Ethnic groupings are imprecise and fluid, so any definitions used must be made explicit at the outset of a research project and will vary according to the requirements of the research. There is no universal consensus on assigning ethnic groups. In the UK, consistent groups are commonly created by permitting people to assert their own identity, in addition to using parental and grandparental place of birth. The contrasts in approaches to measuring ethnicity between the UK and the USA are shown in Table 8.1.

Ethnicity classifications provide a way of highlighting important differences and inequalities between groups. Some ethnic groups are at a higher risk than the majority population of nutrition-related chronic diseases such as obesity, cancer, type 2 diabetes and cardiovascular conditions. These ethnic differences are a key feature of health inequalities in developed countries and have been well documented among migrant and indigenous minority groups in the UK, USA and Australia (Nazroo 2010). Many populations are dynamically enriched by a range of ethnic groups and such minority groups should be key targets when tackling inequalities in health.

The main problems associated with using ethnicity as a research variable include errors of measurement and the degree of heterogeneity of groups. Senior and Bhopal (1994) suggest a number of ways in which the value of ethnicity as a research variable can be improved. Recommendations include recognising the limitations of classification methods; giving at least equal weight to consideration of socio-economic differences as to ethnic differences; not generalising the results of research on ethnicity and health over time, generations or populations with different histories of migration; and following descriptive research with detailed examination of environmental, lifestyle, cultural and genetic influences.

Measuring socio-economic status among minority groups presents its own challenges. Occupational socio-economic status alone may not be an adequate measure among minority ethnic groups, as some groups are more likely to be in manual occupations than the general population; in some cases this is despite being well educated and having previous professional jobs in their home countries (this is known as downward mobility). Multicomponent proxy measures may be more appropriate: in addition to occupation and education measures, these can include access to standard-of-living items. To ensure consistency, multicomponent and traditional measures can be compared in the same study and with national data on, for example, socio-economic status.

Conducting ethnicity and health research

Agreeing core principles for the conduct of ethnicity and health research has been a challenge to date, with a number of disciplines represented in this field. A Delphi exercise was undertaken among a range of academics and practitioners to explore the possibility of deriving common principles for ethnicity and health research (Mir et al. 2013). A Delphi exercise involves the creative exploration of ideas in a structured process of collecting information from an expert panel, usually with a series of questionnaires and opinion feedback, with the aim of attempting to reach consensus (Adler and Ziglio 1996). The ten key principles devised by this group for framing, undertaking and using ethnicity and health research (known as the Leeds Consensus Principles) are shown in Table 8.2. The group acknowledged that study design and data-collection decisions are likely to be shaped by disciplinary norms and practical constraints. However, there was widespread consensus on the need to avoid the replication of patterns of social exclusion, stereotyping and stigmatisation, and that research efforts should contribute to better healthcare experiences and better health outcomes.

Study design

Nutrition-related research among minority ethnic groups can include the whole range of study designs: quantitative (cohort, intervention and so on) and qualitative (interviews, focus groups). However, a number of design issues need to be considered and addressed. Commonly there are limited sampling frames for minority ethnic groups and therefore convenience, pragmatic and quota sampling is common (Chaturvedi and McKeigue 1994). Alternative methods for identifying target groups include name-recognition software for surname analysis, which is useful for those groups (mainly South Asian) whose names are most distinctive. Surname analysis cannot disaggregate populations according to place of origin, however, and as exogamy (marriage outside one’s own group) increases, additional methods will need to be developed. Frequently used sampling frames include general practitioner lists in the UK, which now feature ethnic monitoring, electoral registers (to determine areas of high ethnic density) and samples of occupational groups, as well as mailshots with registered mail envelopes, followed by door-knocking or screening to confirm the ethnicity of household members. The latter screening method was used to good effect to boost ethnic minority samples in the Health Survey for England 2004 (http://www.hscic.gov.uk/pubs/hse04ethnic), which included diet and nutrition related health indicators.

Additional activities are often needed to gain support for studies, particularly in terms of community acceptance and trust. Community participation or community assets approaches refer to improving the understanding of community needs and assets among partners, service providers, local policymakers, other influential community members and the broader community. These approaches are particularly useful for engaging hard-to-reach groups in order to establish trusting relationships between researchers and communities. As the ethos of the community assets approach is one of mutual benefit, improving community health and enhancing community involvement should also be goals in addition to achieving the academic research objectives. The overarching aim in this approach is therefore equitable community participation with the highest level of community involvement that is feasible or practical. Non-traditional venues for targeted community research and interventions to engage minorities, such as places of worship and barber shops, have been used extensively in the USA, but require further exploration in the UK and elsewhere (Maynard et al. 2009). Frameworks for engaging communities (including minority communities) such as Ownership, Control, Access and Possession (OCAP) in Canada, Community Based Participatory Research (CBPR) in the USA and the UK’s People-Centred Public Health principles could usefully be applied to the nutrition research context (Schnarch 2004; Minkler and Wallerstein 2010; South et al. 2013).

Cultural influences

Culture – the behaviours, traditions, beliefs and values of a particular group – has a major influence on the diet choices, food preparation, food preferences and eating habits of different ethnic groups. Traditional foods can play an important part in the diet as they contribute to the maintenance of cultural customs, social networks in destination countries and links with home countries. Among many ethnic groups it is also traditional for families to eat together.

Religion is a key cultural influence on food-related habits for a number of ethnic groups. Religious dietary laws proscribe the consumption of non-halal foods (known as haram) among Muslims (mainly Pakistanis and Bangladeshis, but also other South Asian and African groups). Other related practices include washing the hands and the mouth before and after meals and before prayers, and using the right hand for eating, the left for washing the body. Among Hindus (mostly Indians), the cow is sacred so beef is not consumed. Strict Hindus are vegetarian due to their belief in reincarnation; some avoid additional foods such as onions, garlic, eggs and coconut, often on specific days. The religious laws of Sikhism, originating from the Punjab region of India, advocate refraining from alcohol and ritually slaughtered meat. Vegetarian meals (langar) served free of charge at Sikh gurdwaras (temples) can be an important component of the diet of this group. Seventh-Day Adventism is one of the largest Christian denominations among Africans, Caribbeans and African Americans. Many Adventists are ovo-lacto vegetarians and non-vegetarians avoid pork and fish without fins or scales. Some avoid tea, coffee and other caffeinated drinks and most avoid alcohol.

Acculturation describes the multidimensional and dynamic process by which a minority adopts the behavioural patterns of the majority group (Satia 2010). A number of studies have shown the consequences of dietary acculturation to be detrimental. Examples include the increased intake of saturated fat and energy-dense convenience foods among Black Caribbeans in the UK, potentially increasing disease risk (Sharma et al. 1999), and similarly the increase in ‘Western’ eating habits and decrease in traditional food practices among the Inuit of the Canadian Arctic, consistent with a rise in diet-related disease (Hopping et al. 2010). Nevertheless, it is important to bear in mind that the influences are not all negative or one-way: mutual exchange of food habits means that while minorities are influenced by the food culture of the majority, they contribute to the availability of an ever-increasing range of new food items in shops and restaurants. There are varying degrees of acculturation between ethnic groups, and within-group variation such as in different generations should be considered, with minority children more likely to engage in majority food cultures than their parents (Khokhar et al. 2013).

While considering ‘culture’ as an explanation of differences in health outcomes or behaviours between groups, structural social processes such as socio-economic status, income and education should also be borne in mind.

Dietary intake assessment

The above cultural and other influences have a number of practical implications for assessing dietary intake among minority groups (Vyas et al. 2003; Ngo et al. 2009; Sharma 2011). Concomitantly, weaknesses in food and nutrient assessment in ethnic groups most commonly include limitations in recipe information, portion sizes and food composition data. Improvements in these areas are needed in order to carry out accurate dietary assessments and the requirement for standardised dietary assessment tools needs to be balanced with specific cultural tailoring. The acceptability and appropriateness of techniques and instruments as well as their validity are under-researched in nutrition-related studies among some minority groups. Translating study materials (see Section 8.4) may be a necessary part of the cultural tailoring of methods for those who do not speak the language of the host country. It may also be important as a mark of respect for some groups, particularly among elders. Clear descriptions of methodologies and their adaptation are required in research dissemination materials (papers, reports and so on) to enable replication and wider implementation.

Lists of the most commonly reported foods, portion sizes and food preparation methods need to be field tested for the development of retrospective methods such as food frequency questionnaires (FFQs). Confirming recipes and cooking practices with the target populations is essential, as public sources (cookery books, the internet) may differ significantly from day-to-day practices at home. A number of foods will have the same name across cultures, but have different compositions and nutrient intakes, indicating the importance of using the correct terminology for both foods and portion size. Account needs to be taken of the increasing availability of culturally specific convenience and fast foods; additional prompts may be needed in assessment methods for these and for high-energy snack foods such as patties, fried dumplings and pakoras, as well as sugary drinks.

It should not be assumed that data on ‘average’ food portion size for the majority population are applicable to minority populations, either for commonly consumed or ethnic-specific foods. There is no consistent evidence on the most appropriate method of portion size assessment among ethnic minority groups. However, use of visual aids such as food photographs or food models is considered important to avoid assumptions based on majority population data. For prospective methods employing estimated portions, differences between groups in common household measures will need to be documented.

Food composition databases designed for the majority population usually lack foods and dishes commonly consumed by minority ethnic groups. Cross-country collaborative efforts such as the European Food Information Resource Network (EuroFIR) have increased the scope, reliability and validity of minority ethnic food composition data in national databases in France, Israel, Spain, Denmark, Italy, the Netherlands, Belgium and the UK. The UK arm of EuroFIR, focusing on South Asian diets, has demonstrated that key sources of nutrients (based on chemical analysis of foods), portion sizes and eating practices differ significantly among children and adults in South Asian ethnic groups compared to Caucasian British (Khokhar et al. 2013). Among Black Caribbeans in the UK, varying portion sizes have been identified compared to the Caucasian population for local ‘British’ foods such as potatoes, and average portion size data have been established for local and ethnic-specific foods and composite dishes (Sharma et al. 2002). Ethnic differences in trends in the portion size of energy-dense foods have been identified in the USA, with the greatest increases among African American and Hispanic groups. There is, however, much scope for further development of portion size and food composition data for a number of ethnic groups. The diets of the growing UK Black African populations represent a distinct gap in the literature in Britain. The largest of these groups are from Nigeria and Ghana, with long-established communities in the UK, other European countries, the USA and Canada, but relevant dietary data are sparse. Research has been conducted updating UK food tables with the nutritional composition of commonly consumed Ghanaian dishes in London and Accra, Ghana, including folate content determined by microbiological assay (Owusu et al. 2010). The sodium content of Nigerian, Ghanaian and Caribbean takeaway meals has also been analysed. Studies of East Asian minorities have largely been carried out in the USA, with few studies in Europe, and there are a limited number of large-scale studies among migrant Eastern Europeans. Extensive development of food composition databases has been carried out for minority groups in Bangladesh, American Indian and Alaskan Native populations in the USA, for diverse ethnic groups in Hawaii and for indigenous people in Canada (see for example Kolonel et al. 2000; Hopping et al. 2010; Illner et al. 2012).