Article Text

Download PDFPDF

Social inequalities in health among the elderly: a challenge for public health research
  1. Lucía Artazcoz1,
  2. Silvia Rueda2
  1. 1Agència de Salut Pública, Barcelona, Spain; Red de Investigación de Género y Salud, Barcelona, Spain; Ciber de Salud Pública, Barcelona, Spain
  2. 2Universitat Pompeu Fabra, Barcelona, Spain
  1. Correspondence to:
 Dr L Artazcoz
 Agència de Salut Pública de Barcelona, Pl. Lesseps 1, 08023 Barcelona, Spain; lartazco{at}

Statistics from

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.

Health inequalities among the elderly

Older people have tended to be neglected in research on health inequalities compared with people in other stages of life. Similarly, there has been a lack of research on how class interacts with gender in later life. These omissions are difficult to understand since health needs and the use of health services are greater among older age groups. Moreover, with the improvements in living and working conditions, as well as quality of healthcare services, the proportion of elderly people has risen significantly and will continue to increase during the coming decades. By 2050, it is expected that 30% of Europeans will be aged >60 years and whereas the very old constitute 3% of the European population today, 11 of the former EU15 member states will have at least 10% of their population aged ⩾80 years by 2050.1

There is no consensus about the best measure of socioeconomic position (SEP) in older ages—for example, the use of indicators of SEP based on occupation among people who are no longer in the labour market has been questioned. On the other hand, in trying to identify the best measure of SEP among the elderly, the importance of using multiple indicators has been emphasised. The study of Hyde and Jones2 (see page 533) addresses this issue and overcomes some limitations of previous research. It examines a broad range of SEP measures, and also restricts the study population to retirees, therefore eliminating the problem of potentially different meanings of occupational class-based measures for people who are still working and those who have left their jobs. In addition, it analyses the hypothesis that the impact of occupation-based indicators decreases with time since labour market exit (LME).

The results derived from multiple logistic regression models adjusted for age and marital status for each of the selected indicators partially support the hypothesis that time since LME affects the strength of association between SEP measures developed for use among working-age populations, and self-perceived health status. They also support results from other studies pointing out that the use of occupation-based indicators as a measure of SEP can be problematic among post-working populations.

The results of the analysis of mutually adjusted SEP measures do not confirm the association between occupation-based indicators of SEP and self-rated health status. Gender differences emerge from the study with wealth being the most important factor for health among men, whereas for women, it is subjective social status. The authors provide an interesting discussion involving the survivor effect, job characteristics and the historical experience of labour market attachment to explain these results. Additionally, certain issues related to multicollinearity may to some extent explain the final results. Future research about health inequalities in older ages should strengthen the efforts to develop theoretical frameworks about the adequacy of different measures of SEP, as well as the potential different meanings of SEP indicators depending on gender.

In older age groups, the proportion of women is higher than men and increases with advancing age. Therefore, when studying older people, it is essential to study gender as a basis of differentiation—for example, it has been suggested that older women’s much higher level of functional impairment coexists with a lack of gender differences in self-assessed health. Some studies have reported no gender differences in self-reported health status among elderly people, while higher levels of disability existed among women.3,4 On the other hand, it has been shown that gender differences in health depend on the indicator and the age stratum analysed.5 It could very well be that among older adults the impact of different measures of SEP differ by health indicator and that disability-related indicators may be more sensitive to gender inequalities than broad indicators of general health such as self-perceived health status.

Beyond SEP, family roles and the persistence of sexual division of domestic work at older ages can be important determinants of health status and gender inequalities in health. In incorporating the gender perspective in the analysis of social determinants of health, the consideration of family characteristics is crucial. Although marital status constitutes one of the most used indicators in analysing social determinants of health from a gender perspective, it has been found that the association between marital status and mortality or morbidity is weaker among the elderly.6 Using the most-frequent categories of marital status among younger adults may be inadequate in trying to examine the role of family characteristics in elderly people’s health owing to the fact that most of them are married or widowed. Household composition, on the other hand, is considered to be one of the most important determinants of the well-being of older adults.7 Although considerable interest has been focused on whether living alone increases the risk of negative health outcomes among older people, less attention has been paid to other types of living arrangements such as widows living with their adult children, whether or not as household head, or widows living with their adult children and their spouses (again, whether or not as head of household) which are very common among the elderly. Moreover, in this population caring for disabled people, primarily the spouse, should be taken into account as a determinant of health status.8 The position in the family and the associated family responsibilities are likely to be related to people’s health with patterns that probably differ by gender and SEP.9

As in most studies about the health of the elderly, the study of Hyde and Jones2 is based on a representative sample of residents living in private households. However, among the elderly, the proportion of people living in institutions is increasingly important with advancing age. Moreover, this is particularly important when focusing on the analysis of gender inequalities in health, because gender is closely associated with entry into residential care. The probability of living in residential care is influenced by an interaction among marital status, as a proxy for the availability of informal carers, SEP and age. As older women live longer than their male counterparts, they are less likely to have a spouse to provide informal care when they become disabled and therefore more likely to enter a residential setting. As older women are more likely to live in institutions than men, gender differences based on community samples will underestimate the disadvantaged health status of elderly women because of the exclusion of those living in residential settings.4

With the ageing of the population, analysis of health inequalities among the elderly has become a priority in public health. However, research on this topic is still scarce and should address some current limitations such as the often atheoretical analysis of social determinants of health status, the examination of the best measures of SEP at older ages, the identification of the best health indicators, the study of the impact of living arrangements and caring tasks, as well as their potential interaction with SEP, and the inclusion in the samples of those living in residential settings. Moreover, longitudinal analyses would help overcome some empirical problems related to the direction of causality and cohort effects.


This work was partially supported by the Red de Investigación de Salud y Género and by de Ciber de Epidemiología y Salud Pública.

Health inequalities among the elderly



  • Competing interests: None.

Linked Articles