This paper aims to analyze the use of the variables race, color and ethnicity in epidemiologic studies, carried out with Brazilian populations. This is a systematic review, conducted in the PubMed bibliographic database, on papers published between January 2000 and July 2010. A data extraction form was used to obtain data from all individual studies, such as their objectives, the relevance of the racial/ethnic classification in their analyses, participants' socio-demographic characteristics, including aspects related to the methods of racial classification, as well as the adherence to a set of recommendations on the use of race, color and ethnicity in biomedical publications. After initially identifying 1,174 references, 151 were included in the review. Higher proportions of each of the following results were observed among papers in which the racial/ethnic classification was central to their analyses - of these, 18% justified the use of racial/ethnic categories; 16% regarded racial/ethnic classifications as context-dependent and fluid; 65% described the methods adopted for racial/ethnic classification; 17% took the racial/ethnic classification as a proxy for genetic variation; 26% considered such classification as a risk factor for health outcomes; 47% considered socio-economic factors in the interpretation of racial/ethnic inequalities in health; and 27% adjusted these racial/ethnic disparities for socio-economic factors in their statistical models. Only two studies elucidated the concept underlying the use of race, color or ethnicity. An expressive amount of the reviewed epidemiologic studies does not follow minimum established criteria on the use of variables regarding racial/ethnic classification, such that this should be urgently improved in public health research.
race and health; ethical and race relations; epidemiological measures
