Collection of data on ethnic origin in England

A recent BMJ article marking the 60th anniversary of the NHS noted that “the patchy collection of ethnic data by the NHS needs to be improved.”1 Why are such data important, what is the current position, how is it likely to change in the near future, and how does data collection in the UK compare with other countries?

Around 7.7 million people in England (15% of the population) belong to ethnic minority populations, defined as all ethnic groups other than white British (and therefore including white Irish and white other). In some areas—such as the London boroughs of Newham, Tower Hamlets, and Ealing—ethnic minorities comprise more than 50% of the local population. In 46 of the 354 English local authorities, more than 20% of the population is not white British. The population of the United Kingdom is also ethnically diverse and changing, as exemplified by the recent inward migration from Eastern Europe.

Government targets for reducing inequalities in life expectancy focus on “spearhead” areas with the highest mortality and greatest deprivation, but the latest figures show that inequalities are widening.2 In England, about 37% of people from ethnic minority groups live in these areas, and they make up about 20% of the spearhead population. Furthermore, significant epidemiological differences exist between ethnic minority groups, and—contrary to the spirit of personalisation central to the Darzi review3—people in minority ethnic groups sometimes report poorer experience of using services.4 Some health disadvantages in ethnic groups are related to socioeconomic determinants, but these effects often cannot be distinguished from other effects because of inadequate data.

Information is the bedrock of a good health service. Improving health and reducing inequalities require information to support needs assessment, which should then be reflected in the planning and commissioning of services. The Local Government and Public Involvement in Health Act 2007 places a duty on local authorities and primary care trusts to undertake joint strategic needs assessments to inform commissioning priorities.

Effective commissioning requires information about population numbers, health status, risk factors, disease prevalence, and access to health care by key determinants of need such as ethnic origin. Such information is essential also for monitoring the effect on access and equity of policy initiatives such as patient choice, the 18 weeks target, and introduction of NHS funded independent healthcare services, such as independent sector treatment centres.

The Race Relations Amendment Act 2000 places duties on public bodies to promote race equality and undertake race equality impact assessments on all policies. The Darzi review highlights the role of information in patient choice and quality improvement.3 Without suitable data, we will not know whether ethnic differences in access, quality, outcomes, and choice are being ameliorated or exacerbated.

Historically, however, information about the ethnic origin of patients in health datasets has been poor, as noted in the Cabinet Office’s equalities review.5 The equality review of the national screening programmes, for instance, noted that ethnic inequalities in access and uptake cannot be measured because of a lack of information.6 With a few exceptions, such as hospital episode statistics and the mental health minimum data set, ethnic origin is not routinely recorded in most NHS datasets or in primary care, although more than 90% of all patients’ contacts are with general practitioners. Ethnic origin is not known for sentinel outcomes such as the 636 000 births and 470 000 deaths annually in England, because only the mother’s country of birth is recorded at birth registration and the deceased’s country of birth at death registration.

However, important developments are in the pipeline. In response to the equality review’s recommendations,5 the Office of National Statistics conducted a cross government review of the availability of data for monitoring equality, and the Department of Health’s Equality Monitoring Group, chaired by the permanent secretary, Hugh Taylor, is dealing with filling in the gaps. Mandatory recording of ethnic origin in commissioning datasets for outpatients, people attending accident and emergency departments, and births was announced in June, for implementation from April 2009. The “standard” contracts for commissioners introduced by the department include uniform information requirements across both NHS and (NHS funded) independent health care. Proposals announced by the department in May 2008 to improve access to general practitioner services include consultation with the BMA on the recording of ethnic origin and first language by general practitioners. The report recommends that practices should “collect data as indicated in the national minimum ethnicity dataset” and the department should “promote ethnicity data monitoring in primary care and measure progress towards achieving equality.”7 Although parliament rejected legislation enabling the collection of ethnic origin at registration of births and deaths some years ago as being too complex to implement, the Office of National Statistics has used data linkage to compile infant mortality for England and Wales according to ethnic group.8 Similar approaches could be used to analyse other deaths by ethnic origin.

But, at what stage of contact between patients and the healthcare system should ethnic origin be recorded? Currently the position on this is unclear and it can happen at multiple stages. However, recording ethnic origin in primary care and the electronic care record system of Connecting for Health may obviate the need for repeated collection of this information at different stages in the patient’s journey.

Governmental commitment to reducing ethnic differentials in health is not unique to the UK, and neither is the call for better data on ethnic origin. For example, much work has been done on racial and ethnic disparities in health and access to care in the United States, and calls have been made for standardised data stratified by race, ethnic origin, socioeconomic status, and language.9 10 In New Zealand ethnic differences in mortality over two decades have been examined through creative use of data linkage.11 However, here too, the need for better data to support monitoring by ethnic origin is acknowledged. To overcome limitations in data, retrospective record linkage has also been used in Scotland for examining ethnic differences in morbidity and mortality.12 In contrast, major differences in ideology and interpretation of data protection laws exist between EU member states. For example, official resistance to, and restrictions on, collecting data on ethnic origin is seen in France and Germany, and the UK is one of the few European countries that records ethnic origin in national censuses and officially recognises the need for ethnic data for monitoring purposes.

Although the collection of data in England is improving, the data are worthless unless they are used to target need and reduce inequalities. In this respect the NHS can show greater imagination. Datasets such as hospital episode statistics can already be used for monitoring access to inpatient care, outcomes, and general practitioner referrals by ethnic origin, but we have little evidence of this being done comprehensively at a local level. When current initiatives to improve information yield results, the challenge will be for managers, clinicians, commissioners, and providers to use the information to good effect.