Race and ethnicity in research on infant mortality.

BACKGROUND AND OBJECTIVES Race and ethnicity are variables frequently used in medical research. However, researchers employ race and ethnicity in different ways and with differing intent. This leads to confusion over the interpretation of racial or ethnic differences. This study sought to determine how race and ethnicity are used in research on infant mortality. METHODS We did a structured literature review of original research related to infant mortality published between January 1995 and June 1996 and indexed in the Core Contents section of MEDLINE. RESULTS The majority of articles (54%) mentioned race and ethnicity. US studies mentioned race or ethnicity more than non-US studies (80% versus 22%). Only one study defined the method used to determine the ethnicity of patients; no study defined race or the methodology used in determining patients' race. Researchers primarily used race and ethnicity to describe study populations. Some racial and ethnic identifiers may have been stigmatizing to the subjects studied. The second most common use of race or ethnicity was as a potential confounder. Only one article discussed racism as a contributing factor in infant mortality. CONCLUSIONS There are several problems and ambiguities in the use of race and ethnicity in clinical research. Researchers who use racial or ethnic categories should do so for specified reasons and adopt clear definitions of the categories used.