Close window to return

Issue 2, July 1998, pp.12-14.

Changing health of women in England: 1984-1993
Mel Bartley and Amanda Sacker

Introduction
Social class differences in mortality among women increased during the 1980s, despite general improvements in health and longevity. Death rates fell in all social classes, but the decrease was 22% among those in higher professional and managerial households and only 6% in women in non-skilled manual households.(1) At the same time, women's domestic lives and occupational roles have changed. More women are now living alone, lone parenthood is more common and more married women are going out to work. Despite greater employment opportunities for women, many are still concentrated in clerical and sales jobs or in semi-skilled and unskilled manual work.

Measuring health and social position
In this first report from our study, we looked at ill-health, rather than mortality, and asked whether health inequalities have changed for women in the 1980s and 1990s as a result of changes in their domestic and employment circumstances.

It was possible to do this because two similar health surveys were done about ten years apart: the 1984 Health and Lifestyle Study (HALS) and the 1993 Health Survey for England (HSFE). Both of these asked about women's (and men's) employment, marital and family situations as well as their health. Women in poor health were identified according to their response to a survey question 'how is your health in general?' The answer to this rather vague sounding question, usually referred to as 'general self-rated health' has been found in many other studies to be a good predictor of later heart disease and other causes of mortality.

Our study took a measure of social position which is widely used in the UK to study things like voting and political beliefs, but which has not yet been used very much in health studies. Because nowadays we think of health as being importantly influenced by people's ideas, attitudes and behaviour, it is useful to have a measure of social position which is known to be related to these aspects of life. The Erikson-Goldthorpe (E-G) system classifies occupations into social classes according to the conditions and relationships which people in each job typically experience: for example the amount of job security they have, whether or not there is a career structure with pay rises each year and chances for promotion, and how much autonomy a person has over their own working time. There is strong research evidence from other studies that job security and autonomy are related to health. E-G classes range from professionals and managers in Class 1 to non-skilled manual workers in Class 5.

Women were also classified according to their domestic roles and how they divided their time between home and paid work: either working full-time, working part-time or keeping house. Parental status was split into those with and without children under 16 years living in the household. The changes in the proportions of women in different combinations of social roles between 1984 and 1993 can be seen in Table 1.

Findings so far: widening inequalities
There has been no decline in the scale of health inequality in women between the 1980s and 1990s. After taking account of age, women in (or previously in) non-skilled manual occupations were over twice as likely to report that their health was less than good than women in professional occupations in 1984. The same comparison in 1993 showed that the gap had widened further: women in the non-skilled manual class were nearly three times more likely than women in the professional class to be in poor health. This inequality in women's health remained, even after differences in occupational and domestic roles were taken into account. In 1984, after adjusting for differences between women in professional and non-skilled manual work in terms of their social roles (employment, marital and parental status), professional women were still half as likely to have poor health as women in non-skilled manual work. In 1993, the likelihood of poor health for women in non-skilled manual work reduced from 2.7 times to 2.3 times that for professional women before and after taking these factors into account.

Even though social roles explain a lot of the differences in health between women, they clearly do not explain why women in different occupational classes had more or less ill-health. Moreover, despite the marked increase in employment amongst women between 1984 and 1993 (see Table 1), the social gradient in health, that is, health differences according to occupational class, had not decreased.

Table 1: Distribution of variables (percentages) for women aged 20 to 59 in the Health Survey for England (HSFE) and the Health and Lifestyle Survey (HALS).

The influence of children, employment and marital status
But what happened to health differences between women in different social roles? Work and children affected women differently in the 1980s and 1990s. Women without children at home and in full-time paid employment had better health than those remaining at home, with part-time workers holding an intermediate position. However, women without childcare responsibilities who had no paid work tended to have worse self-rated health than their working counterparts in 1993 than in 1984 (see Figure 1). For women with children, a different trend could be seen. In 1984, women with children at home had the best health if they were working part-time, with higher rates of poor health reported by women in full-time work or keeping house. In 1993, working full-time with children was associated with having the poorest health whilst women remaining at home with their children had the best health.

The health of women was also affected by marital status, particularly for those who had no paid jobs. Married women had the best health overall. But in 1984, single women without paid work were nearly 5 times as likely to have poor health as married women working full-time, with previously married women close behind (see Figure 2). In 1993, the health gap between single women without paid work and working married women was smaller than it had been in 1984, but divorced, widowed or separated women without paid employment remained far less healthy than those with both jobs and partners/husbands.

Figure 1: Relative risk for poor self-assessed health by employment status and maternal status

Women at risk of poor health
By looking at each combination of social roles in the two time-periods, we identified a group of women who may be particularly at risk for poor health in the 1990s.

This conclusion was reached on the basis of a statistical model, but it can be illustrated by drawing a picture of a 'typical' member of the group. It would be accurate to imagine her as having been out of the labour market for many years while her children grew up and then finding herself without the resources to return to work once on her own through divorce or widowhood.

Whilst there may have been many women in this position in the 1980s, the expectation in the 1990s is that women on their own will seek work. However, women who were over 45 or 50 in 1993 were too old to have benefited fully from the changes in education and law which reduced the level of discrimination against women who wished to acquire skills. Those who, for this reason, are without up-to-date occupational skills may be increasingly disadvantaged in today's labour market when competing against younger people for work. In the 1993 survey we used in our study, only 6% of previously married women without children at home remained non-employed if they had at least GCSE level qualifications compared with 23% of those with no qualifications. Against prevailing trends towards better health in the 1990s, the health of this small group has remained poor, with 54% reporting poor health in both surveys.

Figure 2: Relative risk for poor self-assessed health by employment status and marital status

Conclusions
It is clear from these results that social position defined in terms of employment relations and conditions is related to health, and this inequality is largely unaffected when other aspects of women 's lives are also considered. However, the contribution of domestic roles and occupational characteristics is complex. It has significantly changed over the period 1984-1993 in ways which need to be understood in the context of rapid changes in the opportunities and expectations of women. In particular, the evidence would seem to point to the particular vulnerability of older women without work, as a group who are at greatest disadvantage in terms of health in the 1990s.

Mel Bartley and Amanda Sacker both work in the Department of Epidemiology, University College, London, and run the project in the Health Variations Programme on 'Social variation in women's health: work or way of life'. Mel Bartley is also a member of the project team for 'Dimensions of health variations over persons, time and place.'

Reference:
1. Drever, F. and Whitehead, M. (eds.) (1997) Health Inequalities: Decennial Supplement, London : The Stationery Office.