract/FREE Full Text
    35. ↵
      1. Hobcraft J,
      2. Menken J,
      3. Preston S
      . Age, period, and cohort effects in demography: a review. Popul Index 1982;48:4–43.
      OpenUrlCrossRefPubMedWeb of Science
    36. ↵
      Statistics Sweden. De äldres ekonomiska välfärd. [The Economic Welfare of Older People]. Stockholm: Statistics Sweden, 2004.
    37. ↵
      Ministry of Health and Social Affairs. Gyllene år med silverhår – för visa eller för alla?. [Golden Years with Silver Hair – For Some or For Everyone?]. Stockholm: Ministry of Health and Social Affairs, 2002.
    38. ↵
      1. Lynch JW,
      2. Smith GD,
      3. Kaplan GA,
      4. et al
      . Income inequality and mortality: importance to health of individual income, psychosocial environment, or material conditions. BMJ 2000;320:1200–4.
      OpenUrlFREE Full Text
    39. ↵
      1. Marmot M,
      2. Wilkinson RG
      . Psychosocial and material pathways in the relation between income and health: a response to Lynch et al. BMJ 2001;322:1233–6.
      OpenUrlFREE Full Text
    40. ↵
      1. Marmot M
      . The Status Syndrome. New York: Times Books, 2004.
    41. ↵
      1. Wilkinson RG
      . The Impact of Inequality. New York: The New Press, 2005.
    42. ↵
      1. Sapolsky RM
      . The influence of social hierarchy on primate health. Science 2005;308:648–52.
      OpenUrlAbstract/FREE Full Text
    43. ↵
      1. Lundberg O,
      2. Yngwe MA,
      3. Stjarne MK,
      4. et al
      . The role of welfare state principles and generosity in social policy programmes for public health: an international comparative study. Lancet 2008;372:1633–40.
      OpenUrlCrossRefPubMedWeb of Science
    44. ↵
      1. Lundberg O,
      2. Åberg Yngwe M,
      3. Kölegård Stjärne M,
      4. et al
      . The Nordic Experience: Welfare States and Public Health (NEWS). Stockholm: Centre for health equity studies, 2008.

    Footnotes

    • Funding The UBCoS Multigen study is supported by grants from the Swedish Council for Working Life and Social Research (grant number 2007-1010) and the Swedish Research Council (2006-7498). SF gratefully acknowledges financial support from the Swedish Council for Working Life and Social Research (grant 2005-0624 and 2007-2012).

    • Competing interests None.

    • Ethics approval This study was conducted with the approval of the regional ethics committee in Stockholm.

    • Provenance and peer review Not commissioned; externally peer reviewed.