Socioeconomic status and health among the aged in the United States and Germany: A comparative cross-sectional study
Introduction
Research in social epidemiology and medical sociology has consistently shown socioeconomic status (SES) differentials in morbidity and mortality in general populations. Those with lower SES generally experience higher morbidity and mortality rates than those with higher SES (Mackenbach, Kunst, Cavelaars, Groenhof, & Guerts, 1997; Marmot & Wilkinson, 1999; Mielck, 2000; Pappas, Queen, Hadden, & Fisher, 1993). Investigations into the determinants of social inequalities in health have documented that the following explanatory frameworks are of major importance: first, an increased risk of socially patterned unhealthy lifestyles in lower SES groups (Lantz et al., 2001; Lynch, Kaplan, Cohen, Tuomilehto, & Salonen, 1996); secondly, unequal access to and quality of health care received (Feinstein, 1993; Mackenbach, Stronks, & Kunst, 1989), and thirdly, differential exposure to material deprivation and a stressful psychosocial environment over the life course, with particular emphasis on early life (Kuh & Ben-Shlomo, 1997; Lynch et al., 1994) and on midlife (work, family, social exclusion; Blane, Brunner, & Wilkinson, 1996; Marmot, Theorell, & Siegrist, 2002; Shaw, Dorling, & Davey Smith, 1999).
Despite this progress, research on the determinants of social inequalities in health among elderly populations has been less frequent. In this respect, several studies have confirmed the hypothesis that SES differences in morbidity and mortality diminish with age, in particular as a result of selective survival into old age (House et al., 1994; Manor, Eisenbach, Peritz, & Frielander, 1999; Stolpe, 1997). However, contrary to this evidence, a number of investigations have documented a continuation of the social gradient of morbidity and mortality into old age (Berkman & Gurland, 1998; Lampert, 2000; Liao, McGee, Kaufman Cao, & Cooper, 1999; Marmot & Shipley, 1996; Parker, Ahacic, Thorslund, & Lundberg, 1999). It may be that part of the variation in these results is due to different indicators used to define socioeconomic position (Robert & House, 1996). Additionally, the size and consistency of SES differences in health varies according to the health indicators under study. Obviously, a social gradient in morbidity and mortality is not present in all chronic diseases and impairments, though it has been documented for the leading causes of death (Davey-Smith, Wentworth, Neaton, Stamler, & Stamler, 1996; Marmot, Shipley, & Rose, 1984). Finally, the social gradient of morbidity and mortality among elderly populations to some extent may vary according to organisational features of the health care system, such as access to health care and quality of care according to one's economic standing (Marmot & Nazroo, 2001).
In an attempt to disentangle these different sources of variation in results, this paper analyses SES differences in health among elderly populations in the following way. First, two samples of elderly people from two countries with different health care systems are studied—Germany and the United States. There is at least one far reaching difference between the two systems. While in the United States a significant proportion of the population is not covered by health insurance, this proportion is extremely low in Germany. If similar social differences in both countries are observed, this might suggest that access to health care is not a crucial factor to explain inequalities in health. Secondly, different SES indicators are included in the current analyses to explore the consistency or relative importance of single indicators, as proposed by Robert and House (1996). Thirdly, information on different health indicators has been collected in an attempt to examine the stability of the SES–health association. Finally, we analyse intra-elderly age differences in the association between SES and health.
Before describing the study design, these specifications of our approach are explained in more detail. As mentioned, the United States and Germany are especially interesting countries for comparative analyses because there are distinct differences regarding the organisation of health care (Lüschen, Cockerham, & Kunz, 1989). The German health care system represents an approach that is located between a national health care system like the United Kingdom at one end and a market system at the other end (e.g. United States of America). It is characterised by decentralised care delivered by social security agencies. Statutory sickness funds (“Gesetzliche Krankenversicherung”) are the most important financing institutions in the German health care system. Some 88.5% of the population is covered by statutory sickness funds, while the remainder either has a private (8.9%) or a special health insurance (2.4%). The proportion of people without any health insurance is very small (0.2%) (Knesebeck & Siegrist, 2001). In the United States, patients have higher rates of co-payments, and nearly 17% of the population have no health insurance (Pescosolido & Boyer, 2001). People 65 years and older are insured by Medicare program with a co-payment of about $3,000 per year (Klein & Unger, 2001). However, there are studies showing that the poor, elderly, and others with chronic illnesses receive less appropriate treatments under more recently developed managed care models (Ware, Bayliss, Rogers, Kosinski, & Tarlow, 1996).
In studies on SES differentials in health, income, occupational status and education are usually employed as indicators of socioeconomic position. In addition to these traditional indicators, two alternative indicators (assets and home ownership) were applied in our study. These indicators are increasingly used in socioepidemiological research (Graham, 2000; Shaw, Dorling, & Davey Smith, 1999), as they are expected to be more appropriate measures for older adults, assessing economic advantage or disadvantage accumulated over the life course (Robert & House, 1996). Moreover, most of the previous studies have used only one indicator of health when exploring the relationship between SES and health. Because SES might differentially affect different dimensions of health, our analyses include three health indicators that were shown to be important in gerontological research: self-rated health, depression and functional limitations (Adam, 1998; Beckett et al., 1996; Idler & Benyamini, 1997). Finally, we analyse intra-elderly age differences in the SES–health relationship. This is done because the age subgroups within an elderly population may differ with respect to an association of social inequalities with health. For instance, if a powerful effect of selective survival operates, this association is expected to be present in early old age, but to diminish or disappear among the oldest old.
Section snippets
Methods
This study uses data from two national surveys conducted in Germany and the United States in the years 2000 and 2001. Using probability samples of non-institutionalised persons 60 years or older, data were obtained by telephone interviews (computer assisted telephone interviews) with a mean duration of about 40 min. As the proportion of unlisted telephone numbers among the aged in Germany is small, a probability sample was drawn from telephone directories. In the United States, the proportion of
Results
In Germany, income has the strongest effects on all three health indicators net of age and gender (Table 2). Education, occupational prestige, and assets have significant effects on self-rated health and functional limitations, but only having low income is significantly related to depression (model I). Moreover, there are no significant associations between home ownership and the health outcomes in Germany. In the United States, we found significant effects of all indicators of SES for
Discussion
This study examined the association of different indicators of SES with three measures of health based on self-reports (self-rated health, depression, and functional limitations) in two samples of elderly people in the United States and Germany. Associations were less consistent than previously reported for middle-aged populations. However, we found that income is the best SES predictor of the three health measures among the aged in Germany net of age, gender, and the other four SES indicators.
Acknowledgements
The project “Social Status and Health among the Aged in two Welfare Systems—A Comparative Study in Germany and the USA” was supported by the Humboldt-Stiftung, Germany. We also acknowledge fruitful discussions within the network of the European Science Foundation Scientific Program on “Social Variations in Health Expectancy in Europe”. The authors are grateful to Peter Dübbert, Volker Hüfken, Lucy Lewis, and Karim Abu-Omar for helpful assistance and to two anonymous reviewers for their
References (57)
- et al.
Internationally comparable measures of occupational status for the 1988 International Standard Classification of Occupations
Social Science Research
(1996) - et al.
Socioeconomic disparities in health change in a longitudinal study of US adultsThe role of health risk behaviors
Social Science & Medicine
(2001) - et al.
Childhood and adult socioeconomic status as predictors of mortality in Finland
Lancet
(1994) - et al.
Socio-economic inequalities in morbidity and mortality in Western Europe
Lancet
(1997) - et al.
The contribution of medical care to inequalities in health
Social Science & Medicine
(1989) - et al.
Inequalities in death—Specific explanation for a general pattern
Lancet
(1984) Place, social exchange and healthProposed sociological framework
Social Science & Medicine
(2000)Depressive Störungen im Alter
(1998)- et al.
Toward understanding the association of socioeconomic status and health
Psychosomatic Medicine
(1995) - et al.
Analyses of change in self-reported physical functioning among older persons in four population studies
American Journal of Epidemiology
(1996)