Ethnicity is difficult to define, but most definitions reflect self-identification with cultural traditions that provide both a meaningful social identity and boundaries between groups433. In this section we have considered evidence which uses various definitions of ethnicity. However, in the main these definitions are those which people apply to themselves. Thus ethnicity as used here includes cultural identity, place of origin and skin colour, and so includes white and non-white groups. Because country of birth rather than ethnic group is recorded on death certificates, mortality data are restricted to migrants only.

In the 1991 Census just over 3 million people, 5.5 per cent of the population, identified themselves as belonging to one of the non-white minority ethnic groups. Almost half had been born in the United Kingdom434. White minority ethnic groups were not counted on Census night. However recent estimates show that the Irish form the largest minority ethnic group, comprising 4.6 per cent of the population435. Data collected at the 1991 Census show that people from minority ethnic groups are more likely to live in South East England (especially London), the West Midlands, West Yorkshire and Greater Manchester. These places are home to 75 per cent of the minority ethnic population compared to only 25 per cent of the majority population434,436.

The age and gender distribution of minority ethnic groups is different from the majority population. Some minority ethnic groups have more men than women, and all are relatively young437. African Caribbean and South Asian communities have a higher proportion of households with children than the white population. Around 3 in 10 households with a white head of household contain children under the age of 16 years. Comparable figures for minority ethnic groups are over 4 in 10 for African Caribbean households, 5 in 10 for Indian households and 8 in 10 for Bangladeshi and Pakistani households. Pakistani and Bangladeshi families also have more children than families in the majority white population, whilst African Caribbean, Indian and Chinese families have similar numbers of children. Pakistani and Bangladeshi households are also larger because they are more likely to have 3 or more adults, whilst African Caribbean households are more likely to be headed by a lone parent66,438.

Inequality
Country of birth rather than ethnic group is recorded on death certificates. Thus mortality data presented below are restricted to migrants, but include migrants from Scotland, Northern Ireland and the Republic of Ireland.

Mortality
In 1989-92 mortality ratios for deaths, including perinatal mortality, from all causes for nearly all migrant groups were higher than average. However, those born in the Caribbean had a lower than average mortality ratio. For each group, except women born in Scotland, mortality from all causes fell between 1971 and 199161. Cause and age specific mortality varies by country of birth. For instance, mortality from coronary heart disease is higher than average for people born in South Asia, Ireland and Scotland and lower than average for those born in the Caribbean and men born in West Africa. Mortality ratios for cerebrovascular disease are significantly higher than the average for all migrant groups except those born in East Africa. By contrast, mortality ratios for lung cancer are low in migrant groups born in the Caribbean, Asia and Africa and high in people born in Scotland or Ireland, whereas cervical cancer mortality is high for women born in the Caribbean61,439,440. Mortality from suicide is also unusually high in young South Asian women born in India356. Mortality ratios for accidents in people under the age of 15 years and over the age of 65 years are greater in migrants from Ireland and the Indian sub-continent than those born in England and Wales252.

Morbidity
This section uses self identified ethnic group as the basis of analysis. Due to lack of data, unless stated otherwise white minority groups are included in "whites".

Overall people from minority ethnic groups are more likely to describe their health as "fair" or "poor" than the ethnic majority, although this difference comes from the poorer self-reported health of Pakistani and Bangladeshi people, and, to a lesser extent, African Caribbean people441. Chinese people consult less with their general practitioner (GP) than whites and African Asians are as likely to have consulted with their GP as whites. All other groups consult more441.

A variety of conditions show differences between ethnic groups. For example, South Asians have a tendency to central obesity and insulin resistance which may pre-dispose them to diabetes and coronary heart disease442. On the other hand, African Caribbean people have low death rates from coronary heart disease despite their high prevalence of diabetes and hypertension61. Depression appears to be more common in African Caribbeans than in whites441. Tuberculosis is more common in Pakistanis, Bangladeshis and Black Africans than in whites, and the incidence of tuberculosis in these groups is rising443.

There are limited data on morbidity in white minority ethnic groups, and they were included with the white majority in the Fourth National Survey of Ethnic Minorities, the source of much of the recent data on the health of minority ethnic groups. However available data support the view that Irish people have higher rates of morbidity as well as mortality. Analysis of the long standing illness question in the Census, for example, shows rates are higher for those born in Ireland444. Rates of hospital admission for psychiatric disorder are also high in Irish people445.

Smoking is more common in African Caribbean and Bangladeshi men where the rates of smoking (42 per cent and 49 per cent) exceed those in white men (34 per cent). Indian and African Asian men report the lowest rates (19 per cent and 22 per cent). By contrast, in women rates of smoking are low (5 per cent or less) for all groups, except African Caribbean women, where the rates (31 per cent) are similar to those in white women (37 per cent). Alcohol consumption tends to be lower in all minority ethnic groups for both men and women compared to that in the white population. Comparable information is lacking for Scots and Irish people living in England and Wales. Total abstinence is common amongst Muslim groups, predominantly within the Pakistani and Bangladeshi communities441.

In a survey of reported physical activity fewer men and women aged 16 to 74 years from minority ethnic groups than from the general population reported levels of activity which would benefit their health (defined as at least 30 minutes of moderate intensity physical activity on at least five days per week). For instance, amongst South Asian men aged 16 to 74 years, 67 per cent of Indians, 72 per cent of Pakistanis, and 75 per cent of Bangladeshis reported that they did not take part in enough physical activity to benefit their health, compared with 59 per cent of men in the general population. For South Asian women, the corresponding figures were 83, 86 and 82 per cent compared with 68 per cent of women in the general population. Furthermore, men and women from minority ethnic groups were more likely to report being sedentary than men and women from the general population446.

Socioeconomic status
(Due to lack of data, unless stated otherwise white minority groups are included in "whites".)

There are important differences between ethnic groups in factors which are associated with health, and which the Inquiry has taken the view are determinants of health. Firstly, socioeconomic status is different between ethnic groups. Compared to the majority white population (unemployment rate 6.5 per cent), Labour Force Survey estimates of rates of male unemployment are slightly higher in Indians (7.4 per cent), and considerably higher in African Caribbean (20.5 per cent), and Pakistani and Bangladeshi groups (15.9 per cent)45. The sample size is too small for reliable estimates of rates in Chinese people. Surveys of minority ethnic groups have higher absolute percentages of people out of work, but the same pattern of differences between groups66,441. Part of these differences is due to the relatively young average age of these minority ethnic groups, and the associated high rates of unemployment in young age groups in general.

Social class distribution shows similar patterns, with Pakistani and Bangladeshi groups showing a more disadvantaged profile. Perhaps most striking is the number of people from all minority ethnic groups who are living in poverty, as defined by less than half the average income. Just under a third of white households have incomes below this level, compared to a third of Chinese, two-fifths of African Caribbean and Indian households and four-fifths of Pakistani and Bangladeshi households66. Minority ethnic groups are also much more likely to be reliant on Income Support67. Although the 1991 Census showed a worse socioeconomic profile amongst Irish people, a recent survey shows that there are differences within the Irish, with men born in the Republic of Ireland being more likely to be in social class V than any other group. Men born in Northern Ireland, however, were more likely to be in social class I than men born in England, and as likely to be in social class II435. Another recent survey found relatively high proportions of Irish people amongst those earning more than £30,000 or with a university degree. Thus there may be some polarisation within the Irish group to different parts of the socioeconomic spectrum.

About four fifths of Indian and Pakistani households are owner occupied, compared with about two-thirds of white households and half of African Caribbean, Bangladeshi and Chinese households. Overcrowding is relatively common in minority ethnic households - one in ten African Caribbean and Indian households, and more than one in three Pakistani and Bangladeshi households compared with roughly one in fifty white households. Housing quality also varies. About a third of Pakistani and Bangladeshi people live in households which lack a basic amenity, for example, exclusive use of an inside toilet66,441.

Evidence
The contribution of socioeconomic inequalities to inequalities in health both within and between ethnic groups has been much debated over the last twenty years. In an examination of migrant mortality data from the 1970s, there was no socioeconomic gradient for those born on the Indian sub-continent, and an association between higher socioeconomic status and higher mortality for those born in the Caribbean. Differences in socioeconomic status did not account for differences in mortality between migrant groups447. However, analysis of more recent data (1991-93) on migrant mortality has shown a relationship between socioeconomic status and health for some migrant groups439.

All cause mortality was higher in men from manual classes than those from non-manual classes for all migrant groups, except those from West or South Africa, where the difference was present but smaller. In general this pattern was similar for the major causes of death, with the exception of coronary heart disease in men born in the Caribbean. Similar gradients have been found for self-reported health in a recent survey of minority ethnic groups441. Thus within minority ethnic groups in Britain, lower socioeconomic status is associated with higher rates of both mortality and morbidity.

However, it is not clear to what extent socioeconomic status accounts for differences in health between ethnic groups. The most recent analysis of migrant mortality suggests that socioeconomic differences, as measured by social class, do not explain the different rates of mortality between groups born in different countries439. However coding of socioeconomic status using occupation, as in social class, may be a particularly inappropriate proxy in migrants because of the high proportion of young people and women amongst them who have never worked. Within any band of social class, minority ethnic groups tend to be less advantaged than the majority white population. For example, mean income for Pakistanis and Bangladeshis is about half that found for whites in the same social class category441.

The Fourth National Survey of Ethnic Minorities used an alternative index, standard of living, which took into account material deprivation, measured by housing problems, and ownership of cars and consumer durables. Socioeconomic status as measured by this index did account for some of the differences in reported health between most ethnic groups, whilst occupational class did not441. Thus socioeconomic inequalities contribute to the inequalities in health within ethnic groups, and may contribute to the inequalities in health between ethnic groups.

The diversity of experience of health between different ethnic groups may reflect different causes of poor health; differential susceptibility to these causes; differential access to factors which ameliorate cause or susceptibility, for example, preventive health care services; or a combination of these. The Inquiry has decided to make recommendations in two general areas. Firstly, there are recommendations aimed at reducing the inequalities across ethnic groups in the socioeconomic determinants of health, given the clear evidence that these are important determinants of health in people from minority ethnic groups as they are for the ethnic majority. Secondly, there are recommendations addressed at considering the needs of people from minority ethnic groups in using services, particularly health services, which will ameliorate their experience of ill-health.

The main examples of recommendations to address differences in health across ethnic groups are considered here with a summary of their relevance to inequalities across ethnic groups, in addition to their earlier consideration in appropriate parts of the preceding text. Other recommendations may have relevance to inequalities across ethnic groups to some degree.

Benefit
In general, the benefits from policies considered here would be expected to decrease inequalities in health within ethnic groups. The benefit of such policies should be relatively wide, and be felt by those within each group who are least well off. Because minority ethnic communities typically contain a higher proportion of households with children than the white population, these communities should benefit from policies targeted at mothers, children and families (recommendations 2 and 21-23) and those related to education (recommendation 4-7). The extent to which such policies will decrease differences between ethnic groups is unknown. But the balance of evidence would favour a reduction in inequalities between groups.

Considering the needs of people in different minority ethnic groups

Although separate mechanisms might be set up to consider policies which affect inequalities in health amongst minority ethnic groups, this risks marginalising minority ethnic issues. It also implies that the health problems in minority ethnic groups are, in the main, different from those in the ethnic majority, with different causes and different solutions, whereas in fact the similarities are greater than the differences448,449. However, failure to make specific consideration of minority ethnic issues risks increasing ethnic inequalities by unintentionally favouring policies that benefit the ethnic majority. Thus policies to consider inequalities in health should include consideration of the application of these policies to minority ethnic groups as a matter of course, including ways of ensuring that racial prejudice and harassment are overcome. This requires that the structures and processes of policy-making are sensitive to the position and needs of people from minority ethnic groups. One way of achieving this is to ensure that minority ethnic groups are represented on appropriate decision-making and advisory bodies, and that other opportunities are taken to seek their views. As well as the direct effect of such representation, the visibility of people from minority ethnic groups in such positions may reduce the sense of exclusion felt by some group members.

31. We RECOMMEND that the needs of minority ethnic groups are specifically considered in the development and implementation of policies aimed at reducing socioeconomic inequalities.

Reducing poverty

People from minority ethnic groups have higher than average rates of unemployment40,45. Within minority ethnic groups, there is a clear association between material disadvantage and poor health. Very high proportions of people from some minority ethnic groups are living on low levels of income, and are dependent on state benefits. There are a number of ways in which members of minority ethnic groups may be disadvantaged by the social security system. Some of the potential problems are related to the structure of the system and its assumptions. The State pension, for example, is based increasingly on the assumption that retired people should have built up occupational or other personal provisions over their working lives, but this would be impossible for people who migrated to Britain well into their working lives. Other problems are due to a failure to consider the specific needs of members of minority ethnic groups, for instance for translated or additional information450-452. Lack of these may lead to under-claiming of benefits. The younger demographic structure of many minority ethnic groups means that policies which improve the welfare of women of childbearing age, expectant mothers and children are of particular importance.

We recommend policies which will further reduce income inequalities, and improve the living standards of households in receipt of social security benefits (recommendation 3).

We recommend policies which improve the opportunities for work and which ameliorate the health consequences of unemployment (recommendation 8).

Improving housing, safety and the material environment

Although owner occupation is quite high in some minority ethnic groups, housing quality is often poor, regardless of tenure441. Overcrowding and lack of basic amenities is more common in some minority ethnic groups. Furthermore, current housing policy supports construction of homes for relatively small households, whereas for some minority ethnic groups, including Bangladeshis and Pakistanis, requests for housing are to accommodate extended family households. In addition, some minority ethnic groups find that their choice of area of residence is restricted by fear of crime and harassment453.

We recommend policies which improve the availability of social housing for the less well off within a framework of environmental improvement, planning and design which takes into account social networks, and access to goods and services (recommendation 10).

We recommend policies which aim to improve the quality of housing (recommendation 12).

Responses to the Fourth National Survey on Ethnic Minorities suggested that more than one in eight people from minority ethnic groups had experienced some form of racial harassment in the past year66. Although most of these comprised racial insults, many respondents reported repeated victimisation and a quarter of all respondents reported being fearful of racial harassment. The British Crime Surveys have shown that South Asians and African Caribbeans are at greater risk of being victims of crime than whites. Although much of the difference in relation to African Caribbeans was explained by social and demographic factors, these did not explain the greater risk of victimisation for South Asians222.

We recommend the development of policies to reduce the fear of crime and violence, and to create a safe environment for people to live in (recommendation 13).

The use and effects of transport on ethnic minorities has been little researched, partly because of a lack of relevant data454,455. Areas, particularly inner urban areas, with high proportions of minority ethnic residents are often characterised by markers of disadvantage434,436. These include on-street parking, higher traffic volumes and lack of areas for play, and are associated with a high rate of traffic accidents amongst children from some minority ethnic groups456.

We recommend the further development of a high quality public transport system which is integrated with other forms of transport and is affordable to the user (recommendation 14).

We recommend further measures to encourage walking and cycling as forms of transport and to ensure the safe separation of pedestrians and cyclists from motor vehicles (recommendation 15).

We recommend further steps to reduce the usage of motor cars to cut the mortality and morbidity associated with motor vehicle emissions (recommendation 16).

We recommend further measures to reduce traffic speed, by environmental design and modification of roads, lower speed limits in built up areas, and stricter enforcement of speed limits (recommendation 17).

We recommend concessionary fares should be available to pensioners and disadvantaged groups throughout the country, and that local schemes should emulate high quality schemes, such as those of London and the West Midlands (recommendation 18).

Improving services

A number of studies have suggested that people from minority ethnic groups do not receive the same quality of care as the ethnic majority. Overall use of primary care is similar or greater amongst minority ethnic groups to the ethnic majority but people from minority ethnic groups are more likely than whites to: find physical access to their general practitioner (GP) difficult; have longer waiting times in the surgery; feel that the time spent with them was inadequate; and be less satisfied with the outcome of the consultation441,457. They may also be less likely to be referred to secondary and tertiary care441,458-460. Part of these differences may be related to problems with communication. A significant number of people from minority ethnic groups, particularly South Asian women and Chinese people, find it difficult to communicate with their GP441,457. There may also be cultural differences in the expression of symptoms, making the use of Western diagnostic approaches inappropriate for some groups, especially for mental illness461. Women from some minority ethnic groups, notably Pakistanis and Bangladeshis, prefer to consult with female doctors and in order to overcome communication difficulties, female doctors with the same minority ethnic background as themselves441. Given the younger demographic structure of many minority ethnic groups, the provision of sensitive maternal and child health services is of particular importance.

An illustrative example is ethnic differences in cervical screening. A national survey carried out recently found that South Asian women, especially Pakistani and Bangladeshi women, were less likely to have had a cervical smear in the past five years. About half of the Pakistani and Bangladeshi non-attenders lacked basic information about cervical screening, that is reported that they had not received an appointment or did not know what the test was457. Yet in a qualitative study carried out in the same period in East London, women from minority ethnic groups were enthusiastic about cervical cytology screening once they understood the purpose of the test and its procedures. Administrative and language barriers were important factors in participation in the screening programme, as were the adequacy of surgery premises462.

One solution is to train health workers in "cultural competency". This involves acquiring the skills to understand and be sensitive to cultural differences in the presentation of illness and treatment, and other dimensions of health463. Bilingual link workers can act as translators and advocates for people from minority ethnic groups who experience communication problems with health care professionals. Support for health professionals such as general practitioners and health visitors who are themselves from minority ethnic groups is a further strategy to increase the quality of services to people from minority ethnic groups.

People from minority ethnic groups tend to congregate in specific geographical locations, which are frequently areas of multiple disadvantage434,436. Place as well as individual disadvantage may affect health464,465. However, the concentration of people from minority ethnic groups in particular areas may also be protective of health, by preserving levels of social support and a sense of community466-468. The advantages and disadvantages are likely to be conditional upon the place, and the minority ethnic group living there, making local consideration of policies to reduce inequalities essential.

32. We RECOMMEND the further development of services which are sensitive to the needs of minority ethnic people and which promote greater awareness of their health risks.

33. We RECOMMEND the needs of minority ethnic groups are specifically considered in needs assessment, resource allocation, health care planning and provision.

There are limitations on data currently collected to assess inequalities in health across ethnic groups. Death registration collects only country of birth, and so only mortality of migrants can be considered. Yet almost half of those who identified themselves as belonging to a minority ethnic group on Census night in 1991 had been born in the United Kingdom. The high proportions of young people and women who have never held a job, and the downward social mobility that may accompany immigration, mean that classification based on occupation is inappropriate as a measure of socioeconomic status in minority ethnic groups, particularly in comparison to the ethnic majority441. Grouping of minority ethnic people, such as Black or South Asian, may be inappropriate, merging together people who have different cultures, religion, migration history, socioeconomic status and geographical location.

We recommend a review of data needs to improve the capacity to monitor inequalities in health and their determinants at a national and local level (recommendation 1.2).