BLACK AND MINORITY ETHNIC GROUPS

Dr PS Gill,* Dr J Kai,* Professor RS Bhopal,** Dr Sarah Wild***

* Department of Primary Care and General Practice
University of Birmingham
The Medical School
Edgbaston
Birmingham
B15 2TT

** Bruce and John Usher Professor of Public Health
Public Health Sciences
University of Edinburgh Medical School
Teviot Place
Edinburgh
EH89AG

*** Lecturer in Public Health Medicine
Health Care Research Unit
University of Southampton
Level B South Academic Block
Southampton General Hospital
Southampton
SO16 6YD

1 Executive summary

Statement of the problem/Introduction

This chapter provides an overview of needs assessment for the Black and Minority Ethnic Groups (BMEGs). These groups are so diverse in terms of migration history, culture, language, religion and disease profiles that in this chapter we emphasise general issues pertinent to commissioning services. This is not a systematic review of literature on all diseases affecting BMEGs - the reader is referred to other chapters in this needs assessment series for details on specific disorders.

A number of general points are first provided as background to the chapter.

Top of page

1. As everyone belongs to an ethnic group (including the 'white' population), we have restricted our discussions to the non-white ethnic groups as defined by the 1991 census question. In addition, we do not cover needs of refugees and asylum seekers, whose number is growing within the UK.
2. Principles of data interpretation are given to highlight important problems such as the interpretation of relative and absolute risk - the relative approach guides research, while the absolute approach guides commissioning.
3. In the past, data on minority groups has been presented to highlight differences rather than similarities. The ethnocentric approach, where the 'white' group is used as the ideal, and partial analyses are made of a limited number of disorders, has led to misinterpretation of priorities. BMEGs have similar patterns of disease and overall health to the ethnic majority. There are a few conditions for which minority groups have particular health needs, such as the haemoglobinopathies.
4. The majority of the research on health status and access and utilisation of health services has been skewed towards the South Asian and Afro-Caribbean populations, with little written on the other minority ethnic groups.
5. There is an assumption that BMEGs' health is worse than the general population, and this is not always the case.
6. The evidence base on many issues related to minority health is small and needs to be improved.

The historical and current migration patterns are important to local commissioning of services. Migration of communities from minority ethnic groups has been substantial during the latter half of the twentieth century, particularly from British Commonwealth countries such as Jamaica and India.

Problems of defining ethnicity, 'race' and culture are outlined, as they are complex concepts. Ethnicity is multi-dimensional and usually encompasses one or more of the following:

It is also used as a synonym for 'race' to distinguish people with common ancestral origins. Indeed, 'race' has no scientific value and is a discredited biological term, but it remains an important political and psychological concept. Culture is briefly defined. An individual's cultural background has a profound influence on their health and health care, but it is only one of a number of influences on health - social, political, historical and economic, to name but a few.

Ethnic group has been measured by skin colour, country of birth, name analysis, family origin and as self-identified on the census question on ethnic group. All these methods are problematic, but it is accepted that the self-determined census question on ethnic group overcomes a number of conceptual limitations. For local ethnic monitoring, it is good practice to collect a range of information such as religion and languages spoken. There is a marked variation in quality of ethnic minority data collection and caution is advised in interpreting such data. Further training of staff is needed, together with mandatory ethnic coding clauses within the health service contracts.

Top of page

Subcategories

As BMEGs are not a homogeneous group, it is not easy to categorise them using standard format as in other chapters. For pragmatic reasons, we have used the following categories in this chapter:

• Indian
• Pakistani
• Bangladeshi
• Afro-Caribbean
• Chinese
• White.

Black and minority ethnic communities comprised, in 1991, 5.5% of the population of England and have a much younger age structure than the white group. It is important to note that almost half of the non-white group was born in the UK, which has important implications for future planning of services. BMEGs are also represented in all districts of Great Britain, with clustering in urban areas.

Prevalence and incidence

This section emphasises the importance of interpreting data on ethnic minority groups with care. One of the major issues is the comparison of health data of minority ethnic groups with those of the ethnic majority (i.e. 'the white population'). This ethnocentric approach can be misleading by concentrating on specific issues and diverting attention from the more common causes of morbidity and mortality. For example, while there may be some differences between ethnic groups in England, cardiovascular, neoplastic and respiratory diseases are the major fatal diseases for all ethnic groups. Even in the absence of specific local data, this principle is likely to hold.

In this section, two approaches are combined to give the absolute and relative disease patterns. Mortality in the UK can only be analysed by country of birth, and analysis has been carried out for people born in the following countries or groups of countries: India, Pakistan, Bangladesh, China/Hong Kong/Taiwan, the Caribbean islands and West/South Africa. In addition, lifestyle and some morbidity data are provided for Indians, Pakistanis, Bangladeshis, Chinese, Afro-Caribbean and white populations.

Due to the diversity and heterogeneous nature of all of the minority ethnic groups, it is not possible to give details of each specific disease by ethnic group. The top five causes of mortality (by ICD chapter) in all BMEGs are:

• diseases of the circulatory system (ICD 390-459)
• neoplasms (ICD 140-239)
• injury and poisoning (ICD 800-999)
• diseases of the respiratory system (ICD 460-519)
• endocrine, nutritional and metabolic diseases, and immunity disorders (ICD 240-279).

Mental health and haemoglobinopathies, which are specific to a number of minority ethnic groups, are also discussed.

Top of page

Services available

This section provides an overview of services available and their use by minority groups. It focuses upon key generic issues (such as bilingual services) and specific issues (such as the haemoglobinopathies) which are of concern to minority ethnic communities.

On the whole there is no disparity in registration with general practitioner services by ethnic group except that non-registration seems to be higher amongst the African-Caribbean men. Data, from national surveys, show that - in general - minority ethnic groups (except possibly the Chinese) do not underuse either general practitioner or hospital services. After adjusting for socio-economic factors, minority ethnic respondents are equally likely to have been admitted to hospital. However, it appears that use of other community health services is lower than the general population. It is still not clear to what extent institutional racism and language and cultural barriers affect service utilisation.

Even though ethnic monitoring is mandatory within the secondary sector, there still is lack of quality data for adequate interpretation.

Data on cost of services for BMEGs is not available except for language provision and the haemoglobinopathies.

Effectiveness of services and interventions

In general, current evidence on the effectiveness and cost-effectiveness of specific services and interventions tailored to BMEGs is limited. As most studies have excluded individuals from the black and minority ethnic communities, there is a dearth of data on the effectiveness and cost-effectiveness in these groups. The reader is referred to other chapters for details of effectiveness and cost-effectiveness of specific services and interventions aimed at the whole population.

The quality of care provided is considered generally and with reference to cardiovascular disease and the haemoglobinopathies. In addition, specific services, such as communication, health promotion and training interventions, relevant to minority groups are mentioned.

Top of page

Models of care recommendations

This section provides a generic framework for service development which includes the following points.

1. Services for BMEGs should be part of 'mainstream' health care provision.
2. The amended Race Relations Act should be considered in all policies.
3. Facilitating access to appropriate services by:
   • promoting access - this will entail reviewing barriers to care and provision of appropriate information on services available
   • providing appropriate bilingual services for effective communication
   • education and training for health professionals and other staff
   • appropriate and acceptable service provision
   • provision of religious and dietary choice within meals offered in hospitals.
   • ethnic workforce issues, including addressing racial discrimination and harassment within the workplace, and promoting race equality and valuing diversity in the workforce.
   • community engagement and participation.
4. Systematising structures and processes for capture and use of appropriate data.

Details of all services are not covered, as the above framework outlines the principles underpinning them. Service specifications (e.g. cervical screening) that are pertinent to BMEGs are given as examples and can be adapted to other conditions.

Outcome measures, common targets, information and research priorities

The importance of principles guiding further action on priorities are covered in this section, which include:

• national standards of quality of health care to be applied to BMEGs
• emphasis on basic needs, irrespective of similarities or differences between ethnic minority and majority populations
• emphasis on quality of service rather than specific conditions
• focus on a number of priorities rather than a large number
• being guided by priorities identified by, and for, the general population, e.g. Saving lives: Our Healthier Nation Strategy for England, as the similarities in the life problems and health patterns of minority ethnic groups exceed the dissimilarities
• considering the impact of policies and strategies in reducing health inequalities amongst BMEGs.

Top of page

As the development of outcome measures for each disease/condition and ethnic group is in its infancy, existing outcome measures need to be adapted and validated before use.

Further, to improve the quality of care for the BMEGs, the following dimensions of heath services need monitoring: access, relevance, acceptability, effectiveness, efficiency and equity.

National targets for commissioners to achieve have been set and cover:

1. the development of a diverse workforce
2. specific diseases and
3. service delivery issues.

There is a need for further information by ethnic group from primary care, as well as community and cancer screening services. The quality and completeness of ethnic monitoring data from secondary care needs to be improved. There is a need to include ethnic group data on birth/death certificates.

There are many gaps in knowledge and the following are the main priorities for further research:

• the need for incidence data on the major conditions affecting mortality and morbidity
• the evidence base by ethnic group on health status, access to services, health outcomes and cost-effectiveness of interventions is poor and needs to be addressed by all national commissioning bodies
• further evaluation of different models of providing bilingual services, such as physically present interpreters and advocates compared to telephone and telemedicine interpreting
• assessing the effect of racism on health and health care.

Top of page

2 Statement of the problem/introduction

In this chapter we are not dealing with a specific disease category but a group. Black and Minority Ethnic Groups (BMEGs) are heterogeneous - they are populations grouped together by a concept - that of 'ethnic group'. There are conceptual difficulties with defining the latter and a pragmatic definition has been adopted. We can only provide an overview of the issues that commissioners of health services need to consider to meet the needs of these diverse groups. The reader is referred to other sources for details of particular ethnic groups as well as to chapters in this series for specific diseases or services. Some specific areas mentioned in Saving Lives: Our Healthier Nation (http://www.ohn.gov.uk/ohn/ohn.htm) will be discussed, but in addition we want to highlight other priority areas which are also important for these groups.

There are some general points we want to emphasise:

1. Everyone belongs to an ethnic group (including the 'white' population). We cannot provide a comprehensive review and have restricted our discussions to the non-white ethnic groups as defined by the 1991 census question. In addition, we do not cover needs of refugees and asylum seekers, whose number is growing within the UK. Refugees, again, are a diverse group who have wide-ranging health, social and educational needs, and the reader is referred to Aldous et al(1) and Jones & Gill.(2)
2. Principles of data interpretation are given to highlight important problems such as the interpretation of relative and absolute risk - the relative approach guides research, while the absolute approach guides commissioning.(3)(4)
3. In the past, data on minority groups has been presented to highlight differences rather than similarities. The ethnocentric approach, where the 'white' group is used as the ideal, and partial analyses are made of a limited range of disorders, has led to misinterpretation of priorities.(4)(5) BMEGs have similar patterns of disease and overall health to the ethnic majority.(6)(7) There are a few conditions for which minority groups have particular health needs such as the haemoglobinopathies.
4. The majority of the research on health status and access and utilisation of health services has been skewed towards the South Asian and Afro-Caribbean populations,(8)(9) with little written on the other minority ethnic groups.
5. There is an assumption that BMEGs' health is worse than the population and this is not always the case.
6. The evidence base on minority health is now sparse and needs to be improved.

Needs assessment is a relatively new concept and the process is outlined in Chapter One and by Wright et al.(10) This is a complex process for minority ethnic groups due, for example, to cultural diversity, languages spoken, and their genetic susceptibility to specific diseases. These health needs also change with time after migration.(11) This chapter builds upon previous work undertaken on needs assessment and minority ethnic groups which provides further insight into this complex area.(12)(13)

Top of page

Migration

Migration to Britain has been occurring for the past 40 000 years from all over the world so that everyone living in Britain today is either an immigrant or descended from one.(14)

It is important to note that immigrant and ethnic group are not synonymous, and nor should it be assumed that for all minority ethnic groups, immigration is for settlement purposes.(15) 'Immigrant' refers to someone who has arrived in this country for at least a year. Figure 1 shows the growth of ethnic minority population within the last 30 years with data derived from the Labour Force Survey.(16) Note that this survey underestimates the BMEG population in comparison with the 1991 census.

Figure 1: Trend in total ethnic minority population, 1966-7 to 1989-91

Top of page

The reasons for this migration are complex and specific to groups.(17-21) During the late 1940s there was a need for labour, and British Commonwealth citizens were encouraged to come to Great Britain. This migration started with migrants from Jamaica, then the Indians arriving in the 1960s.(22) Under the British Nationality Act of 1948, citizens of the British Commonwealth were allowed to enter Britain freely, to find work, to settle and to bring their families. Many chose this option as a result of employer and government-led recruitment schemes. However, successive immigration policies since the 1960s have significantly reduced this option for persons from the New Commonwealth and Pakistan.(23) Political changes in East Africa ('Africanisation') stimulated a flow of 'Asian' refugees of Indian origin in the late 1960s and early 1970s.(24) The more recent migrants have come from the Sylhet region of Bangladesh, but most migration during the past 30 years or so has consisted of families of the earlier, mainly male, South Asian migrants coming to join their relatives.

Data for international migration for the UK are partial and complex.(25) Most of the data are based on administrative systems - related to control - rather than migrant numbers.(15) However, there is annual variation in net international migration, which contributed a third of the overall population growth.(26) Migration occurs from as well as into the UK.

The majority of people leave the UK due to work, whereas those arriving do so to accompany or join their families. Migrants to the UK are younger than those leaving. Within the UK, Chinese in their twenties are the most mobile group.(26)

Defining ethnicity, 'race' and culture

In this section, an overview of the problems of defining and describing ethnicity is highlighted, together with its measurement. A great deal of confusion surrounds the meaning of 'ethnicity' and it is commonly interchanged with 'race'. The latter is now a discredited biological term but it remains an important political and psychological concept.(27) Social scientists have been debating for some time on what different ethnic groups should be called(28)(29) - the so-called 'battle of the name'.(30) This debate has also featured in health services research.(31-35)

Top of page

What is ethnicity?

Ethnicity is also a multi-dimensional concept that is being used commonly in medical research.(34) It is neither simple nor precise and is not synonymous with 'race'. It embodies one or more of the following: 'shared origins or social background; shared culture and traditions that are distinctive, maintained between generations, and lead to a sense of identity and group; and a common language or religious tradition'.(4) It is also usually a shorthand term for people sharing a distinctive physical appearance (skin colour) with ancestral origins in Asia, Africa, or the Caribbean.(36) This definition also reflects self-identification with cultural traditions and social identity and boundaries between groups. Several authors(4)(37) have stressed the dynamic nature and fluidity of ethnicity as a concept.

What is race?

Both race and ethnicity are complex concepts that are appearing in an increasing number of publications.(33) In the United States, the collection of data on race is well established and used extensively for epidemiological, clinical and planning purposes.(38) Buffon in 1749(39) first introduced race into the biological literature. It was explicitly regarded as an arbitrary classification, serving only as a convenient label and not a definable scientific entity. Race, however, carries connotations of genetic determinism and possibly of relative value.(40) It is known that 85% of all identified human genetic variation is accounted for by differences between individuals whereas only 7% is due to differences between what used to be called 'races'.(41) Current consensus is that 'race' has no scientific value(27) as there is more genetic variation within than between groups.(42)

What is culture?

The notion of culture was first defined by Taylor in 1871(43) as:

Anthropologists have further refined this.(44)(45) It is seen as a set of guidelines which state 'how to view the world, how to experience it emotionally, and how to behave in it in relation to other people, to supernatural forces or gods, and to the natural environment'.(45) These guidelines are passed on to the next generation to provide cohesion and continuity of a society.

Hence culture is a social construct that is constantly changing and notoriously difficult to measure.(46) 'Culture' is further complicated by societies consisting of subcultures(43) in which individuals undergo acculturation, adopting some of the attributes of the larger society.(45) Although an individual's cultural background has profound influence on their health and health care, it is only one of number of influences on health - social, political, historical and economic, to name but a few.(33)(45)(47)

Top of page

Operationalising ethnicity

Given the importance of ethnicity on health, there are pragmatic grounds for assigning people into ethnicity groups. We would suggest the benefit of collecting data on ethnic group is to help reduce inequalities in health and health care. For the latter, guidelines have been recently produced for studying ethnicity, race and culture.(48)

A number of descriptions have been given to these ethnic groups - i.e. 'ethnic minorities', 'ethnic minority groups' or 'minority ethnic groups'. Note that these groups are not simply minorities in a statistical sense: they are both relatively small in number and in some way discriminated against on account of their ethnic identity.(47) As the title of this chapter states, we have used the term 'minority ethnic groups' to emphasise the question of population size. As stated earlier, we recognise that all individuals in all groups belong to an ethnic group(36) - it is simply that these groups vary in size, and the focus in this chapter is on the non-white group. In addition, the term 'black' has also been used as an inclusive political term to counter the divisive aspects of racism. Debate and controversy continues amongst other minority ethnic groups, as 'black' does not allow them to assert their own individuality in historical, cultural, ethical and linguistic terms.(49)

Several methods used to allocate individuals to ethnic groups are discussed briefly below:

1. skin colour
2. country of birth
3. name analysis
4. family origin and
5. 1991 census question on ethnic group.

Skin colour

A classification based on physical traits (phenotype) seems an obvious way to measure ethnicity. Skin colour is subjective, imprecise and unreliable.(4) For example, colour cannot distinguish between the majority 'white' group (i.e. between the Irish and English) and minority ethnic groups (i.e. between Indians, Pakistanis and Bangladeshis).

Country of birth

The country of birth has been commonly used as a proxy for ethnicity,(50)(51) as this was readily available - particularly on death certificates. A question on country of birth has been included in each census since 1841. It is an objective but crude method of classification. For example, it does not take account of the diversity of the country of origin of the individual; neither those 'white' people born in countries, such as India, ruled by the British Empire not the children of immigrants (i.e. 'second-generation immigrants') are identified by this method.(4)

Top of page

Name analysis

Name analysis has been used in several studies.(52)(53)(54) South Asian* names are distinctive and relate largely to religion,(55) where endogamy is the norm.(56) The validity of this method has been shown to be good,(55)(57) though this will diminish with increasing exogamy.(56)

A software package, developed by Bradford Health Authority and the City of Bradford Metropolitan Council, is available which can identify South Asian names.(58) This program has been shown to have 91.0% sensitivity, 99% specificity and a positive predictive value of 87.5%.(59)

Family origin

This has been used in combination with the census question in a recent study.(60) This approach, based upon country of origin, is relatively straightforward and stable, 'though individuals within particular groups cannot be considered homogeneous in respect of factors related to self-determined ethnicity and health.'(49) Both self-perception and family origin are well related.(60) The difficulty with this approach occurs when an individual responds that they have mixed family origins.(60)

1991 census question on ethnic group

Despite the inclusion in the 1920 Census Act of 'race' as an issue upon which questions might be asked, there has been a long history to the acceptance of an 'ethnic question' in the 1991 census.(61)(62) The 1991 census question on ethnic group is a pragmatic, self-determined ethnic group question which was found to be acceptable despite conceptual limitations.(63)

The 1991 census was the first in Great Britain to include a question on ethnic group. Before this, reliable information on ethnic groups was derived from data on country of birth, the Labour Force and General Household Surveys (see http://www.data-archive.ac.uk/ for further details).

The census ethnic question may not meet the needs of all researchers and commissioners, and several authors have suggested that extra information is collected, such as languages spoken and religion, to describe the groups being studied.(4)(37)(48)

Top of page

The question also does not deal adequately with people of mixed parentage(64) - most of whom have one minority parent and one white.(60) In addition, the white group conflates a number of groups which have distinct cultural, geographical and religious heritages, i.e. those of Irish, Greek or Turkish origin.

It has been estimated that the census missed 2.2% of the resident population (about 1.2 million people) due to such factors as non-response, one-person households and transient populations, and unpopularity of the community charge.(65) This undercount was not uniform across ethnic groups, age, gender, or geographic areas. To adjust for this, imputed data has been developed (Appendix B).(66)

Why collect data on ethnic group?

There are two main reasons for this. First, national data was needed to assess the scale of disadvantage and discrimination amongst the Black and Minority Ethnic Groups.(67) Secondly, primary data was required, as it was no longer viable to rely on surrogate measures, i.e. country of birth, for planning.(68)

Coding of ethnic group in the 1991 census

The 1991 census question (Box 1) included two categories - 'Black other' and 'Any other ethnic group' - to allow individuals to describe their ethnic group in their words if they felt none of the pre-coded boxes (numbered 0 to 6) was suitable. To deal with these 'written' answers and also with multi-ticking of boxes, the Census Offices developed an extended classification containing 35 categories in all (Appendix A).

Box 1: The ethnic group question in the 1991 Census of Great Britain Ethnic group White 0 Please tick the appropriate box Black-Caribbean 1 Black-African 2 Black-Other please describe...     Indian 3 Pakistani 4 Bangladeshi 5 Chinese 6 If the person is descended from more than one ethnic or racial group, please tick the group to which the person considers he/she belongs, or tick the 'Any other group' box and describe the person's ancestry in the space provided. Any other ethnic group please describe...

Top of page

Due to a number of limitations,(69) including lack of recognition of the significant Irish group resident in this country, the 2001 question as been modified as shown in Box 2. A question on religion and country of birth, but not proficiency in English language, has been also added.(69)

Box 2: Ethnic group categories proposed for 2001 Census

What is your ethnic group? Choose ONE section from A to E, then tick the appropriate box to indicate your cultural background. a. White British Irish Any other White background, please write in b. Mixed White and Black Caribbean White and Black African White and Asian Any other mixed background, please write in c. Asian or Asian British Indian Pakistani Bangladeshi Any other Asian background, please write in d. Black or Black British Caribbean African Any other Black background, please write in e. Chinese and other ethnic group Chinese Any other, please write in

Ethnic monitoring

Ethnic monitoring was introduced in all hospitals in 1995 to enable the NHS to provide services without racial or ethnic discrimination. Currently, the use and the delivery of services vary on these grounds, with or without intent, which hinders the achievement of equity in the NHS.(36) As the census categories may be insufficient to meet the needs of the local population, these categories should be adapted for the particular service and may include items such as religion, language, or dietary requirements.(70)

As there is marked variation in quality of data collection by speciality, particularly mental health services,(71) caution is advised in using this data. Further training of staff is needed together with mandatory coding clauses within contracts.(71)

There is a call for ethnic monitoring to be implemented within the primary care setting,(72) as feasibility has been demonstrated.(73)(74)

For local purposes, it is good practice to collect a range of information,(48) such as:

• self-assigned ethnicity (using nationally agreed guidelines enabling comparability with census data)
• country or area of birth (the subject's own, or that of parents and grandparents, if applicable)
• years in country of residence
• religion
• language.

Top of page

3 Subcategories

Who are they?

However we define ethnicity (see 'What is ethnicity?' above), the 'ethnic label' is a crude indicator of need. For pragmatic reasons we have used the census ethnic question to define ethnic group in this chapter. The more detailed classification is used for the majority of tables in the printed Country/Region Reports and the Local Base Statistics released in computer-readable form for further analyses by local authorities and researchers. The fourfold classification is used in the Small Area Statistics, a computerised dataset for the 145 000 Enumeration Districts and Output Areas in Great Britain.(75) These are the smallest areas for which census data is released, each containing approximately 200 households.

Pragmatic categorisation

BMEGs are not a homogeneous group, so it is not easy to categorise them using standard format as in other chapters. For pragmatic reasons we have therefore used the following ethnic group (self-assigned/country of birth) categories:

• Indian
• Pakistani
• Bangladeshi
• Afro-Caribbean
• Chinese
• White.

How many are there?

In the 1991 census over 3 million people (5.5% of the population) identified themselves as belonging to one of the non-white ethnic groups (Table 1). South Asians (Indians, Pakistanis, Bangladeshis) together formed 2.7% of the British population. 'Black' ethnic groups accounted for 1.6% of the population, with Black-Caribbeans being the largest group. Chinese were 0.3% of the population (Table 1).

Top of page

Age and sex structure

Figure 2 presents age-sex pyramids by ethnic group in which the black shading in each population pyramid represents the percentage of each ethnic group born outside the UK.(16) First note that the minority ethnic groups have a much younger age structure than the white group. The Black-Caribbean population has an hour glass structure, with the bottom half of the structure representing the UK-born children of the first-generation immigrants. Secondly, almost half (46.8%) of the non-white group were born in the United Kingdom. Note: darker shading represents persons born outside the UK.(16)

Figure 2: Age and sex distributions of persons born within and outside the UK by ethnic group 1991

Top of page

Also note that Bangladeshi men outnumber the women in the older age groups and the Pakistani pattern is similar, albeit less pronounced. Black-Caribbean women outnumber Black-Caribbean men, though part of this may be due to underenumeration of young Black-Caribbean men (see '1991 census question on ethnic group' above). Among other Asians, there is again a preponderance of females.

For further details on the major ethnic groups, see Peach 1996.(21)

Estimating future population size of an ethnic group is complicated and has to take into account not only fertility, mortality and net migration, but also ethnic identity.(76) There will, for reasons obvious in Figure 2 and Figure 3, be more elderly Black-Caribbeans and Indians. This has major implications for health and social care.(77)(78)

The assumption that minority elders have supportive extended families is false(79) - the need for health and social care will grow.

Where are they living?

Black and Minority Ethnic Groups are represented in all districts of Great Britain.(80) The geographical distribution varies across the country, with clustering in urban areas.

Top of page

Over 70% of the combined ethnic minorities are clustered in two regions of Great Britain, the South East and the West Midlands, which together contain 40% of the total population of Great Britain. These are the only regions of the country where the region's share of minority groups is higher than its share of the total population (Table 3). The Black-Caribbean and Black-African groups reside predominantly in the Greater London area. The Indians also reside in the Greater London as well as the East and West Midlands. On the other hand, there is a relatively low proportion of Pakistanis in Greater London with their greatest concentration in West Yorkshire and the West Midlands Metropolitan County. The Bangladeshis are found predominantly in Greater London particularly in Tower Hamlets.(81) The Chinese community is much more evenly distributed throughout Great Britain. Detailed geographical spread by district is given in Rees & Philips (1996).(80)

Top of page

Social class profile

Table 4 shows that socio-economic position of the minority groups differs significantly. The Chinese, Black-African and Indian males are strongly represented in class I. On the other hand, Black-Caribbean, Pakistani and Bangladeshi are over-represented in classes IV and V.

Females are less well represented in class I than males. The Chinese fare better, with nearly 70% in the higher socio-economic groups (classes I-III (NM)).

Note that this data needs to be interpreted cautiously, as it is recognised that measurement of social class by these groupings is limited. These groupings are not internally homogeneous, so that ethnic minorities could be found in lower occupational grades.(82)

Top of page

Unemployment

Figure 3 shows the variation in unemployment rates by ethnic group with the Black-Caribbean unemployment rate double the national, Black-African rates three times as high, while Pakistani and Bangladeshi rates being highest of all (29 and 32% respectively).

Figure 3: % Unemployment by ethnic group, Great Britain 1991

4 Prevalence and incidence

Epidemiological approaches

Traditional epidemiological approaches have defined priorities using data on actual and relative mortality, years of life lost, morbidity and loss of social functioning. Ethnicity and race have been used as variables for measurement of such needs by ethnic group. The most popular approach has been to compare the health statistics of ethnic minority groups in relation to those of the population as a whole or the ethnic majority - i.e. in Britain, the 'white' population. Essentially, a disease that is commoner than in the white population is declared a problem and a relatively higher priority than one that is less common than in the white population. This comparative perspective, which is ethnocentric, has some intuitive merit but can also mislead. By concentrating on specific issues, attention may be given to a narrow range of services and drawn away from ensuring that all services are equitable and available to all. This approach has led to some needs of ethnic minorities being ignored, e.g. respiratory diseases and lung cancer.

Top of page

This is shown in Table 5, which contains data originally presented by Marmot and colleagues.(51) The two columns give radically different perspectives on disease patterns. Generally, when presented using the number of cases, major health problems for minority groups are seen as similar to those of the population as a whole. When presented using the SMR, the differences are emphasised. For example, while there are some differences between ethnic groups in Britain, circulatory diseases, cancer and respiratory diseases are the major fatal diseases for all ethnic groups. Even in the absence of specific local data, this principle is likely to hold: that the important diseases and other health problems of the population generally will also be important to ethnic minority groups. The relative risk approach, which focuses on diseases more or less common in ethnic minority groups, can refine the analysis and interpretation of conclusions reached using simple counts of cases. Interpretation of data has often been misguided by an excessive emphasis on:

• differences rather than similarities
• the uncritical use of 'white populations' as a standard to which minority populations should aspire
• the use of partial analysis and datasets, e.g. looking at a limited number of conditions or particular age groups, leading to misinterpretation of the priorities.

The pattern of disease and interpretation of priorities and needs depends on the mode of presentation of data. The recommendations arising are the following.

• Base the epidemiological component of the needs assessment on ranked causes based on case numbers and disease rates.
• Refine understanding by looking at comparative indices such as the SMR, which will focus attention on inequalities and inequities.
• Draw causal hypotheses based on differences with care, and with due emphasis on social and economic deprivation as explanatory factors.
• Be aware that inferences of biological difference between ethnic/racial groups may be particularly prone to error and misinterpretation, and may harm the standing of minority groups.

In this section we combine the two approaches and give the actual and relative disease patterns. In studying the pattern of disease for health needs assessment, the following are basic items of information:

• the number of cases
• the disease rate, i.e. number of cases per unit of population per unit time, for example 10 cases per 1000 population per year
• the rank position of the disease in question based on the number of cases or rates
• the rate relative to that expected, e.g. the SMR or relative risk
• the rank based on SMRs.

Unfortunately, most existing reports and papers neither present analyses in this format nor provide the information to permit readers to extract it themselves.

Top of page

Collecting and interpreting epidemiological data for health needs assessment

Questions which are essential to the process of health needs assessment include the following.

• Which ethnic groups are to be studied? Are the ethnic categories used to define population sub-groups acceptable, ethical and accurate?
• What data need to be collected? Have we collected accurate, representative data?
• How do we derive from the data a true picture of the health and health care needs and priorities?

The answer to the first question is usually dictated by the classification used at census. For national studies reliant on census data for denominator information, this is invariably the case. While we may be interested in the pattern of health and disease in Muslims, Punjabis, Hindi-speakers, or those from the Gujarat, such patterns are unlikely to be available, at least from national data. The nearest we can get is the appropriate category at census. Clearly this is a weakness, but the census is the key to building a picture of the ethnic minority communities and analysing and interpreting most epidemiological data, and its limitations are noted (section '1991 census question on ethnic group' above).

Using pragmatic categories can be misleading. For example, one ethnic category that is commonly used is 'South Asian' or 'Asian' as a label for people from India, Pakistan, Bangladesh and Sri Lanka. This label leads to an erroneous view that South Asians are ethnically homogeneous - which may have adverse consequences for health. For example, Bangladeshi men had an extremely high prevalence of current smoking (49%) compared to all South Asian men (26%).(84) Indian men reported a prevalence of 19%, and white men 34%. The same survey showed many important differences by religious affiliation too.

The answer to the second question depends on the underlying purpose. In health needs assessment the challenge is to provide both professionals and members of ethnic minority communities with balanced information to allow them to make informed choices about priority issues and to make rational judgements on the actions to be taken. The value of mortality and morbidity data is self-evident. Despite a national policy for ethnic health monitoring, reliable national statistics on hospital utilisation are not available. Information on the patterns of (non-fatal) ill-health is difficult to obtain. Cancer registrations include country of birth and are published for some areas.

Top of page

Except in some health authorities with very large ethnic minority populations, local information on causes of death will be hard to make sense of, simply because the numbers of deaths per year will be small. Knowing the make-up of the local ethnic minority community, it is possible to gauge the major health problems by applying the findings from national data to the local populations. Even in the absence of any data on the causes of death in the ethnic group of interest, disease patterns are likely to be similar to the general population, e.g. coronary heart disease, strokes and cancers are major fatal diseases for all ethnic groups in Britain.

Lifestyle is a major determinant of health. All aspects of lifestyle which are important for the general population are important for ethnic minorities, including smoking, alcohol, exercise, diet in relation to chronic disease, and stress. These must not be overlooked when undertaking health promotion with ethnic minorities (there is evidence that this can happen). Other lifestyle issues worth noting in some communities include, e.g. the use of traditional substances such as eye cosmetics that may contain heavy metals, self-treatment with herbal and other remedies, and a strong sense of modesty, especially among women, which may affect the health (vitamin D deficiency) and health care (physical examination).(85) Many such traditional customs have been recorded and much attention has been given to them. However, their overall importance to health is small in comparison with the issues in the above paragraph.

Statistics on self-reported health status and on aspects of lifestyle are in some respects easier to interpret than disease rates, in other respects more difficult. In the two main nationally relevant sources of data - the surveys by the Health Education Authority(86-88) and by the Policy Studies Institute(84) - the main focus is on presenting numbers and percentages, usually giving the figures for the 'white' ethnic majority population. With some simple manipulation of the statistics, ranks can be ascertained and comparisons made. The interpretation of such data in the context of health needs assessment requires the same wary approach outlined for the SMR.

Note that the Health Survey for England for 1999 is focusing on BMEGs and will produce further useful data. The full anonymised dataset for this survey is available through the Data Archive at Essex University (http://www.data-archive.ac.uk/).

There are some subtle difficulties in comparing ethnic groups in lifestyle and self-reported health. The most important questions to ask are the following.

• Are the populations comparable? It is common practice to draw samples for different ethnic groups using different methods. Differences are inevitable, and may have no relation to ethnicity per se, if the samples are different. For example, if some of the ethnic populations are inner city ones, and others are a mix of urban and rural populations, differences will inevitably result.
• Are the data collected equally valid in the different ethnic groups? The concepts underpinning questions (let us say on angina) may be interpreted differently in different ethnic groups. Where questions need translating, the potential pitfalls are magnified.

Top of page

These limitations need to be remembered in health needs assessment. The validity of health statistics for minority ethnic groups is based on several assumptions: that ethnicity categories and specific ethnic group designations are not only valid but that they are consistently defined and ascertained; also that such categories and designations are completely understood by the populations questioned; that participation and response rates are high and similar for all populations questioned; and that people's responses are consistent over time.

Data

Available data on mortality and lifestyles can be re-analysed or extracted from published documents to provide a foundation in the epidemiological contribution to the health needs assessment process. The demonstration of missing gaps is important to guide future work. National hospital data are not available, and information on disease incidence, as opposed to mortality and prevalence, is unavailable.

Mortality analyses

Limitations of mortality analyses

The accuracy and validity of the numerator (death data) and denominator (population data) and the possibility of numerator-denominator bias should be considered. Death data include information on any person dying in England and Wales and thereby include deaths of visitors, but only include information on residents of England and Wales who die in other countries if these are notified to consulates. Such reporting probably varies across different populations. Recording of country of birth on death certificates, which is reliant on an informant, may be less accurate than on the census, when the person is still alive to provide the information, leading to the possibility of numerator-denominator bias (i.e. where country of birth is recorded differently in census and mortality data). Previous analyses of mortality by country of birth have grouped together countries for which this is a particular issue (e.g. South Asian countries),(51) but this approach obscures potentially important differences between countries of birth. Death certificates do not provide an accurate reflection of prevalence of certain conditions in the general population e.g. diabetes mellitus.(89) Variation in accuracy of cause of death described on death certificates by country of birth has not been studied but may exist. The census excludes people who are not normally residents, but deaths of visitors are included in the numerator. Census data is not complete and no data were obtained for 2.2% of the population in 1991. Underenumeration varied by population and was greatest for Afro-Caribbean men of 20-29 years of age.(90) The effect of underenumeration is to increase apparent mortality rates. As the census occurs only every 10 years, information on population size becomes rapidly inaccurate. Restricting the mortality analyses to the years around the census minimises the effect of population variations. In these analyses we have used four years of mortality data to increase the number of deaths to allow meaningful analysis.

Top of page

At present, analyses of mortality are limited to the use of country of birth because ethnic group is not available on death certificates. Country of birth is an inexact measure of ethnicity as demonstrated by the cross-tabulation of country of birth by ethnic group given in the 1991 census.(91) For example, of people born in West Africa, 73% described themselves as being of black African origin and 22% described themselves as being of a white ethnic group. Several studies of immigrant populations have suggested that mortality experience tends to approximate to that of the host population with both time and succeeding generations.(51)(92) The healthy migrant effect is a term used to describe the fact that migrants as a whole tend to be healthier than the populations they leave and join. There is also, however, the possibility that people migrate as a consequence of ill-health. Country of birth provides no indication of length of stay in that country. Mortality by country of birth is a particularly poor measure of health in children - very few children living in this country were born abroad and mortality statistics are a very incomplete measure of health of children. Socio-economic factors are also likely to influence migration and health.

Some of these limitations can be overcome by analysing data from the Longitudinal Study, a 1% sample of people enumerated by the 1971 census (http://www.cls.ioe.ac.uk/Research/jclr.htm). Unfortunately, the number of deaths in this dataset is too small for accurate interpretation. We have provided two tables (Tables 18 and 19) showing the major causes of death by ethnic group as a means of corroborating the general findings on the major causes of death from the national data.

Methods

The Office for National Statistics provided population and death data for England and Wales. Population data were available from the 1991 census by sex and country of birth in five-year age groups. Death data for the four-year period around the census 1989-92 were available by sex, age, country of birth and underlying cause of death coded using the ninth revision of the International Classification of Diseases (ICD-9).

For this analysis, six countries or groups of countries were studied, as for many countries the numbers of deaths were too small to permit separate tables. West/South Africa denotes data from people born in the Gambia, Ghana, Sierra Leone, Nigeria, Botswana, Lesotho, Swaziland and Zimbabwe. The term Caribbean is used to cover the following countries: Barbados, Jamaica, Trinidad & Tobago, Guyana, Belize, West Indies and other Caribbean islands. Data for people born in Hong Kong, China and Taiwan were combined into a single group that we call Chinese. Data for people born in Bangladesh, India and Pakistan are analysed for individual countries.

Top of page

Death data are presented in various forms (see 'Epidemiological approaches' and 'Collecting and interpreting epidemiological data for health needs assessment' above). The average number of deaths per year over the four-year period is given to provide information on absolute mortality and to permit the reader to assess the reliability of estimates of rates and SMR. Age-standardised death rates per 100 000 population per year were calculated by using the direct method for each sex by five-year age group with 1991 data on population of England and Wales as the standard. Comparisons between standardised rates for men and women are not directly comparable because age distribution differs between men and women. Comparisons between ethnic groups for each sex separately are possible for directly standardised rates within any age group, e.g. 20-74 in Table 7(a) or 20-44 in Table 6(a). Population data by country of birth for five age groups are given in Appendix C.

Standardised mortality ratios (SMRs) were calculated using the indirect method - i.e. reference rates generated from numbers of deaths and population data for England and Wales as a whole by sex and five-year age group applied to populations by country of birth to estimate the expected number of deaths by cause and sex. The SMR is calculated as the ratio of observed to expected deaths for various causes of death, sex and age groups with 95% confidence intervals calculated using the number of deaths over the four-year period. SMRs for individual causes of death were examined for the 20-74 year age group. SMRs cannot be compared either across the sexes or ethnic groups, as age distributions differ by sex and ethnic group, i.e. the SMR can only be compared in relation to the standard for each sex of 100.

The cause specific mortality tables are presented in rank of the number of deaths by ICD chapter. The main text gives data for the top five causes of death, again at the level of the ICD chapter. In presenting the findings, attention is drawn to the major causes of death, and where the excess is substantial, and the number of deaths is not insignificant, to high SMRs. Readers may also wish to note low SMRs, even though space does not permit the authors to comment In detail.

Mortality patterns

Tables 6-17 summarise the mortality analyses for each country of birth group. The even numbered tables show age-specific death rates for the age groups 0-19 years, 20-44 years, 45-64 years, 65-74 years, 75+ years, and also all age mortality. The odd-numbered tables give the causes of death at ages 20-74 combined. Numbers of deaths in the youngest age group are very small. These tables indicate that SMRs for large age bands can obscure differences that are noted in smaller age bands. SMRs tend to be closer to 100 for older age groups, whereas for younger age groups SMRs tend to exceed 100. As a consequence of smaller numbers of deaths at younger ages, confidence intervals around SMRs tend to be wider. The data confirm that major causes of death are not necessarily associated with high SMRs. Some of the findings of interest are discussed below for each country of birth group.

Top of page

Indian-born

Table 6(a) shows that while death rates were highest in Indian men aged 75 years and more, most deaths actually occurred in the age group 45-74, reflecting the relatively small size of the population over 75 years. The overall SMR was marginally above the population average (103), with the SMR varying by age - the value of 112 in the 20-44 age group being the most notable finding.

Table 6(b) shows fewer deaths (and lower death rates) in each age group than in Table 6(a), largely reflecting women's better survival compared to men. The overall SMR was 113, indicating that Indian women had higher mortality than the whole population of women. (Men and women cannot, for reasons already discussed, be compared on the SMR or the all age standardised rate).

Top of page

Top of page

Top of page

Circulatory diseases, and specifically ischaemic heart disease, were the dominant causes of death in men (Table 7(a)) and women (Table 7(b)). These SMRs corroborate past analyses showing these diseases as 30-50% more common in Indians compared to the population as a whole.(5)(6) The rates/100 000 show that Indian men have much more circulatory disease than women, a point obscured in SMR analyses.

Neoplasms were a dominant cause of death, even though the SMR is lower than in the whole population, and in contrast to the little attention they sometimes receive, the commonest neoplasms in Indians are lung cancer in men and breast cancer in women.

Injury and poisoning was the third ranking cause of death in men, and the fifth in women (Tables 7(a) and 7(b)). The SMR for women was raised.

Death from diseases of the respiratory system is common and only slightly less common than in the whole population. The importance of digestive disorders as a cause of death is noteworthy, as are the high and relatively high rates of cirrhosis in men (but low in Indian women, see Table 7(b)).

Diabetes mellitus is substantially commoner in Indians, men and women, than in the population as a whole, and a major killer. For all these diseases the cardiovascular risk factors, including smoking, are of prime importance in either initiating or promoting disease.

The sizeable variations in the SMRs in various conditions are worthy of note, particularly for cirrhosis in men, tuberculosis in men and women and nephritis.

Pakistani-born

Table 8(a) shows, strikingly, that while death rates are highest in the oldest age groups, most deaths occurred in 45-64 year olds (reflecting the population structure). The overall SMR was lower than the population average for men, with an excess only in the under 20 year age group.

Table 8(b) shows that the number of deaths and death rates were lower in women than men. Again, in comparison to the population average for women, there was a raised SMR in the under 20 year age group but overall the SMR was substantially lower than the population average.

Top of page

Top of page

Top of page

Top of page

In Pakistani men, and to a lesser extent in women, circulatory diseases dominate (Tables 9(a) and 9(b)). In women, the SMR for ischaemic heart disease was only 11% higher than the whole population, with a bigger excess in cerebrovascular disease. As for Indians, neoplasms were the second ranking cause of death. Diabetes mellitus outranked respiratory diseases in men and women. Cirrhosis was, unlike Indians, not especially common in Pakistanis. Injury and poisoning were high in Pakistani men (Table 9(a)), but not so in women (Table 9(b)) .

It is noteworthy that infectious and parasitic diseases, though relatively very common (SMR = 335), were the fifth ranking cause of death in Pakistani women.

The data demonstrate the vital importance of controlling cardiovascular risk factors, including smoking, and better control of diabetes in Pakistanis.

Bangladeshi-born

Table 10(a) shows a huge preponderance of deaths in the 45-64 age group, though, as before, death rates rose with age. The SMR was raised, compared to the population average, in this age group, but was substantially lower in the others. For women (Table 11(b)), numbers of deaths and death rates were substantially lower than in men. The overall SMR, and SMRs within each age band, were substantially lower than the population average.

Table 11(a) shows that in men, the disease patterns were similar to Indians and Pakistanis (circulatory disease and neoplasms dominating), with an exceptionally high SMR from liver cancer and diabetes. Cirrhosis was a relatively common cause of death in men but not woman.

Table 11(b) shows that the number of deaths in women were small, but neoplasms and circulatory diseases were the commonest cause of death. In women, coronary heart disease rates were relatively low in comparison to the whole population.

Bangladeshi men are in urgent need of interventions to reduce their cardiovascular risk and control diabetes.

Top of page

Top of page

Top of page

Chinese/Hong Kong/Taiwan-born

As shown in Table 12(a), deaths were mostly in the 45-74 age group in Chinese men, though death rates were highest in the older age groups. The high number of deaths over 75 years in Chinese women reflects the substantial population in the age group (Appendix C). The number of deaths (and death rates) were higher in men than women (Tables 12(a), 12(b)). The SMR was lower in Chinese men and women, compared to the population average, in virtually every age group.

Top of page

Top of page

Top of page

Top of page

In Chinese men and women (Tables 13(a) and (b), neoplasms were the top ranking cause of death (lung cancer being in the commonest single cancer in men, and breast cancer in women), with circulatory diseases second. In men, the commonest circulatory disease was ischaemic heart disease, but in women it was cerebrovascular disease. Injury and poisoning was the third ranking cause of death. In both men and women, infections, though an uncommon cause of death, were relatively common, with high SMRs, including for tuberculosis. SMRs for some specific causes were very high, e.g. for liver cancer, nasopharyngeal cancer and lip/oral/ pharynx cancer (Tables 13(a) and (b)).

Caribbean-born

As shown in Tables 14(a) and (b), most deaths occurred in the 45-64 age group, but the death rates were higher in older age groups and in men at each band.

The SMR for men overall shows mortality rates similar to the population average, though the SMR was substantially higher in the age group 20-44 years and substantially lower in those over 75 years. In women, the overall SMR was higher than the population average for women, with a substantial excess in the age groups 20-44 and 45-64.

Top of page

Top of page

Top of page

Tables 15(a) and (b) show that in both Afro-Caribbean men and women, circulatory disease, neoplasms and endocrine diseases (mainly diabetes) were dominant causes of death. It is worth emphasising that ischaemic heart disease (IHD), which has a low SMR, is the commonest of the circulatory diseases in Caribbean-born men, particularly as this disease may be overlooked in favour of stroke, which has a high SMR. In a similar vein, the low SMR for cancer, including for lung and breast cancer, must not obscure their importance as common causes of death. Endocrine diseases, mainly diabetes, were exceptionally common, with extremely high SMRs in men and women.

The infrequency of deaths from respiratory disease (in absolute and relative terms, especially in women) is notable (Tables 15(a) and (b)). High SMRs were particularly notable for hypertensive heart disease and stroke, liver cancer, prostate cancer, tuberculosis, nephritis and deaths from symptoms/ill-defined conditions.

West and South African-born

Tables 16(a) and (b) shows that in men and women most deaths were in the 45-64 age group, but with the usual pattern of rising mortality rates with age. Relative to the whole population of men, the mortality rate was high, especially in the younger age groups.

For women, too, most deaths were in the 45-64 age group, and the number of deaths and death rates was lower than in men. The SMR shows death rates higher than the population as a whole in those aged up to 64 years, and lower thereafter.

Top of page

Top of page

Top of page

The disease pattern in men and women was different, as shown in Tables 17(a) and 17(b). In men, the usual pattern was observed. with circulatory diseases and neoplasms dominant, though IHD had a low SMR. Hypertensive disease and cerebrovascular disease were both common, and had very high SMRs. Injuries and respiratory disease were major killers. Diabetes was relatively common. The high SMRs for liver cancer, infections, symptoms and ill-defined conditions and genito-urinary disorders were noteworthy.

In women, the number of deaths were small but, nonetheless, neoplasms dominated (breast cancer being the commonest) over circulatory diseases. Ischaemic heart disease comprised a small fraction of circulatory deaths and was relatively uncommon, being exceeded by cerebrovascular deaths. Although the SMRs were high for several specific conditions, the number of cases was too low for accurate interpretation (Tables 17(a) and 17(b)).

A note on 'South Asians' and the inclusion of 'East Africans'

A common practice over the last 15 years is the combination of Indians, Pakistanis, Bangladeshis, and sometimes Sri Lankans and East Africans too, into one category, 'South Asians'. As the above tables show, there are similarities and dissimilarities in mortality. Overall, it is probably wise to recognise the substantial heterogeneity in these populations' health needs, even though the study of the separate groups poses additional challenges of smaller population size, and fewer deaths.

We have examined the data for Indians, Pakistanis, Bangladeshis and Sri Lankans as a single group of 'South Asians' together and East Africans separately. The data are not presented here, but we conclude that study of such a South Asian group is reasonable for diabetes, but not for several other causes.

Mortality by ethnic group - the Longitudinal Study

One per cent of the enumerated 1991 census population of England and Wales was identified for the Longitudinal Study (LS) (http://www.cls.ioe.ac.uk/Research/jclr.htm). Table 18 shows the numbers of Indians, Pakistanis, Bangladeshis, Chinese, Black-Caribbean, black Africans and Whites in the longitudinal study (for this chapter, and, analysis, the categories 'black other',' other Asian' and 'other' are excluded). These populations are 'flagged' and traced at the NHS Central Register, from where mortality data are obtained. Table 18 shows that the population size for the ethnic minority groups is small, especially for Bangladeshi, Chinese and Black African populations. The patterns are likely to be least reliable for them.

Top of page

Table 19 ranks the causes, giving numbers of deaths defined by ICD chapter. Table 19 shows that circulatory diseases were the top ranking cause of death with the exception of Black-Caribbeans, in whom this place was taken by neoplasms.

Top of page

Table 20 gives the numbers for a small number of specific causes and confirms the burden placed by the specific causes of ischaemic heart disease, stroke, diabetes and the two common cancers The number of deaths is too low to permit valid sex- and age-specific rates, and hence age-sex adjusted rates, to be calculated. In view of the substantial differences in population structure, rates unadjusted for age and sex would be potentially misleading. The important point is that the ranking of causes of death, as summarised in Table 20, is similar to that arising from country of birth analysis. This gives confidence in undertaking health needs assessment for adults based on the data in Tables 6-17.

Lifestyle, measures of health and self-reported health

Table 21 summarises the studies from which the data have been extracted. The general findings are summarised below. In comparing different groups, the reader needs to remember that different methods of sampling and questioning in different languages makes precise comparisons between ethnic groups difficult. Tables 22-27 summarise key data on lifestyles, biochemical measures, anthropometric measures, and self-reported and self-assessed health in six ethnic groups. These data are a sample of the extensive information available. Readers are advised to read the original source to understand the method before utilising the data.

The paucity of research on racism on health is discussed by Bhopal,(7)(96) though it is acknowledged as a factor in terms of housing(97) and education.(98) One study from the US found an association between racial discrimination and hypertension,(99) possibly operating via the 'psychosocial pathway'.(100)

Top of page

Indians

Indians are extremely heterogeneous, so findings are likely to differ in different places, and communities. In particular, religion has an important effect. For example, smoking is much less common in Sikhs than Hindus. The reverse applies to drinking alcohol. That said, the data in Table 22 show that there are substantial needs in relation to smoking, alcohol and lack of physical activity. In women, the cultural taboo against smoking is holding, for the present.

Top of page

Lipid profiles in Indians change dramatically after immigration, moving from very low levels towards the high levels of cholesterol in the white population.(101) Vigorous action to alter lipid profiles is warranted.

Indians are relatively short and obesity (particularly central) is common. Indians born in the UK are growing taller than their parents. Blood pressures vary in different Indian communities, with the best judgement being that levels are similar to the white population - i.e. hypertension is a common disorder.

Diabetes and the associated syndrome of insulin resistance are exceptionally common in men and women. The presence of cardiovascular symptoms is high, and in some studies reflects mortality data.

Mental health problems are present in a substantial proportion of the population.

These data, together with the mortality patterns and other findings in the research literature, show that Indians present health needs that are similar to the population as a whole. Special emphasis is needed to sustain the low prevalence of smoking in women, and vigorous control of all the risk factors for diabetes and cardiovascular diseases.

Pakistanis

Pakistanis are mainly Muslims, whose religion impacts in ways important to health. Although heterogeneity between Pakistani communities should not be overlooked, this is less than in Indians. As with Indians, there are substantial needs in relation to smoking (men) and in promotion of physical activity (Table 23). Few people drink alcohol, though the taboo against it may lead to underreporting. Those Pakistanis who do drink may have special difficulties due to social problems arising from admitting to an alcohol problem.

Top of page

The comments above on lipids and physical measures of health including obesity in Indians, apply with even greater force in Pakistanis, whose rates of heart disease and diabetes are slightly higher than in Indians. The reduction of cardiovascular and diabetes risk factors is the prime health need in Pakistani adults. The indicators of mental health status suggest major needs, as does the high prevalence of self-reporting poor health/longstanding illness.

Overall, these data, combined with the knowledge that Pakistanis are relatively poor, indicate an especial challenge in meeting the health needs of this population.

Bangladeshi

Of the South Asian populations in the UK, the Bangladeshis are the most homogeneous, having in common a single major religion, Islam, and origins from a small country, Bangladesh, and within that many Bangladeshis come from Sylhet. Table 23 shows that smoking prevalence in Bangladeshi men is exceptionally high, making this the priority public health issue. Although the prevalence of smoking is relatively low in Bangladeshi women, tobacco chewing (with betel nut or paan) is a common practice, and much more so than in Indian or Pakistani women.

The points made on alcohol use in Pakistanis apply to Bangladeshis, too. The exceptionally low rates of physical activity (a major issue) need to be interpreted in the knowledge that most men are in manual occupations.

Lipid patterns in Bangladeshis are problematic, with the apparently low total cholesterol being a result of very low HDL cholesterol. This, together with high triglycerides, signifies a need for dietary advice and change.

Bangladeshis are very short, a reflection of poor nutrition in childhood. In comparison with other ethnic groups, Bangladeshis have less obesity and a lower mean blood pressure. This should not lead to complacency, for their risk of developing cardiovascular disease and diabetes is the highest of all the ethnic groups considered here. It may be that cardiovascular risk is triggered at a lower threshold than in other ethnic groups.

Self-reported health problems are common, though surprisingly, the prevalence of mental health problems is comparatively low. This may simply reflect difficulties of translating questions in comparable ways, or it may arise from social and cultural factors yet to be studied. As Bangladeshis are the poorest of the ethnic minority groups studied here, and the most recent immigrants, one might anticipate their mental health to be worse.

Top of page

Top of page

Afro-Caribbean

While Afro-Caribbeans come from a diaspora of Caribbean Islands, each with their distinctive characteristics, they have in common a language (English), and are predominantly Christian.

The need for services relating to smoking cessation, alcohol drinking and exercise uptake is clear from the data in Table 25. The cholesterol levels are high, but triglycerides are low. The reasons why Afro-Caribbeans have a comparatively low mortality from coronary heart disease despite their unsatisfactory risk profile is unclear. The possibilities of data artefact, or a temporal trend, need to be considered, and the view that African Americans were protected from coronary heart disease (CHD) has not been sustained.(102) An epidemic of CHD may be imminent.

Obesity is common, as in the population as a whole, and weight control is a priority in the light of the high blood pressure and high prevalence of diabetes.

Mental health problems are extremely common, especially in women, and the prevalence of suicidal thoughts is significant. The problem of poor self-assessed health and longstanding illness is an indicator of high levels of health need.

Top of page

Chinese

China is a vast territory, yet it is surprisingly homogeneous, mainly as a result of its long history as a single political entity and ancient civilisation. Chinese people in Britain are either agnostic, Christian or Buddhist, and most speak Cantonese (87%).

The smoking prevalence is substantial in men, though low in women. There is a need for smoking cessation activity for men, and actions to maintain the low levels in women. The low prevalence of physical exercise is problematic.

The lipid profiles and measures of physique come from a single survey in Newcastle in the early 1990s.(94) In the absence of other data, the cautious interpretation is that the lipid profiles are favourable and Chinese people's physique is slim. This accords with the comparatively low rates of CHD mortality. The challenge for services is to maintain or improve upon this comparatively advantaged position. Mortality data show cardiovascular disease as the second commonest cause of death in Chinese. On self-report, (Table 26) cardiovascular disease and diabetes are common. There is no room for complacency.

The prevalence of symptoms indicating mental health problems is high in Chinese (excepting suicidal thoughts).

Top of page

White population

The difficulties in making comparisons have been discussed above. Nonetheless, for interest and reference, some of the comparative data are in Table 27. While assessing the health needs of the white population is beyond the remit of this chapter, it would be remiss not to point out that there are multiple and diverse populations captured by the term 'white', and these populations may have distinctive health needs.

Top of page

A synthesis of current knowledge on the patterns of disease in ethnic minority groups

The following synthesis is based on a reading of the literature, particularly the reports summarised in Table 27, and examination of the data tables. Note that preliminary analysis of data collected during the first months of 1999 Health Survey for England broadly substantiate the conclusions presented below and in other sections (for further details, see http://www.doh.gov.uk/public/hs99ethnic.htm).

Ethnic minority groups are heterogeneous in their health. In terms of both overall health (say, measured by the all-cause SMR or self-reported health) and specific causes (say, coronary heart disease or oral cancers) there is marked heterogeneity. There is also great heterogeneity within ethnic groupings.

There is a common assumption and oft-stated view that the health of Britain's ethnic minorities is worse than expected (judged by the standard of the ethnic majority (white) population). This is at best simplistic, and sometimes wrong. First, such conclusions need to be cautious in the light of the possible weaknesses in the underlying data, particularly those based on mortality statistics. Second, even on the basis of the published statistics, overall measures such as SMRs are often around and sometimes less than 100 in some ethnic minority populations. There is the subtle question of how we judge the level of expected health. Is it right to base the expected level on the white population which, on average, has much higher economic standing? Might it be that, taking into account social and economic factors, the health of ethnic minority groups is about that to be expected? Certainly, overall SMRs in ethnic minority groups tend to be on a par with people in social classes IV and V in the general population. It is worth noting that some of the highest all-cause SMRs are not in the ethnic minority groups but in a sub-group of the white population - Irish and Scots living in England.(103)(104)

In many if not most respects, for mortality and morbidity, the ethnic minority groups have similar patterns of disease and overall health to the ethnic majority. This is plain when disease rankings are based on frequency as in Tables 5, and 7-18. In their detailed community-based study of South Asians in Glasgow, Williams et al(105) concluded that 'South Asians were consistently disadvantaged only in terms of anthropometric measures. Otherwise, the many differences were balanced, with disadvantage being concentrated only among South Asian women.' This general conclusion holds in this analysis.

There are some differences in disease pattern that need attention, but not at the expense of potentially more important diseases that show no striking differences (such as respiratory diseases). Conditions which are less common in minority ethnic groups than in the white population tend to be ignored (e.g. lung cancer, the leading cancer in men in most ethnic groups, and among the leaders for women) but may be worth more attention than conditions which are actually less common (though relatively more common than in the white population), e.g. liver cancer.

Top of page

The differences are complex and vary over time and between ethnic groups. Simplifications may easily mislead. It should be noted that information is most readily available for Afro-Caribbean and South Asian groups, is poor for Chinese origin people, and unavailable for most other groups, e.g. those from the Middle East and many groups of refugees.

With the above provisos, the following generalisations seem to be sound, consistent across studies, and unlikely to be explained by artefacts: the major cause of death, and both the serious and minor health problems, of most ethnic minority communities differ little from those of the population as a whole. For example, coronary heart disease, stroke and cancer are the commonest cause of death, and accidents, poisonings, digestive disorders, respiratory infection and circulatory problems the main reasons for admission to hospital, whichever community you consider. Health professionals caring for ethnic minority patients will usually be confronted with these common problems, and will see the conditions specific to ethnic minorities infrequently. Their problem will be to make the correct diagnosis in the face of communication barriers of one kind or another. However, both health authorities and individual practitioners need to know of the conditions that are rare in the population as a whole and yet sometimes seen in minority ethnic communities. Health authorities may need to modify their service priorities and practitioners may need to consider their approach to diagnosis.

Some of the conditions that are much commoner in one or more minority ethnic groups than the indigenous community include:

• infectious diseases including tuberculosis and malaria
• diabetes mellitus
• perinatal mortality
• hypertension and cerebrovascular disease
• cancer of the oropharynx; cancer of the liver; cancer of the prostate
• haemoglobinopathies
• vitamin D deficiency.

Equally, there are some conditions which are less common in one or more minority ethnic groups relative to the population as a whole, including:

• many cancers, including the common ones of lung and breast
• mental disorders
• diseases of the nervous system and sense organs.

For most specific conditions, the SMR is not consistently high in every ethnic group, for example, ischaemic heart disease is relatively common in Indian, Pakistani, and Bangladeshi populations but relatively uncommon in the Chinese and Afro-Caribbeans

The above lists are not comprehensive. Health authorities have the difficult task of ensuring that their services cater not only for the common causes of death and disability but also take account of any unusual patterns of disease in their population. Some specific diseases that merit discussion include the following.

Top of page

• Diabetes: This is much commoner in Afro-Caribbean and South Asian minority groups than in the population as a whole. In the Chinese, the prevalence (in Newcastle) is on a par with the white population but on the basis of a higher prevalence of impaired glucose intolerance,(94) there is evidence that a rise is imminent. The causes of the high rates are likely to be a mix of genetic, lifestyle, environmental and economic factors.
• Coronary heart disease: This is moderately higher in South Asian groups than in the population as a whole, with increasing evidence that the poorest groups, of Pakistani and Bangladeshi origin, have the highest rates. The causes of the excess are incompletely understood. Recent work(84)(93)(94) indicates that socio-economic factors are important. The role of the classic risk factors (high blood pressure, lipids, smoking) is clearly important. Central obesity and insulin resistance are two other factors of especial note. Coronary heart disease is one of the foremost killers of other ethnic groups, including Afro-Caribbean and Chinese, even though the rates are lower than in the population as a whole.
• Stroke: This is highest in Afro-Caribbean populations, but also the rates are relatively high in the Chinese and South Asian groups. The major known associated risk factor is high blood pressure, which is extremely common in Afro-Caribbeans but not in the others. This tendency to stroke is commonly attributed to genetic factors. Other causes, including racism, are being investigated. Stroke is an extremely important cause of death in all other ethnic minority populations.
• Respiratory diseases: These tend to get little attention. The mortality and morbidity from these diseases is usually a little less than in the white comparison populations, which makes them extremely common and important problems which ought not to be neglected.
• Neoplasms: Overall, cancers tend to be less common in ethnic minority groups than in the 'white' comparison population (but a dominant problem, nonetheless). Some cancers are strikingly less common, e.g. lung cancer - relating to lower smoking prevalence. Nevertheless, this cancer remains the top ranking cancer in men. For some cancers, the SMRs are strikingly different from the population as a whole. Oropharyngeal cancers are commonest in South Asian groups and prostate cancer in African origin groups. Cancer variations are usually attributed to environmental factors.
• Infections: The common respiratory and gastrointestinal infections are dominant and important in all ethnic groups. Diseases that are associated with warm climates, such as malaria, are much more likely in ethnic minority groups. Tuberculosis is dramatically commoner in most ethnic minority groups, particularly South Asian ones. The causes are complex - relating to opportunities for exposure (travel, migration, etc.), immunity and living conditions in the UK. The latter seems to be an important factor maintaining the high level of tuberculosis in South Asians settled in the UK.
• Haemoglobinopathies: The haemoglobin disorders - thalassaemias and sickle cell disorders - are important genetic conditions that affect people who originate from Africa, the Caribbean, the Middle East, Asian and the Mediterranean. It is important to distinguish between carriers of haemoglobin disorders, who are very numerous, and people who have a major haemoglobin disorder, who are relatively few.(106) Carriers are healthy but due to recessive inheritance there is risk of having a child with a major disorder. The risk of these disorders in some ethnic groups is shown in Table 28.

Top of page

Top of page

• Childhood mortality: Perinatal and neonatal mortality rates, and those in the age group 1-14, tend to be higher in most studies. The exception is the comparatively low incidence of sudden infant death syndrome demonstrated in some ethnic minority groups. The causes are complex and poorly understood. The high perinatal mortality rate amongst Pakistanis maybe linked to consanguineous marriages,(111) but this remains controversial.(112) Consanguineous marriages are defined as between close relatives, usually between second cousins or closer. These kinship patterns are found throughout the world and not restricted to the Muslim community.(113) These marriages can lead to an increase in rare recessively-inherited disorders, but the effect of this on the disease patterns of the population as a whole has been exaggerated.(113)
• Mental health: Numbers of deaths directly attributable to mental illness are small, and mortality data are not helpful to distinguish ethnic differences in prevalence of psychoses and neuroses. Suicide rates among migrants in England and Wales in 1979-93 generally reflect patterns in the country of origin, and migration does not appear to increase the risk of suicide.(114) SMRs for all age groups for suicide in 1988-92 were significantly lower than 100 in men born in Bangladesh, Sri Lanka and Pakistan and for men and women born in the Caribbean commonwealth, although SMRs for suicide among the Caribbean-born were elevated in the 25-34 year age group. SMRs for suicide are significantly higher among women born in India, with marked excess for deaths by burning.(115) Migrants to and from a variety of countries have higher rates of admission to psychiatric hospitals than native-born populations and in the United Kingdom African-Caribbeans have higher admission rates and receive a diagnosis of schizophrenia more often than do members of other ethnic groups(116-8) A prospective study of incident psychosis found that annual incidence of schizophrenia and other non-affective psychoses was higher than the white population in all other minority ethnic groups studied, but the difference was only significant for the black population.(119) Conflicting evidence exists regarding rates of hospital admission for psychosis among people born on the Indian subcontinent, and this may reflect differences between sub-groups of this diverse population. Possible explanations for differences between migrants and native populations include higher incidence of psychiatric disease in the host country, the effect of migration, selection bias of migrants, confounding by socio-economic factors, differences in seeking medical care, prejudice in medical practice, inequitable service utilisation and drug use. Descriptive epidemiological studies of use of treatments have suggested that African-Caribbeans have low rates of depression and South Asians have low rates of all mental illnesses.(120) Results of the British Fourth National Survey of Ethnic Minorities provided a different picture, possibly as a consequence of the use of incidence rather than prevalence rates.(121) Rates of mental illness among Asians who had been educated in Britain or who were fluent in English were similar to those of the white population, suggesting the possibility that the instruments used to detect depression were less sensitive among other Asian groups.(122) Similar methodological limitations were suspected in a study of the prevalence of dementia and depression among elderly people in black and ethnic minorities in Liverpool. No differences were found between English speaking ethnic groups and the indigenous populations, but dementia was found to be more prevalent among non-English speaking groups.(123) The prevalence of anxiety and depressive illness was similar in African-Caribbeans and whites in a population-based survey in Manchester.(124) Limitations of this survey include the definitions of the ethnic groups and the low response rates, giving a potentially unrepresentative sample.
• Sexual health: Limited information is available concerning the sexual health of ethnic minority populations. South Asian men in Glasgow reported lower use of condoms than non-Asian men.(125) A high prevalence of sexually transmitted disease including gonorrhoea and chlamydia has been reported among Afro-Caribbeans in London, Leeds and Birmingham.(126-9) Data from anonymous seroprevalence surveillance suggest that the risk of HIV among pregnant women from sub-Saharan Africa is much higher than that of other populations, such that 76.4% of seropositive newborn babies were delivered to women from this population in 1997-8. Seroprevalence of HIV was highest among women born in East Africa (2.3%) and Central Africa (1.9%), compared with 0.14% overall. There was little evidence of HIV in women born in Southern Asia (0.0081%) and none within UK-born Asian communities.(130) Data from the General Household Surveys of 1991-5 were used to examine fertility and contraception among ethnic minority women in Great Britain.(131) Fertility in Pakistani/ Bangladeshi women was more than double that of white women and was associated with reduced use of contraception. The study highlighted potential unmet family planning needs and the need for cultural sensitivity in provision of family planning services. Married, non-professional Asian women have been found to experience difficulties in using family planning services, largely due to communication problems with health professionals and low levels of personal autonomy.(132) Abortion data are not recorded by ethnic group but there is a rising abortion rate among Bengali women in Tower Hamlets.(133)
• Nutrition: Whilst some BMEGs have lower incidence of specific diet-related diseases when resident in their country of origin, this difference gradually reduces as they adopt the more 'western' foods and cooking practices.(134) Further, there are also some diet-related problems that are commoner in BMEGs, such as vitamin D and iron deficiencies amongst South Asian children under 2 years.(135)(136) There is an association between low plasma vitamin D and iron deficiency anaemia, particularly in winter.(136)(137)

Top of page

Conclusion

Patterns of health and disease are profoundly influenced by genetic, cultural, socio-economic and environmental factors. Undoubtedly, important differences exist between human populations in such factors. It would be most extraordinary if one of the consequences was not differences in health and disease by ethnicity, which is linked to the factors mentioned. Indeed, such differences between ethnic and racial groups can be shown with ease. The difficulties are not in demonstrating differences but in interpreting their meaning and using them to benefit the population.

Why is a disease more common in one group of people than another? This question lies at the heart of the debate on inequalities in health. Answers to these questions contain essential and unknown truths about the causes of disease. Answers will benefit all populations. Epidemiologists, who attempt to unravel the mystery in the patterns of disease in populations, become intrigued by ethnicity and health research, and particularly the mechanisms by which disease differences occur.

One major explanation, which has had insufficient attention, is the role of socio-economic status. On arrival in Britain most migrants held unskilled jobs. This legacy has been passed to their children (though there are many exceptions) and ethnic minority communities have more than their share of unemployment and low paid work. Much of the health disadvantage associated with ethnic minority groups may not result from their racial and cultural background, but relate to their socio-economic disadvantage. Their health status may be comparable to social classes IV and V in the indigenous population, and the solutions to health problems may also be similar. The problem of inequity and inequality in the health and health care of ethnic minority groups has defied easy solution. The explanation is not simply lack of knowledge, interest or even money. Inequalities may widen in the face of both interest and research - the most clear-cut example being the black/white disparity in life expectancy in the USA.(96)

The challenges of gaining, interpreting and utilising information on the pattern of health and disease in ethnic minority groups are great. To avoid traps, health needs assessors should: understand the strengths and limitations of the concepts of race and ethnicity, and the population sub-groupings derived to categorise people; ensure that all the relevant data and modes of presentation are used to produce a balanced analysis; and give due emphasis to both similarities and differences and draw tentative and careful interpretations of the causes of differences. Above all, they should avoid portraying differences as demonstrating the inferiority of some population group - that path has sustained and nourished a racist scientific literature. For health needs assessment, the common diseases and other common health problems deserve the most attention. Health needs assessors must avoid being deflected by the attention given to controversies generated by ethnic differences.

The approach used here has been to focus first on the important problems and diseases, then to refine the sense of priority using the relative approach. This approach avoids the piecemeal approach to tackling so-called ethnic health issues. Statistics cannot make coherent policy, without principles that guide their interpretation. This section is therefore as much concerned with the principles of data interpretation as with the data itself.

Top of page

5 Services available

Introduction

This section considers services available and their utilisation in so far as data are available. It focuses upon key generic issues of concern to the health care needs of minority ethnic communities. Bilingual services, in particular, are considered. The reader is referred to other chapters for detail upon specific diseases and services, although certain pertinent issues relating to BMEGs are mentioned here.

Access to appropriate services

A central question for health authorities, trusts and Primary Care Groups (PCGs) is the extent to which minority ethnic populations enjoy equality of access to appropriate health services. Variation in effective access to services may be important sources of inequality in the health experience of different ethnic groups, impacting upon quality and outcomes of care.(84) The variations in health described in section 4 might be partly explained by differences in service use. These may reflect demand for services rather than inequality of access to them. However, differences in demand may also result from a failure of health services to appropriately address the needs of minority ethnic groups.

In addition to levels of ill-health, the demand for, and use of, services will depend upon a wide range of factors including knowledge of services and how to use them, health beliefs and attitudes, the sensitivity of services to differing needs, and the quality of care provided.(138) These raise the key issues for health professionals of effective communication, awareness of attitudes, culture, stereotyping and racism within consultations and broader aspects of service delivery.(139)

Variation in availability and use of services

It must be stated again that even though ethnic monitoring is mandatory for some aspects of secondary care, relevant data remains incomplete and of variable quality for interpretation. However, in some localities, data may be of sufficient quality.(71) There appears to be considerable variation in availability and use of services in different localities. This may reflect several factors including:

• historical lack of performance management of, and variable commitment to, appropriate service development for ethnic minorities
• lack of awareness or relevant training about ethnic health and diversity issues
• professionals' differing attitudes towards people from ethnic minorities
• lack of relevant information (including inadequate ethnic monitoring) and therefore mechanisms to inform clinical audit, service planning and delivery
• nature of the population and variation in demand for services.

It is still not clear to what extent institutional racism and language and cultural barriers affect service utilisation and quality of care. Many services, for example bilingual services, are underdeveloped or may be underused by patients or the professionals facilitating their health care.

Top of page

Health service utilisation

Primary care services

In general, a high proportion of people from most ethnic groups appear to be registered with a general practitioner - with registration rates of 99-100%,(86)(140) but African-Caribbean men have higher non-registration rates (4%). Minority groups are also significantly more likely to attend open GP surgeries than those offering appointments(141) and may wait longer to see their GP.(86)(142) With the exception of the Chinese, minority groups have comparable or higher consultation rates with their GP than the general population.(84)(86)(143)(144)

As ethnic monitoring is not yet mandatory within primary care, there is currently little routine data available. However, data (Tables 29-35) is available from the National Morbidity Statistics from General Practice study done in 1991.(145) Essentially, 60 practices in England and Wales provided data for one year on face-to-face contact with 502 493 patients. Two percent of these patients were from ethnic minority groups compared to 6% in the 1991 census. The data in the tables are a re-analysis done by ONS and are not identical to those in the published report. This new analysis includes consultations with a nurse (although the study did not record nurse consultations if a doctor was also consulted during the same visit). The standard population for calculating the standardised patient consulting ratios (SPCR) was the entire study population including those for whom there was no ethnicity code (17% of patients). This group's consultation rates were low. As a result the SPCR for the white population is high at 108 for men and 105 for women. The interpreting of the data requires caution as the sample is not representative, the number of people is small, and 95% confidence intervals are not given (for technical reasons). Nonetheless, these are the best data available that provide a national picture. As with the mortality tables, the causes for consultation are ranked by approximate frequency of consultation (based on the numbers for women at all ages).

For each of men and women, in the three age groups and at all ages, the tables show the number of consultations, the consultation rate (crude), the age-standardised consultation rate (both per 10 000 patient-years at risk), and the age-standardised patient consulting ratio, where the entire population in the study provides the standard i.e. 100. The number of people in each age group was small, and this applied particularly to those over 65 years (the exception to this is the white population). The causes of consultation often varied by age and sex, usually in a predictable way. For example, the standardised consultation ratio for infectious and parasitic diseases was higher in children than in adults, diseases of the blood and genito-urinary systems were commoner in women than men, and diseases of the circulatory system were commoner in men than women. The consultation rate for mental disorders in men was half that in women. In all minority ethnic groups, except the Chinese and White groups, boys aged 0-15 years had a higher consultation rate than girls. The interpretation of the patterns is shown in detail for Indians, as an example, and briefly for other groups.

Top of page

In Indians (Table 29), for all diseases, the standardised rates were higher in women than in men - mainly because of substantially higher consultation rates in women 16-64 compared to men. The standardised ratio shows that Indian men had a 12% excess of consultations compared to the whole population, and for women the excess was 2%. The commonest causes of consultation in Indians were factors influencing health status and contact with health services, respiratory problems, musculo-skeletal and connective tissue disorders, problems of the skin, and problems of the nervous system and sense organs. It is noteworthy that at general practice level, diseases of the circulatory system are not one of the dominant problems, and neoplasms are a rare cause of consultation.

Top of page

The standardised ratio picks out conditions that are relatively common or relatively rare. Surprisingly, the ratio for infectious and parasitic diseases was close to 100. The conditions that were comparatively high were: endocrine disorders; blood; respiratory; circulatory; and symptoms, signs and ill-defined conditions; and those that were comparatively low were: neoplasms; mental disorders; and genito-urinary.

The pattern of consultation for Pakistanis (Table 30), shown in Table 30, was broadly as described for Indians. Overall, consultation rates for women exceeded those for men. In both men and women, compared to the whole population, there was a 9% excess of consultation in men and 8% in women. For most conditions, the consultation rates were slightly higher than in Indians, but this did not apply to the circulatory system. The substantially raised standardised ratio for endocrine disorders, for digestive system disorders and for symptoms and signs were noteworthy.