Investigating disparities in health between more and less advantaged groups requires the accurate identification and categorization of those groups. The definitions of race, ethnicity and SES raise measurement issues which researchers in health disparities must consider. With respect to race and ethnicity, measurement strategies have ranged from use of genetic markers to 3rd
party assignment to self-identification. While self-identification is generally considered the gold standard for non-genetic studies 3
, a recent review found that many authors do not indicate the means of identifying the race and ethnicity of subjects in their manuscripts, and that investigators assign race and ethnicity to subjects in a minority of cases.4
How best to categorize race and ethnicity is another area of concern. The inclusion of mixed-race categories and the degree of granularity used to categorize ethnicity (e.g. whether to group all Hispanic/Latinos as one category versus considering Mexican-American, Cuban-American etc., separately) are both topics of active discussion in the literature. The most commonly used categories are those delineated by the Office of Management and Budget (OMB) in 1997, which includes five race categories (Black or African American, White, Asian, American Indian or Alaska Native, and Native Hawaiian or Other Pacific Islander) as well as one ethnicity choice (Hispanic/Latino or non-Hispanic/Latino) 3
. In addition, the OMB allows for the designation of multiple race categories by each individual. A recent report by the Institute of Medicine (IOM) attempted to further clarify reporting of race and ethnicity by calling for use of the OMB categories along with more precise ethnicity categories in accordance with the geographic area in which data collection occurs 5
. It should be noted that the OMB and IOM categories are not universally agreed upon. One area of disagreement is whether to consider race and ethnicity separately, with some arguing that in fact these two categories are overlapping 6
SES presents different, but equally complex, measurement issues. The concept of SES represents a composite of many different factors, including income, education, childhood income level, parental education, and wealth. In disparities research, this complexity is often distilled down to the use of one, or at most two, factors. This is often inadequate, as analyses have shown that conclusions can differ widely depending on which measures of SES are utilized 7
. Ideally, the study of health disparities by SES should incorporate multiple factors, with attention to those which are most relevant for the research question being studied.
An additional consideration in measuring race, ethnicity and SES in the study of health disparities is how to account for the complex ways in which these constructs can interact with each other 8
. For example, being of low SES may impact the health of African Americans differently than Whites, and consideration of these types of nuances must be incorporated into the conceptualization and study of health disparities.