You are here

Article Text

Article menu

Download PDFPDF

Original research
Widening income inequalities in life expectancy? Analysing time trends based on German health insurance data
  1. Fabian Tetzlaff1,
  2. Jelena Epping2,
  3. Stefanie Sperlich2,
  4. Juliane Tetzlaff2
  1. 1Institute for General Practice, Hannover Medical School, Hanover, Germany
  2. 2Medical Sociology Unit, Hannover Medical School, Hanover, Germany
  1. Correspondence to Fabian Tetzlaff, Institute for General Practice, Hannover Medical School, Hanover, 30625, Germany; tetzlaff.fabian{at}mh-hannover.de

Abstract

Background While social disparities in mortality have been shown in the majority of high-income countries, research on inequalities in the German population is still limited. This applies especially to studies investigating time trends in social inequalities with respect to life expectancy. The aim of this study is to examine income inequalities in life expectancy and whether inequalities have narrowed or widened over time.

Methods The analyses are based on the claims data of a large German health insurance provider, which facilitates the combining of information on individual income and mortality. Life expectancy is calculated separately for three income groups (<60%, 60% to 80% and ≥80% of the average income in Germany) and for sex by applying period life table analyses. Trends are assessed by comparing the time periods 2005–2008 (N = 1 773 122), 2009–2012 (N=1 792 735) and 2013–2016 (N = 1 987 114).

Results Trends in life expectancy differed by sex, age and income group. Especially among elderly men, the gap between low- and high-income groups widened over time, disadvantaging men with low incomes. Among women, a slight reduction in inequalities was observed, which was driven by the increases in life expectancy in low-income groups.

Conclusion Our study shows that not all population subgroups benefitted equally from the continuing rise in life expectancy. The persisting inequalities emphasise the importance of public health efforts concentrating on reducing mortality risks among individuals in lower socioeconomic positions. Special attention should be paid to elderly men with low incomes. Further research is needed on the mechanisms underlying increasing health inequalities over time.

  • life expectancy
  • time trend
  • income inequalities
  • Germany

http://dx.doi.org/10.1136/jech-2019-212966

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

    With a life expectancy of 78.4 years for men and 83.2 years for women,1 Germany is among the top 30 low-mortality countries worldwide.2 Nevertheless, health inequality rooted in socioeconomic disparities represents a substantial burden on the German healthcare system and society as a whole.3 In particular, the cumulative harmful effects of social deprivation and health-related behaviour during the life course determine the outcome of social inequalities in health and mortality.4 Social disparities in mortality and longevity have become evident in the vast majority of industrialised countries.5–13 Compared with other countries, studies on social inequalities in life expectancy in Germany are still rare and the evidence is mixed.14 This holds especially true for studies investigating the development of social inequalities in life expectancy over time.

    The wide body of international research proves that social inequalities in mortality have been a frequently studied issue. In contrast, there is a limited number of studies in Germany investigating this issue due to the fact that the official population statistics do not provide information on mortality by socioeconomic status (SES). Therefore, studies analysing social inequalities in life expectancy are most frequently based on survey data with mortality follow-up14–16 or on data from the German Federal Pension Fund.17 18 Previous studies concur that social gradients are also present in the German population, in terms of occupational groups17 or educational16 or income levels.3 15 17 18 However, due to data limitations, most studies on social inequalities in life expectancy are not directly comparable with international studies. This holds not only for the data source used but also for the methods used to calculate life expectancies. With regard to the restricted case numbers in survey data, life expectancy by SES is usually obtained by using estimations other than classical period life table approaches.14 This includes approaches that combine relative mortality risks between SES groups estimated from survey data with mortality information from official population statistics.3 15 Furthermore, the studies only report life expectancy for middle-aged3 or older individuals,14 15 as data often do not cover the full age range.

    So far, there is little evidence on how trends in socioeconomic inequalities in life expectancy vary by age group. As in other high-income countries, a slowdown in life expectancy increases can also be observed in Germany. The reasons for this slowdown differ by country according to recent trends in causes of death.19 Moreover, international research shows that these trends also vary by age, leading to differing trends in mortality and life expectancy among the younger and older population.19 Recent studies indicate that trends in health also differ according to age. While self-rated health (SRH) has improved among the older German population, younger age groups have tended to show stagnating trends in good SRH.20 Similar results for SRH and functional health have also been reported for other European countries.21–23

    For Germany, previous research on time trends in inequalities in life expectancy showed mixed evidence.15–18 24 Considerable inequalities were reported in studies that combine mortality risks by income estimated from survey data with data from official population statistics.15 16 In the period 1992 to 2016, life expectancy at birth was reported to differ between the highest and the lowest income groups by 8.6 years in males and by 3.7 years in females.15 The results also indicate that inequalities have remained fairly stable over time.15 16

    Other studies examining social inequalities in life expectancy are based on data from the German Federal Pension Fund and focus on trends in the remaining life expectancy of men aged 65 years.17 18 In these studies, income inequalities were assessed considering lifetime earnings registered for calculations of pension payments by the German Federal Pension Fund.17 18 For the period 1995–1996, a gap of 3.2 years between pensioners in the lowest and the highest income group was reported. In contrast to the studies mentioned above, this gap increased over time, amounting to 4.9 years in 2007–2008. Smaller but also widening inequalities were found in terms of occupational groups.17 Current research suggests that these inequalities in life expectancy have continued to increase until the recent past.18 The analyses reveal that especially younger cohorts of East German men with low income were disadvantaged compared to men with higher income, who benefitted most from mortality reductions over time.18 On the other hand, it has been shown that regional disparities in life expectancy in Germany have decreased over time, despite persisting large economic inequalities.24

    The limited number of studies and the mixed evidence stress the importance of further research on trends in life expectancy inequalities in the German population. The aim of our study was to investigate these trends based on health insurance claims data. The major advantage of using health insurance data is that these datasets contain complete records of all deaths in the insurance population. Furthermore, the datasets include information on individual income, which facilitates the calculation of life expectancy by applying classic period life table analyses based on the observed deaths by individual income. The study is guided by the following research questions:

    • Are there differences in life expectancy between income groups?

    • Are there differing time trends in life expectancy between income groups that lead to widening or narrowing inequalities over time?

    METHODS

    Data

    In Germany, it is mandatory for all inhabitants to be insured either by a statutory health insurance or by a private health insurance company. As statutory health insurance coverage is part of the welfare-state system, approximately 90% of the German population is insured by a statutory health insurance provider.25

    For this study, we used the claims data of a large statutory health insurance provider located in the federal state Lower Saxony (AOK Niedersachsen (AOKN)), which insures more than one-third of the inhabitants of the state.26 The data cover the years 2005 to 2016. Single years of observation were combined into time periods, as case numbers of deaths within income groups are limited. For this study, the data that containall insured individuals aged 20 and older for the time periods 2005–2008, 2009–2012 and 2013–2016 were analysed. The data include sociodemographic information (eg, age, sex, date of death and income), outpatient and inpatient diagnoses, treatments and medications of all individuals insured by the AOKN. An earlier study has shown that the AOKN population is representative of the total German population in terms of sex and age distributions, while individuals with lower incomes and lower occupational status are overrepresented.27

    Definition of income groups

    As insurance fees depend on income, the dataset contains detailed information on individual income. This information includes the annual pre-tax income from salaries and pension payments. Income groups were defined in relation to the average income in Germany, which is annually reported as part of the official statistics.28 Because of increases in general income levels, the average income varied over time. Over the observation period, the increases in the average income clearly exceeded the inflation rate in Germany.28 29 Thus, a person with an average income during the third period (2013–2016) could afford a much higher standard of living than a person with an average income during the first period (2005–2008). To increase comparability over time, we adjusted the individual income reported in the data for the annual inflation rate. In a second step, we classified individuals relative to the average income level of the first observation period.

    After these adjustments, individuals were assigned to three groups. The low-income group comprises a pre-tax annual income of <60% of the average income, the high-income group includes individuals with ≥80% of the average income and the middle-income group was defined as falling in between these categories (table 1).

    View this table:
    Table 1

    Characteristics of the study population: number of insured individuals, exposures in person-years and number of deaths by time period and sex

    Since we used annual income information, individuals who were not insured over a full-year period were excluded from analyses. An exception was made by including individuals who died in the respective year. With a proportion of 12% to 15%, only a minor share of the insurance population could not be included in the analyses (Table A1, supplement). During the three periods, a total of 5 552 971 individuals, 375 506 deaths and 15 504 553 person-years of exposure were included in the analyses, with the highest case numbers in the low-income group (table 1).

    Supplemental material

    Statistical analyses

    To analyse the development of income inequalities in life expectancy over time, we calculated age-specific mortality rates by income group, sex and time period. Using these rates as input, period life tables were calculated.30 31 To determine whether differences between income groups and over time were significant, 95% CIs of life expectancy were estimated by drawing 1000 bootstrap samples. To depict trends in life expectancy over time, we calculated absolute differences between periods within income groups as well as between income groups. As the general level in life expectancy differed substantially between income groups, relative differences were also reported. All analyses were stratified for sex and income group and were performed in Stata 1432 and R Version 3.5.1.33

    Sensitivity analyses

    To test the robustness of the reported time trends, sensitivity analyses were performed using different types of redistribution strategies for individuals with missing information on income: (a) redistribution according to the age-specific proportion of the low-, middle- and high-income group among the study population with income information, (b) redistribution to the low-income group and (c) redistribution to the high-income group.

    RESULTS

    Figure 1 displays life expectancies for men and women in the three periods. Comparing our calculations with the total German life expectancy, the life expectancy of the low-income group is below that of the high-income group above the total German life expectancy. This holds for both sexes and all periods (Tables A2 and A3, supplement). In both sexes, we found significant increases in life expectancy over time (figure 1). These increases exist in all income groups except for women in the high-income group, for whom no significant changes occurred. The strongest gains in life expectancy over time were observed for men in the middle-income group. In this group, the change in the expected number of remaining life years varies between 0.7 years at age 65 and 2.5 years at age 20. With increases of 0.6 years at age 20 and 0.3 years at age 65, the highest gains for women emerged in the low-income group (figure 1).

    Figure 1
    Figure 1

    Life expectancy by income group and period as well as changes in life expectancy between periods within income groups by age and sex.

    Note: Absolute change refers to the difference in life expectancy between the third and the first time period in life years. Relative change refers to the difference in life expectancy between the third and the first time period divided by life expectancy at the first time period in percentage points. *Significant difference in life expectancy between periods at the 5% level.

    Absolute and relative differences in life expectancy between income groups indicate whether income inequalities have widened or narrowed over time (table 2). Overall, the trend observed in the first and second periods continued in the third period. Among younger and middle-aged men, inequalities between the high- and low-income groups remained quite constant. This does not hold for men of age 65 years, for whom a substantial increase in income disparities was observed (+1.0 life years, +6 percentage points). However, driven by the strong increase in life expectancy among younger and middle-aged men belonging to the middle-income group, the difference in the high-income group narrowed over time, whereas the gap in the low-income group increased. Among women, the gap in life expectancy between the low- and high-income groups narrowed over time, irrespective of whether age 20, 40 or 65 was considered. In each of these age groups, the differences were reduced by approximately 1 to 2 percentage points, which equals a reduction ranging from 0.9 years at age 20 and 0.2 years at age 65 (table 2). While differences between the middle- and low-income groups remained constant over time, decreases were observed between the high- and middle-income groups. Due to the clear increases in the low- and middle-income groups, women benefitted from decreasing inequalities in life expectancy (table 2).

    View this table:
    Table 2

    Absolute and relative income inequalities in life expectancy between income groups by period, age and sex

    DISCUSSION

    The aim of this study was to investigate the development of income inequalities in life expectancy over time. The analyses show that life expectancy between 2005–2008 and 2013–2016 increased in both sexes. This holds true for men, irrespective of income group, and for women belonging to the low- and middle-income groups. However, not all individuals benefitted equally from this trend. In particular, elderly men with low income are disadvantaged, as the gap between the high- and low-income groups widened over time, exacerbating the existing inequalities in life expectancy at age 65. This trend did not occur in men at ages 20 and 40. Similar results were not found for women, where a slight reduction in inequalities in life expectancy was observed in all age groups. This reduction is driven by the clear increases in life expectancy among women in the low- and the middle-income groups.

    This paper is one of the rare studies on the time trends of social inequalities in life expectancy in the German population. Previous research on this topic shows mixed evidence.14 15 17 18 While other studies demonstrate substantial social inequalities in life expectancy as well, the reported trends indicate either constant15 16 or increasing disparities17 18 over time. Similar to Kibele et al and Wenau et al, the present analyses are based on individual income information. Our results are in line with these studies, which also reported widening inequalities among elderly men over time based on pension fund data.17 18 In accordance with previous research,5–13 15 17 18 34 our analyses reveal higher social disparities in life expectancy for men than for women. This finding might be due to greater differences in occupational burden between SES groups among men than among women.14 35 36 Overall, the increase in life expectancy for women is much weaker than for men, leading to a narrowing gender gap, which can also be observed for the total German population.1 Due to the considerably small increase in female life expectancy, changes in inequalities over time were expected to be less pronounced for women than for men, which holds true for our study population.

    Strengths and limitations

    Our study is based on a large number of cases, which allowed us to investigate time trends among different subgroups of the study population by stratifying for observation periods. The data represent a complete health insurance population that includes all insured individuals regardless of their current health status. Thus, the analyses are unaffected by health-related nonresponse, which occurs in survey data if individuals refrain from study participation for health reasons.37

    Previous research on the development of income inequalities in life expectancy among the German population is either restricted to trends in the remaining life-years at older ages among men17 18 or based on survey data containing limited case numbers of deceased individuals.3 14–16 An advantage of using health insurance data derives from the completeness of information on mortality. As death terminates health insurance, the date of death of all deceased individuals can be identified precisely. Furthermore, the dataset includes detailed information on annual individual income. These data characteristics allowed us to estimate the life expectancy of different income groups by applying classical period life table analyses based on a single dataset rather than using estimation strategies that combine information from different datasets.15 16 Furthermore, life expectancy can be calculated for both sexes and at various ages, as the data include information on mortality and individual income from age 20 up to the oldest old.

    Considering sex and age distributions, the data are comparable with the total population of Germany and Lower Saxony but differ in terms of the social structure.27 Whereas the average income in Lower Saxony differs only slightly from that in Germany,38 high-income earners are underrepresented in our data.27 However, the general trend in income inequalities is comparable with that of the German population,39 tending towards increasing inequalities. Due to the low-income level, the total life expectancy of the insurance population is lower than that of the total German population (Tables A2 and A3, supplement). This is taken into account, as all calculations are stratified for income. Furthermore, comparisons show that the life expectancy of the low-income group is below and that of the high-income group above the total German life expectancy. This holds for each of the three periods and for both sexes (Tables A2 and A3, supplement).

    Since the data contain only information on the annual income for individuals who are insured over a full year, individuals with shorter insurance periods were excluded from the life table analyses. Shorter insurance periods mainly result from changing the health insurance provider or a change in residence to another federal state, which is more frequent in younger age groups (Table A4, supplement). The question of whether the SES characteristics of the excluded individuals differ from those included in the analyses cannot be answered straightforwardly, because information on annual incomes in this subgroup is missing. However, as the proportion of individuals who were not insured over a full-year period is small and remained stable over time (Table A1, supplement), we expected the reported time trends in life expectancy to be robust. We tested this assumption by applying different redistribution strategies for individuals with missing information on income. Furthermore, the effect of excluding individuals without income information was assessed by comparing the total life expectancies of the insurance population with and without this exclusion. While total life expectancy in these additional analyses tended to be higher, the general time trend in life expectancy remained nearly unchanged (Table A5, supplement). The same applies with regard to time trends of the income-specific life expectancy, irrespective of the strategy used for redistribution (Table A6, supplement).

    As our dataset does not facilitate the matching of the income information of spouses, and since no information on household composition is available, household income could not be used in this study. This may have led to an underestimation of financial resources, especially among women, as their general income level is lower than that among men.28 Therefore, the results for women should be interpreted with some caution. However, research on the effects of income type on health shows that individual income is an appropriate measure to analyse SES disparities in health, although social gradients are more pronounced if household income is applied.40

    Among men, our findings point towards a growing effect of income inequalities on mortality with increasing age. This could partly be explained by the cumulative effect of social deprivation over the life course. From a short-term perspective, income determines the quality of the current living conditions and health-related behaviour. Furthermore, belonging to a high-income group increases the chances of accumulating property and wealth over the life course,4 which may contribute to the higher effect of income among older men.

    Since case numbers within income groups are limited, the reported trends are based on a comparison of three time periods. Analyses based on single years may provide more differentiated information on short-term developments in inequalities in life expectancy, but would affect the robustness of the results. To gain deeper insight into these developments, the analyses are not limited to the highest and lowest income earners but also include the group in between. Focusing only on the highest and the lowest income groups would have provided an incomplete picture, as a considerable share of the population cannot be assigned to these extremes. Since the case numbers within the income groups are limited, no more than three different groups could be defined and no differentiation above ≥80% of the average income could be made.

    To increase comparability over time, income has been adjusted for inflation, and individuals are assigned to groups relative to the average income in the German population in the first observation period. Without applying these adjustments, gains in life expectancy in the low- and middle-income groups would have been somewhat higher, as income heterogeneity within these groups would have increased over time. This holds especially for the lowest income group. However, as the differences between these two income definitions are limited, the trends in income inequalities in life expectancy are similar. Furthermore, individuals were assigned to income groups according to their relative position to the German average income. An assignment based on the income distribution within our dataset (eg, income quintiles) would have reduced the comparability with the total German population, as low-income earners are overrepresented in the data. This was prevented through the use of the external criterion of the average German income as a reference for the assignment to the income groups.

    CONCLUSION

    The study shows persisting or even widening income inequalities in life expectancy over time. While life expectancy increased for both sexes, not all individuals benefitted equally from this trend. Our analyses indicate that time trends differ not only with regard to income groups but also according to age groups. Thus, further research should not be limited to studying time trends in life expectancy at birth or retirement age but should include a broader age range to account for these differences. Moreover, changes over time occurred not only in differences between the high- and the low-income group but also in comparisons between the middle- and the high-income group or between the middle- and the low-income group. More research is needed to reveal the underlying mechanisms causing widening inequalities in mortality. As gains in life expectancy are not only caused by changes in morbidity but also by increasing survival after disease onset, studies aiming to investigate these mechanisms should not be limited to social inequalities in all-cause mortality. Health insurance data are an appropriate data source for such analyses, as they allow for the combination of individual information on specific diseases, mortality and socioeconomic position.

    What is already known on this subject

    • Social inequalities in mortality have been shown in high-income countries. Previous research indicates that social inequalities in mortality are also present in the German population. As official population statistics do not provide mortality information by SES, studies on the time trends of social disparities in life expectancy in Germany are still rare and the evidence is mixed.

    What this study adds

    • The aim of the study is to investigate time trends in social inequalities over time based on German health insurance data. As the data contain individual information on income as well as on mortality, the classical period life table approach based on a single dataset can be applied. Our results show that not all population subgroups benefitted equally in terms of life expectancy gains. In particular, older men with low incomes are disadvantaged, leading to increasing inequalities over time. The findings emphasise the importance of public health efforts aimed at reducing health inequalities.

    Acknowledgments

    We thank the AOKN (Statutory Local Health Insurance of Lower Saxony) for providing the data. In particular, the support of Dr Jürgen Peter, Dr Jona Stahmeyer and Dr Sveja Eberhard made it possible to carry out this study.

    REFERENCES

    1. ↵
      1. Statistisches Bundesamt
      . Sterbetafel 2015/2017 - Methoden- und Ergebnisbericht zur laufenden Berechnung von Periodensterbetafeln für Deutschland und die Bundesländer. Wiesbaden, 2018. Available https://www.destatis.de/DE/Themen/Gesellschaft-Umwelt/Bevoelkerung/Sterbefaelle-Lebenserwartung/_inhalt.html
    2. ↵
      1. World Health Organization
      . World Health Statistics 2016: monitoring health for the SDGs sustainable development goals. Geneva: World Health Organization, 2016.
    3. ↵
      1. Lampert T,
      2. Kroll LE,
      3. Kuntz B, et al.
      Health inequalities in Germany and in international comparison: trends and developments over time. J Health Monit 2018;3(S1):1–24.
      OpenUrl
    4. ↵
      1. Lampert T,
      2. Hoebel J
      . Socioeconomic inequalities in health in later life. Zeitschrift für Gerontologie und Geriatrie 2019;52:91–9. doi: 10.1007/s00391-018-01487-y
      OpenUrl
    5. ↵
      1. Valkonen T
      . Social inequalities in mortality. In: Jacques CGV, Wunsch G, eds. Demography: analysis and synthesis A treatise in population studies. Amsterdam: Elsevier, 2006: 195–206.
    6. ↵
      1. Corsini V
      . Highly educated men and women likely to live longer. Life expectancy by educational attainment. Statistics in focus. 2010: 24/2010. Brussels: Statistical Office of the European Communities (Eurostat).
    7. ↵
      1. Caselli G DS
      , Wegner-Siegmundt C, Luy M. Future mortality in low mortality countries. In: Lutz W BW KC, ed. World population and human capital in the 21st century. Oxford: Oxford University Press. 2014: 226–72.
    8. ↵
      1. van Raalte AA,
      2. Kunst AE,
      3. Deboosere P, et al.
      More variation in lifespan in lower educated groups: evidence from 10 European countries. Int J Epidemiol 2011;40:1703–14. doi: 10.1093/ije/dyr146
      OpenUrlCrossRefPubMedWeb of Science
    9. ↵
      1. Firebaugh G,
      2. Acciai F,
      3. Noah AJ, et al.
      Why lifespans are more variable among blacks than among whites in the United States. Demography 2014;51:2025–45. doi: 10.1007/s13524-014-0345-2
      OpenUrlCrossRefPubMed
    10. ↵
      1. Aburto JM,
      2. van Raalte A
      . Lifespan dispersion in times of life expectancy fluctuation: the case of Central and Eastern Europe. Demography 2018;55:2071–96. doi: 10.1007/s13524-018-0729-9
      OpenUrl
    11. ↵
      1. Aburto JM,
      2. Wensink M,
      3. van Raalte A, et al.
      Potential gains in life expectancy by reducing inequality of lifespans in Denmark: an international comparison and cause-of-deathanalysis. BMC Public Health 2018;18:831. doi: 10.1186/s12889-018-5730-0
      OpenUrl
    12. ↵
      1. Bronnum-Hansen H
      . Socially disparate trends in lifespan variation: a trend study on income and mortality based on nationwide Danish register data. BMJ Open 2017;7:e014489. doi: 10.1136/bmjopen-2016-014489
    13. ↵
      1. Chetty R,
      2. Stepner M,
      3. Abraham S, et al.
      The association between income and life expectancy in the United States, 2001-2014. JAMA 2016;315:1750–66. doi: 10.1001/jama.2016.4226
      OpenUrlCrossRefPubMed
    14. ↵
      1. Luy M,
      2. Wegner-Siegmundt C,
      3. Wiedemann A, et al.
      Life expectancy by education, income and occupation in Germany: estimations using the longitudinal survival method. Comp Popul Stud 2015;40:339–436.
      OpenUrl
    15. ↵
      1. Lampert T,
      2. Hoebel J,
      3. Kroll LE
      . Social differences in mortality and life expectancy in Germany. Current situation and trends. J Health Monit 2019;4:3–14.
      OpenUrl
    16. ↵
      1. Unger R,
      2. Schulze A
      . Can we really (all) work longer? Trends in healthy life expectancy according to social stratum in Germany. Comp Popul Stud 2013;38:545–64.
      OpenUrl
    17. ↵
      1. Kibele EU,
      2. Jasilionis D,
      3. Shkolnikov VM
      . Widening socioeconomic differences in mortality among men aged 65 years and older in Germany. J Epidemiol Community Health 2013;67:453–7. doi: 10.1136/jech-2012-201761
      OpenUrlAbstract/FREE Full Text
    18. ↵
      1. Wenau G,
      2. Grigoriev P,
      3. Shkolnikov V
      . Socioeconomic disparities in life expectancy gains among retired German men, 1997–2016. J Epidemiol Community Health 2019;73:605–11. doi: 10.1136/jech-2018-211742
      OpenUrlAbstract/FREE Full Text
    19. ↵
      1. Ho JY,
      2. Hendi AS
      . Recent trends in life expectancy across high income countries: retrospective observational study. BMJ (Clinical Research Ed) 2018;362:k2562. doi: 10.1136/bmj.k2562
    20. ↵
      1. Sperlich S,
      2. Tetzlaff J,
      3. Geyer S
      . Trends in good self-rated health in Germany between 1995 and 2014: do age and gender matter? Int J Public Health 2019;64:921–33. doi: 10.1007/s00038-019-01235-y
      OpenUrl
    21. ↵
      1. Wolff JKNS
      . Spuling SM Altern nachfolgende Kohorten gesünder? Selbstberichtete Erkrankungen und funktionale Gesundheit im Kohortenvergleich. In: Mahne KW, JK SJ, Tesch-Römer C, eds. Altern im Wandel Zwei Jahrzehnte Deutscher Alterssurvey (DEAS). Wiesbaden: Springer, 2017: 125–38.
      1. Clause-Verdreau AC,
      2. Audureau E,
      3. Leplege A, et al.
      Contrasted trends in health-related quality of life across gender, age categories and work status in France, 1995-2016: repeated population-based cross-sectional surveys using the SF-36. J Epidemiol Community Health 2019;73:65–72. doi: 10.1136/jech-2018-210941
      OpenUrlAbstract/FREE Full Text
    23. ↵
      1. Johansson SE,
      2. Midlov P,
      3. Sundquist J, et al.
      Longitudinal trends in good self-rated health: effects of age and birth cohort in a 25-year follow-up study in Sweden. Int J Public Health 2015;60:363–73. doi: 10.1007/s00038-015-0658-y
      OpenUrl
    24. ↵
      1. van Raalte AA,
      2. Klusener S,
      3. Oksuzyan A, et al.
      Declining regional disparities in mortality in the context of persisting large inequalities in economic conditions: the case of Germany. Int J Epidemiol 2020 [Epub ahead of print]. doi: 10.1093/ije/dyz265
    25. ↵
    26. ↵
      1. Allgemeine Ortskrankenkasse Niedersachsen
      . Geschäftsbericht 2017. Remshalden-Grunbach: Wacker Offsetdruck GmbH, 2017.
    27. ↵
      1. Jaunzeme J,
      2. Eberhard S,
      3. Geyer S
      . How “representative” are SHI (statutory health insurance) data? Demographic and social differences and similarities between an SHI-insured population, the population of Lower Saxony, and that of the Federal Republic of Germany using the example of the AOK in Lower Saxony. Bundesgesundheitsblatt Gesundheitsforschung Gesundheitsschutz 2013;56:447–54.
      OpenUrl
    28. ↵
      1. Statistisches Bundesamt
      . Arbeitnehmerverdienste - Lange Reihen. Fachserie 16 Reihe 2.4. Wiesbaden, 2019. Available https://www.destatis.de/DE/Themen/Arbeit/Verdienste/Verdienste-Verdienstunterschiede/_inhalt.html#sprg233158
    29. ↵
      1. Statistisches Bundesamt
      . Verbraucherpreisindizes für Deutschland. Jahresbericht 2017. Wiesbaden, 2018. Available https://www.destatis.de/DE/Themen/Wirtschaft/Preise/Verbraucherpreisindex/_inhalt.html#sprg238948
    30. ↵
      1. Chiang CL
      . The life table and its applications. Malabar, FL, USA: Robert E. Krieger Publishing Company, 1984.
    31. ↵
      1. Preston SH,
      2. Heuveline P,
      3. Guillot M
      . Demography. Measuring and modeling population processes. Oxford: Blackwell Publishing, 2001.
    32. ↵
      1. Stata_Corp
      . Stata statistical software. release 14. College Station: Stata Corp, 2015.
    33. ↵
      1. R Core Team. R
      . A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing, 2015.
    34. ↵
      1. Johnson B
      . Deriving trends in life expectancy by the National Statistics Socioeconomic Classification using the ONS Longitudinal Study. Health Stat Q 2011;49:9–51. doi: 10.1057/hsq.2011.2
      OpenUrl
    35. ↵
      1. Bertuccio P,
      2. Alicandro G,
      3. Sebastiani G, et al.
      Mortality by occupation-based social class in Italy from 2012 to 2014. Int J Public Health 2018;63:865–74. doi: 10.1007/s00038-018-1149-8
      OpenUrlPubMed
    36. ↵
      1. Lallo C,
      2. Raitano M
      . Life expectancy inequalities in the elderly by socioeconomic status: evidence from Italy. Popul Health Metr 2018;16:7. doi: 10.1186/s12963-018-0163-7
      OpenUrl
    37. ↵
      1. Geyer S
      , Jaunzeme J. Möglichkeiten und Grenzen von Befragungsdaten und Daten gesetzlicher Krankenversicherungen. Swart E, Ihle P, Gothe H, et al. eds. Routinedaten im Gesundheitswesen: Handbuch Sekundärdatenanalyse: Grundlagen, Methoden und Perspektiven. Bern: Verlag Hans Huber, 2014: 223–33.
    38. ↵
      1. Grabka MM,
      2. Goebel J,
      3. Liebig S
      . Wiederanstieg der Einkommensungleichheit: Aber auch deutlich steigende realeinkommen. DIW-Wochenbericht 2019;86:343–53.
      OpenUrl
    40. ↵
      1. Geyer S
      . Income, income, or income? The effects of different income measures on health in a national survey. J Epidemiol Community Health 2011;65:491–6. doi: 10.1136/jech.2009.103390
      OpenUrlAbstract/FREE Full Text

    Footnotes

    • Contributors FT and JT developed the idea and research questions of the study. FT analysed the data and wrote the first draft of the manuscript. JE and JT were major contributors to the final manuscript. JE, SSp and JT contributed to the conception and discussion of the study and reviewed the work critically. All authors read and approved the final version of the manuscript.

    • Funding The work done by JT was funded by the AOKN (Statutory Local Health Insurance of Lower Saxony) as part of a project on morbidity compression.

    • Competing interests None declared.

    • Patient consent for publication Not required.

    • Ethics approval Our study is based on claims data, that is, on routinely collected data. The use of this sort of data for scientific purposes is regulated by federal law, and the data protection officer of the Statutory Local Health Insurance of Lower Saxony (AOKN) has approved its use.

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

    • Data availability statement The datasets generated and analysed during the current study are not publicly available due to protection of data privacy of the insured individuals by the Statutory Local Health Insurance of Lower Saxony (AOKN).