Department of Health Skip to content

Please note that this website has a UK government access keys system.

You are here:

National standards for ethnic group and related matters

What is an ethnic group?

31. Ethnicity is complex to define as it is multi-faceted. Importantly, ethnicity is subjective: a person should self-assign his or her own ethnic group. While other people may view an individual as having a distinct ethnic identity, the individual's view of their own identity takes priority. Features that help to define ethnic group are as follows :

  • a shared history;
  • a common cultural tradition;
  • a common geographical origin;
  • descent from common ancestors;
  • a common language;
  • a common religion; and
  • forming a distinct group within a larger community.

32. While an ethnic group is sometimes perceived as a minority within a larger community, ethnic groups cover people from all communities not just those of African, Caribbean, Asian or Chinese backgrounds. For example, White British people are an ethnic group. Because they outnumber all other ethnic groups in England, comprising 87% of the population, they are the majority ethnic group at the national level. White Irish people are an ethnic group; because they only comprise 1.3% of the population, they are a minority ethnic group at the national level. Percentage figures for all ethnic groups in England are given in Annex C.

The national standard for ethnic group and its codes

33. From April 2001, DH, Trusts and councils have used, as a National Standard, a set of 16 codes to record the ethnic group of patients, services users and staff (see the box below). The codes are identical to those used in the 2001 ONS census, in accordance with ONS guidance on national standards. They are grouped under five headings  : White; Mixed; Asian or Asian British; Black or Black British; and Chinese or other ethnic group.

34. The headings should not be used as codes for direct data collection. They may be used to feedback broad findings; however, even then, feedback based on the 16 codes will almost always be preferable.

35. The 16 codes are used across Government. Use of the 16 codes helps to maintain consistency between DH central collections and ONS population information. Critically, their use enables ready comparison between NHS and social care information and national and local population counts based on the 2001 census. The codes are robust following much public consultation. It should be noted that the codes may be referred to as the '16+1' codes. The extra code is for 'Not stated', where for various reasons individuals do not, or choose not, to state their ethnic group. When used to record the ethnic group of patients, service users and staff, space should be left after each of the five 'Any other ...' codes so that the individual can describe their own ethnic group.

36. The 16 codes, presented under the five headings, plus instructions for completion taken from the 2001 ONS Census, are as follows :

Ethnic group - 16+1 codes

What is your ethnic group? Choose ONE section from A to E, then tick the appropriate box to indicate your ethnic group.

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 or other ethnic group

  • Chinese
  • Any other (please write in)

Not stated

  • Not stated

Detailed breakdown of the 16 codes

Detailed breakdown of the 16 codes

37. Both questions about, and analyses of, ethnic group may refer to a more detailed breakdown of the 16 codes. A recommended detailed breakdown, based on how ONS classified 2001 census responses, is given in Annex D. When and how local NHS bodies and councils decide to use categories from the detailed breakdown in Annex D depends on local circumstances and issues. Although the 16 codes would suit most NHS and social care situations most of the time, resort to some of the breakdowns shown in Annex D is acceptable.

38. This is how it works. If a Trust or council used the 16 codes only, people who say that they belong to the Greek Cypriot ethnic group would be coded to the Any other White background group. If there is a large Greek Cypriot community in a Trusts or councils area or there is particular interest in that community, when data are being collected the Trust or council should (with reference to Annex D) include a Greek Cypriot code under the White heading. In this way, the Trust and council can explore issues for that community while at the same time being able to re-aggregate the Greek Cypriot code back into the Any other White background code for comparison with local or national population data or with data from other places.

39. Similarly, a PCT may have a relatively large Arab community. If it did nothing, the PCT would expect people of Arabic origin to code themselves under the Any other code of the Chinese or other ethnic group heading. However, the PCT (with reference to Annex D) could include an Arab ethnic group code under the Chinese or other ethnic group heading. The PCT could then explicitly identify Arab patients and staff, while retaining the ability to re-aggregate the Arab code back into the Any other code.  Westminster PCT takes this approach in its Patient Profiling.

40. Trusts and councils should avoid using too many of the sub-codes from Annex D, the reason being that forms and questionnaires seeking ethnic group would be too long and cumbersome, and analyses too complex with the danger that important trends are lost in the detail. However, it is a matter of balance. Too few sub-codes, which mask important aspects of the local population mix, can serve just as poorly as too many sub-codes.

41. The sub-codes of Annex D may be included on forms seeking ethnic group information. Doing it this way would make coding relatively easy. Alternatively, sub-codes may be shown to patients, service users and staff if they opt for one of the five Any other & codes from the 16 codes. Their answers from the fuller code list can then be recorded on forms or data storage systems. Either way, a consistent and clear approach to Annex D sub-codes allows fuller analyses of ethnic group to be carried out consistently and in a way that can build back into the 16 codes.

42. Where records for individual current patients, service users and staff still rely on the 1991 ONS Census codes for ethnic group, Trusts and councils should consider how best they might readily update these records so that individuals ethnic group can be described using the 2001 codes. (This is important as the 1991 codes do not directly translate into the 2001 codes.)

Not stated

43. When providing information about themselves, it is not advisable to give patients, service users and staff the opportunity to record 'Not stated' on the forms and questionnaires they might be asked to self-complete. Hence the 'Not stated code' should not be an option. However, where patients, service users and staff leave the form or questionnaire blank or write on it that they are not willing to give their ethnic group, data processors should ensure that they can enter 'Not stated' on their collection and storage systems.

44. It is important for Trusts and councils to monitor the number of 'Not stated' entries on their systems. Reasons for relatively high numbers should be explored. Trusts and councils should also make provision for instances where staff administering questions about ethnic group fail to, or cannot, ask the question. Such instances are of a different order from those occasions where individuals whose ethnic group is to be coded choose not to supply the information.

45. It is generally true that 'Not stated' codes and other non-entries can be minimised when senior managers are committed to ethnic monitoring, staff are trained and patients and users are given clear explanations on why they are being asked about their ethnic group. NHS bodies will be aware that the proportion of invalid and 'Not stated' codes form a component of the Data Quality Indicator (DQI) and high proportions have an adverse affect on the overall DQI and hence on the organisation's overall performance.


46. DH encourages the appropriate collection of additional information that is often related to a person's ethnicity or culture. Typically, this information concerns religion, diet and language. It should be collected only if it is relevant for a particular facility or local area, and analysed locally. Its collection at an individual level should ensure that when individual patients and service users are receiving support, matters of their religion, diet and language may be taken into account. The benefits to both individuals and those providing the support are obvious. For example, if an individual is to spend any length of time as a hospital in-patient, a care home resident or day facility user, s/he may wish to express their spirituality in prayer or other means. They may wish to receive their care in particular ways. Indeed some forms of care and treatment may be prohibited by their religion.

47. Questions about religion can be asked by using the question and codes from the ONS Census of 2001 :


What is your religion? Tick one box only.

  • None
  • Christian (Including Church of England, Catholic, Protestant and all other Christian denominations)
  • Buddhist
  • Hindu
  • Jewish
  • Muslim
  • Sikh
  • Any other religion (please write in)
  • Not stated

48. As with the ethnic group codes, the above six codes (Christian to Sikh) may be broken down to reflect local population or specific needs. For example, it may be important to distinguish Protestants from Roman Catholics and within Protestant to identify, for example, Jehovah's Witnesses. Useful breakdowns of religion are given in 'The Health Survey for England, 1999' (National Statistics) and in 'Count me in', the census of conducted by the Mental Health Act Commission, the National Institute for Mental Health in England and the Healthcare Commission. These breakdowns are given in Annex E.  Whatever options that are offered to individuals should not be 'Christian-oriented'.

49. ONS makes the useful distinction between religions practice and religious identity. Religious practice comprises active faith or belief and participation in worship and religious identity. Religious identity is about identifying with a particular religious community even though the religion may not be practiced. Data on these two aspects of religion are generally used for different purposes. Data on religious practice would be used to inform and assist actual patient care and the provision of immediate facilities to enable the religion to be practiced, if needs be, in a care setting. Data on religious identity would help to monitor race equality obligations towards religions that are also considered to be ethnic groups in law. (See below.)

50. The 'Not stated' category should not be included on forms asking for religions for the same reasons given above with regard to ethnic group.

51. It is noted that among the detailed breakdown of the 'Any other ethnic group' code in Annex D, there are religious groups - Buddhist, Hindu, Jewish, Muslim and Sikh. If possible, (and despite the fact that for the purposes of the Race Relations Act 1976, both 'Jewish' and 'Sikh' constitute ethnic groups) these five groups should be avoided in recording ethnic group because they mask ethnic group. (For example, a Jew can be from any ethnic group.) If a patient, service user or staff member volunteers one of these or other religions as an ethnic group, they should be encouraged to choose another ethnic group from the 16 codes or other sub-codes from Annex D. However, some people see their religion as central to their identity and hence may feel reluctant to choose an alternative ethnic group. If that is how they feel and perceive themselves, their wish should be respected. Where Trusts and councils are recording religion as well as ethnic group, then individuals can be satisfied that their religion will be recorded, and this might persuade them to choose another ethnic group.


52. Trusts and councils that provide meals for patients and service users in hospitals or other care settings, might wish to consider using checklists when identifying individuals' dietary requirements. If the various diets on the checklist can be coded, then Trusts and councils can use the information to monitor the types of diets that are requested, which may influence how and what food stuffs are procured.

53. Broadly speaking, diets can be grouped into five broad types :

  • No special requirements
  • Vegetarian diets
  • Vegan diets
  • Diets related to religious practice such as Kosher or Halal.
  • Diets related to allergies and medical conditions such as low sugar/fat, high fibre diets for people with diabetes.

Trusts and councils may wish to use these broad types as a basis for their checklists. Staff should guard against assuming that because people are from particular ethnic groups or faiths, they will automatically need or ask for particular foods prepared in particular ways. Using the checklist, staff should discuss dietary preferences with each individual patient or service user.

54. Two recent publications can help Trusts and councils to think through their approach to diets for people from various faiths and minority ethnic communities. They are 'Culturally Competent Care - a good practice guide for care management', Kent County Council, 2002 and 'Community Handbook - a guide to understanding the diverse faith and ethnic communities in the UK', Ambulance Service Association, 2005.


55. For many people from black and minority ethnic communities it will be crucial for Trust and council staff to establish their ability and preferences in speaking and communicating in languages other than English. Trusts and councils should be prepared to ask for this information, and record it, at the earliest opportunity.

56. The NHS is founded on the principles of equal access and equal treatment for all. Providing communications support to service users is not an optional extra; this is driven by the requirement to comply with legislation and supporting guidance. The Disability Discrimination Act 1995, the Race Relations (Amendment) Act 2000 and the Human Rights Act 1998 make it imperative for public organisations, including PCTs, other NHS bodies and local councils, to provide language and other communications support to individuals seeking help. For people who have difficulty in communicating in English or can only readily communicate using sign language, the provision of information about services in different languages and formats is essential. Likewise, when talking through their needs and circumstances with health and social care professionals, patients and service users need to be able to communicate in the most effective way possible. This may call for professional translation and interpretation services or support.

57. Reliance on family members and friends of the individual seeking help for translation and interpretation is not a good idea, as, on the one hand, the individual may feel constrained in talking about personal matters and, on the other, family members and friends may lack the expertise and knowledge to put over the individual's views accurately.

58. Trusts and councils should note that, generally speaking, written translations of hospital or care processes, procedures, treatments and services may not be the cost-effective. Some people, resident in the UK, who cannot read English also cannot read their own preferred language. Different strategies, such as putting verbally translated information on video- or audio cassettes, should be explored. Local community and staff consultations should confirm the most appropriate approaches.

59. The 2001 ONS Census did not ask a question on language or ability in English, although some questions were asked about abilities in Welsh, Scottish Gaelic and Irish languages. However, the 2005 'Count me in', census, organised by the Mental Health Act Commission and partners, included a categorization of language that Trusts and councils may wish to use locally. The codes are as follows :

  • Arabic
  • Bengali
  • Cantonese
  • English
  • Farsi
  • French
  • Gaelic
  • Gujarati
  • Hakka
  • Hindi
  • Korean
  • Mandarin
  • Patois / Creole
  • Polish
  • Portuguese
  • Punjabi
  • Somali
  • Spanish
  • Tamil
  • Turkish
  • Urdu
  • Vietnamese
  • Welsh
  • British Sign Language
  • Any other language (specify)

60. However, local variations will need to be taken into account. For example, in London, according to recent evidence (see P Baker and J Aversely (eds), 'Multilingual Capital, London', Battlebridge, 2000) with respect to black and minority ethnic groups, the 10 most spoken languages other than English are :

  • Bengali and Sylheti (40,400 speakers);
  • Punjabi (29,800);
  • Gujarati (28,600);
  • Hindi/Urdu (26,000);
  • Turkish (15,600);
  • Arabic (11,000);
  • English-based Creoles (10,700);
  • Yoruba (10.400);
  • Somali (8,300); and
  • Cantonese (6,900).

61. Data from NHS Direct indicates that among callers requiring interpretation in the course of 2003 and 2004, the most frequently requested languages included Punjabi, Urdu, Bengali, Gujarati, Hindi, French, Spanish, Polish, Arabic, Turkish, Portuguese, Farsi, Tamil and Somali.

62. Trusts and councils should make the distinction between languages that are spoken (as some people are fluent in more than one language) and where people genuinely need to communicate in languages other than English. The 'Health Survey for England, 1999' (National Statistics) includes questions that can be used to establish competency in written and spoken languages including English.

Help with translation and interpretation

63. NHS Direct (on 0845 45 47 or at, which operates from call centres across England and is available on line, provides advice to callers on symptoms they or others are experiencing and local help that is available. Nurse Advisors or Health Information Advisors answer calls as appropriate.

64. From October 2004, NHS Direct started utilising one national supplier (Bowne Global Solutions) to provide services to callers for interpretation, translation and British Sign Language interpretation.  All callers who have difficulty in speaking English, and prefer to speak in other languages, can have a telephone interpreter on the line during a consultation. The interpretation service provides interpreters in whatever language is required, and the service is available 24 hours a day seven days a week. Callers can request that details of relevant local health resources be read out over the phone or posted to them, and information will be made available in languages other than English. The NHS Direct Access to Information Centre holds stocks of all materials in languages other than English. Information can also be made available in formats other than print, including Braille and audio-tape. 

65. The national contract which NHS Direct has procured also provides a framework contract for all NHS bodies for BSL, translation and telephone interpreting.  Trusts can purchase services through the contract and benefit from the economies of scale, delivery standards and quality assurances that have been incorporated within the contract. The contract also makes provision for a shared web-based database of translated materials, including patients information leaflets, standard letters, and so on.

66. The Government-funded "Health for Asylum Seekers and Refugees Portal" (HARP) website,, is aimed at health professionals working with asylum seekers and refugees. The site is maintained by a non-profit making organisation based at the University of East London. It includes a multi-lingual appointment card that translates appointment information into 31 languages and is freely available to NHS staff and other agencies. It also provides an on-line resource so that local organisations can share locally developed translated material.

67. Sign, the National Charity for deaf people with mental health problems, on, has developed software for use in healthcare settings. This software provides video clips of British Sign Language phrases to support communication. It can be easily adapted to provide translations into minority community languages, and can also be used to produce written information about health conditions, medications and treatments for patients to take away.

68. The importance of providing support to people who have difficulty in communicating in English, and the importance of monitoring changing patterns of language use and the take-up for translation and interpretation services, is shown in the good practice example below.

Good practice example 2: Leicester City Council

The Social Care & Health Department (SCHD) of Leicester City Council arranges for extensive translation and interpreting services so that language support is available for interaction between SCHD and those users whose first language is not English. SCHD regularly uses the interpretation skills of up to 100 sessional interpreters who, between them, speak over 65 African, Asian and European languages. The service also has access to communication in British Sign Language and translation in Braille. During 2003/2004, over 7,144 hours of service were used. Gujarati and Punjabi are the two main languages with over 43.5% usage in that year.  A noticeable increase has taken place in certain languages; these are Afrikaans, Cantonese, French, German, Italian, Japanese, Kurdish, Mandarin, Mongolian, Farsi, Pushto, Romanian, Russian, Serbo-Croat, Swahili, Thai and Urdu. SCHD continues to provide ongoing supervision, support and consultation for interpreters and efforts are made to undertake quality monitoring visits with interpreters to identify any issues and future training needs. Work will continue around identifying the need for increases in languages covered and recruiting interpreters to meet the demand. Currently SCHD is undertaking a survey of all people who have used its interpreters, in order to evaluate and assess the quality and accessibility of the service, and identify whether any equal opportunities issues are reported by service users accessing this provision.  Further work will be undertaken upon the results of the survey to ensure the service has attended to race equality and other equality issues.  

Contact :

Access keys