The collection of data on ethnic groupings has become an increasingly perva
sive feature of contemporary health policy and research in the United Kingd
om, with attention concentrating primarily on monitoring access to and util
ization of services by different ethnic groups, together with epidemiologic
al data on morbidity and mortality. At the same time, the collection of dat
a on ethnic populations by census and health agencies has been the subject
of a wide-ranging and contentious debate and there is a growing critique th
at challenges the collection and use of such data on political, methodologi
cal and practical grounds. This paper explores the nature of these debates
as they apply to health research. Issues of validity and reliability arisin
g from the application of pre-defined ethnic categories, such as those used
within the National Health Service derived from the 1991 census, are consi
dered and alternative approaches which utilize a range of variables such as
language, religion and length of residency in a country suggested. Experie
nces derived from an ethnographic study of the provision of district nursin
g care to patients from different ethnic backgrounds are used to illustrate
some of the practical issues of researching ethnic diversity. Strategies f
or addressing some of the methodological and practical concerns are propose
d.