INEQUALITIES IN HEALTH CHANCES

Unequal health chances: Socio-economic factors The dominant tradition in work on health inequalities in the UK has focused on links between social class and measures of mortality (death rates and life expectancy) and of morbidity (diagnosed illness). Criticisms of the validity of measures of social class based on male occupation have led to the examination of other measures of socioeconomic status (Davey Smith et al. 1990 and 1998b). Paying attention to the non-employed – as the majority of social work service users are (Becker 1997) – reveals increased inequalities as well as the significance of dimensions other than occupation (Judge and Benzeval 1993; Roberts et al. 1997). It is ‘those occupying disadvantaged positions in the hierarchies of class, gender, “race” and disability who are overrepresented among households on low income’ (Graham 1995: 10). Moreover there has been increasing recognition that ‘medical’ statistics of health and illness can illuminate only part of the picture (Graham 1993). This is reflected in evidence based on self-reported health status (such as General Household Survey data) and on qualitative data. We see all these sources as complementary. However, whether relying on classification by occupation or by measures of deprivation, the evidence of substantial differences in mortality between those who are relatively well-off and those who are poor remains consistent across a range of data sources (Davey Smith et al. 1990; Drever and Whitehead 1997). For example, there is no main cause of death for which children in Social Classes IV and V have lower rates than those in Classes I and II (Woodrofe et al. 1993). The chance of a child from Social Class IV or V dying in the first year of life is over 40 per cent higher than for a child in Classes I and II (Independent Inquiry 1998).

In the early 1980s, even before the rapid increase in childhood poverty which has taken place, death rates for children aged 1–15 in Social Class V were more than double those for Classes I and II, while death rates for adults classified as ‘unoccupied’, mostly economically inactive single mothers, who constituted 6 per cent of the population, were three times as great (Judge and Benezeval 1993). Such inequalities exist throughout life (Arber and Ginn 1993) and are reflected in most of the major causes of death, including coronary heart disease, stroke, lung cancer, accidents, violence and suicide. There would have been over 17,000 fewer deaths per year from 1991 to 1993 in England and Wales if all men aged 20–64 had had the same death rates as those of Social Classes I and II (Independent Inquiry 1998). This translates into substantial differences in life expectancy: an average of five years more for men in Social Classes I and II, compared with those in Classes IV and V; a gap of three years for women (Drever and Whitehead 1997). Watt reports even greater differentials between districts. On average ‘people in the most deprived areas of Glasgow die 10 years earlier than people in its affluent suburbs’ (1996: 1026–7). Marmot and Shipley (1996: 1180) concluded that ‘important socio-economic differences in mortality persist beyond retirement age … On an absolute scale these differences increase with age’. As Watt powerfully put it, ‘dying before your time is the ultimate social exclusion’ (1996: 1027). A similar picture of physically embodied social inequalities emerges from diverse sources of evidence linking morbidity to social class. Power et al. (1998), analysing data collected on over 17,000 children born in 1958, found that at ages 23 and 33, men and women in Social Classes IV and V were twice as likely to report poor health. At age 33 this accounted for more than one person in six in the unskilled and semi-skilled groups.

Moreover, many people who do not report themselves to be in poor health are nevertheless living with long-term illness (Bowling and Windsor 1997). Evidence of links between illness and socio-economic status are paralleled in reports of pain, tiredness, sleep disturbance and emotional distress; Davey Smith et al. (1990: 374) concluded that ‘the shorter lifespan in less privileged groups seems to go with a longer period in poor health’. The effects of occupational status and illness are circular. For example, manual workers are more likely than non-manual workers to be forced out of work by chronic illness (Davey Smith et al. 1990). As we discuss in Article 6, even when a diagnosis of terminal illness has been given there are class-related differences in length of survival (Cannon et al 1994; Davey Smith et al. 1990). In ‘developed’ countries, socio-economic inequalities affect rates of ill health and death rates across society as a whole, not just among those in relative poverty (Wilkinson 1996a). For example, the longterm follow-up of a large cohort of civil servants by Marmot and colleagues found that each successive ‘grade’ of the service was linked to better health outcomes than the one ‘below’ (Marmot et al. 1984; Marmot and Shipley 1996). As the Independent Inquiry (1998: xi) put it, ‘these inequalities affect the whole of society and they can be identified at all stages of the life course from pregnancy to old age’. Reducing health inequalities cannot be achieved just by targeting the ‘socially excluded’.

Social class and deprivation do not only impact on health through the effects of income differentials, but can also be seen to mediate the impact of environmental conditions on health. There has been growing recognition of the negative effects on health, both current and potential, of environmental pollution; for example, the impact of global warming on the incidence of skin cancer, and of traffic pollution on respiratory disease (Friends of the Earth 1995). But evidence is accumulating that lower socio-economic position can expose you to greater risks. In Britain the concentration of cheaper, less well-insulated inner-city housing stock close to higher traffic concentrations is implicated in the steep class gradient of the most severe form of asthma (Cochrane et al. 1994). Moreover, the working and domestic environment contains health risks which reflect differentiated social position. Unskilled and other manual workers are particularly vulnerable to a range of pressures increasing the likelihood of workplace accidents. As Quick (1991: 87) shows, ‘weaker unions, “speeding up” processes, more small firms, higher staff turnover, casual labour and contracting all have implications for safety’. At home, inadequate income increases the risk of the disconnection of water and fuel supplies, the ‘voluntary’ restriction of heating and washing or the use of heating and lighting methods which bring increased risks of fire (Ahmad and Walker 1997; Roberts 1997). As Graham (1993: 161) reports, parents, most commonly mothers, act to cut down bills while trying to minimise health costs: ‘I put the central heating on for one hour before the kids go to bed and one hour before they get up. I sit in a sleeping-bag once they’ve gone to bed’; and ‘When the children are in bed, I turn the heating off and use a blanket or an extra cardigan.’ But such strategies are not always successful. The increased rate of death among older people in winter is partly attributed to hypothermia (Independent Inquiry 1998), linked to the combination of low income and a greater chance of living in accommodation which is difficult to heat. Unequal health chances: ‘Race’ A crucial development since the work of the Black Report has been the recognition that other dimensions to inequality, such as ethnic identity, affect people’s health chances, cross-cutting and interlocking with the impact of social class and economic disadvantage. Again there are

limitations in the methods of data collection which have been used. The failure routinely to collect and analyse evidence about mortality and morbidity based on ethnic identity in the last three decades is not just disappointing; it reflects institutional racism (Graham 1995). Statistics collected by place of birth are of limited value in examining ‘racial’ differences in health in the UK when half the Black British population is UK-born (Fenton 1997). Nevertheless they provide evidence of excess mortality among men born in the Indian subcontinent and men and women born in Africa, Scotland and Ireland (Independent Inquiry 1998). Substantially raised rates of stillbirths and deaths in the first week of life are found when mothers have been born in the Indian sub-continent (Smaje 1995). Adult Punjabi Sikhs, Gujarati Hindus and Muslims from India and Pakistan have death rates from coronary heart disease around 40 per cent higher than the majority white population (NHS Centre for Reviews and Dissemination (CRD) 1996).

People born in the Caribbean have twice the incidence of stroke compared to the general population (CRD 1996). Deaths associated with hypertension are four times higher in men and seven times higher in women (CRD 1996). Some groups also have substantially lower mortality rates from particular conditions than the majority population (for example, low rates of death from coronary heart disease amongst men born in the Caribbean), but this too has been the subject of little attention. Analysis of data based on ethnic identity rather than on country of birth shows that members of African-Caribbean, African and Indian groups and, especially, those of Bangladeshi or Pakistani origin, have raised rates of limiting long-standing illness by comparison with the majority white population (Nazroo 1997). This reflects increasing evidence that the main reason why people from Black minority ethnic groups have unequal health chances is the association between ‘race’ and socio-economic status. Smaje (1995) and Modood et al. (1997) record the greater likelihood that people in Black minority ethnic groups will suffer material disadvantage as a result of discrimination than will their white British counterparts. Unemployment rates for most minority ethnic groups are considerably higher than for whites, and the gap grew during the 1980s. Differences are greater still amongst the young and long-term unemployed. When in work, disproportionate numbers of men from minority ethnic groups are in low-paid occupations, taking into account the level of their educational qualifications, and poor working conditions – shift work, nightwork and homeworking – are more common.

People from minority ethnic groups are more likely to have poor social security rights. Housing tenure also exhibits marked ethnic patterns, with the quality of housing in each sector tending to be poorer. It is, therefore, not surprising that findings from Nazroo’s (1997) comparative study of minority ethnic groups’ health suggest that economic status is the key to differential chances of health – not only between members of minority ethnic groups and the majority population, but also between and within different minority ethnic groups. So, for example, people of Pakistani and Bangladeshi origin were found to be, on average, 50 per cent more likely to report ill health than the majority population, reflecting the evidence that over four-fifths of households in these communities have below half the average income (Modood et al. 1997). One of Ahmad and Walker’s respondents described what the combination of poor health and poverty meant for her: ‘It’s a problem finding enough money to properly furnish my house, to help me. And finding enough money to go back to Bangladesh to see my other five children and getting my daughter wed. I need help to re-unite me with at least one of my sons so that he can look after me in my old age.’ (Respondent, a Bangladeshi widow in her late 50s, with chest problems and severe money problems, whose sons have been refused entry to the UK.) (Ahmad and Walker 1997: 151) Indian and Chinese groups, whose income was closest to that of the majority population, were generally as healthy, while rates of heart disease amongst wealthy Pakistanis and Bangladeshis were little different from the majority population. Self-assessed health shows similar substantial inequalities between the majority white population and minority groups (Rudat 1994), again largely attributable to the experience of people of Pakistani and Bangladeshi identity and, to a lesser extent, African-Caribbeans.

Unequal health chances: Gender

Gender inequalities in health chances are significant, but complex and insufficiently understood, with men having a substantially lower life expectancy – about five years (Independent Inquiry 1998) – but also higher rates of self-reported ‘good’ health from childhood onwards. Again the evidence needs careful reading. Differences can be relatively small. Arber and Ginn (1993) reported that for each five-year cohort in old age only about 5 per cent more women than men assessed their health as ‘poor’ or ‘fair’. Bowling and Windsor (1997) interviewed almost 2,000 adults aged over 16 in 1996 and also found limited and variable gender differences in self-reported long-standing illness. Of the illnesses mentioned, those involving the musculosekeletal system affected almost half those reporting ill health, with heart and circulatory, respiratory and digestive problems affecting between one in five and one in seven. While levels of reported illness show little difference between men and women (see also Independent Inquiry 1998), there is evidence that women’s health is more severely affected. By age 75 and over, the physical functioning of three-quarters of women, compared to half of men, was affected by their health status. There were statistical differences favouring men over 75 in the ability to climb stairs or walk half a mile, in role limitations attributable to physical health, in limitations in social functioning, and in pain and energy levels (Bowling and Windsor 1997).