Fertility, marriage and household: associations with the health and mortality of women in later life

Emily Grundy and Cecilia Tomassini, London School of Hygiene & Tropical Medicine

[Project number 30091]

The importance of life course influences on health and mortality differentials in later life is increasingly recognised, not just because past experiences exert strong influences on current circumstances, but also because accumulated stresses and disadvantages may have independent consequences for health outcomes in older age groups.

Associations between family and household circumstances and health in elderly age groups have attracted research attention, not least because of concerns about the implications of the increasing proportions of older people living alone and high rates of marital disruption in cohorts now approaching later life. Among women the currently married appear to have better health than the widowed or divorced, but in some studies the never married, or sub groups of the never married population, appear to be even healthier. These, and other ambiguities, in results from this body of research partly reflect design and data problems in some studies. Much of the work on associations between marital status and health, for example, has been based on analyses of samples which exclude the population in institutions, a major source of bias. Additionally many studies have included poor control for socio-economic characteristics. Interpretation of associations is complicated by the need to allow for selection effects which vary considerably between cohorts.

Associations between fertility behaviour and health and mortality in later life have been less extensively investigated. There are a number of well established associations between parity and timing of childbearing and particular morbidities, e.g. breast cancer. In a recent study using data from the ONS Retirement Surveys Grundy and Holt (2000) found that marital history and parity and timing of childbearing were associated with indicators of health and disability in a cohort in early old age, but sample size was relatively small and findings were suggestive rather than conclusive.

Several studies have reported associations between parity and all cause mortality with the nulliparous and those of high parity having the highest risks. These include a recent study based on analyses of the ONS Longitudinal Study, however in this the only other characteristic allowed for in the analysis was presence/absence of a higher educational qualification so confounding by socio-economic and marital history factors was not adequately allowed for (Dobblehammer 2000). Collaborative work with Professor Wadsworth indicates associations between timing of fertility and health and mortality risks in mid life, these results apply to one birth cohort (1946)and in the first part of this project an important aim will be to see whether mortality risks are similar or not in earlier birth cohorts.

The aims of the proposed project are:

1) To analyse associations between parity, timing of births (including spacing) and health status (limiting longstanding illness) in 1991 and mortality.

Hypothesised associations: 'Deviant' childbearing patterns associated with greater health risks because of stress caused by being 'off time'. Examine differences from cohort average. Early, late, and closely spaced births, and high parity associated with greater risks in all cohorts.

2) To analyse associations between marital history and health and mortality.
Hypothesis: Marital disruption associated with greater risks of mortality and reported longstanding illness in 1991, but this may vary between cohorts and socio-economic groups. Remarried and cohabitees worse health than the married, but better than divorced, widowed or never married. The disadvantage of being never married will be variable according to proportion of and cohort group never married.
(In a later phase of the project we also plan to examine associations between history of living arrangements, mortality and health status as reported in 1991 and 2001. A separate proposal will be submitted to the LSRB for further phases of the work).

Phase I of this project was approved on 4 September 2002; a supplement of Mortality of Mothers of Twins was approved on 14 May 2003; and Phase II was approved on 6 October 2003. An extension to this project was approved on 8 August 2008 and a new project number assigned. This project is supported by CeLSIUS.

Top of page