[Ethnicity and race as variables in epidemiological research about inequity]

Epidemiology analyzes differences in states of health and disease of populations. Public Policies are established considering inequities associated with ethnicity and race. In this context, the identification of vulnerable groups for concentration of resources is relevant. Nevertheless, the lack of a clear definition of these variables might lead to biased results and interpretations. Two problems about the use of these variables are discussed. First, lack of a measurable and objective characteristic, even considering self reference (gold standard), considering that the opinion of a person can change in time. The second problem is a consequence of the former, basing research on a poorly defined variable. Uses of ethnicity and race variables between 1920-1999 in the American Journal of Epidemiology, Health Services Research and American Journal of Public Health were reviewed. In 919 articles, 27 different names identified to describe these variables and more than half did not describe the reason to use these variables. Almost half did not describe analytical methods. Although some articles found statistically significant relationships, less than half discussed those results. It has been suggested that there is enough evidence to exclude these variables in biomedical investigations. However, others propose that they cannot be excluded, given their multidimensional condition that includes social, cultural and genetic features. Therefore, provided the lack of clear definition, the assessment of ethnicity and race effects must be done as rigorously as possible.