We read with interest the recent article by McNeil et al. (Med Pediatr Oncol 39:554-557, 2002), which analyzed SEER update data in order to review the incidence and prognosis of childhood acute lymphoblastic leukemia (ALL). We agree with the article's authors and also the authors of the accompanying editorial (Med Pediatr Oncol 39:551-552, 2002) that the article illustrates the value of the SEER program in understanding population trends in cancer generally and, more specifically, ALL.
We were struck, however, by the fact that McNeil et al. used the population characteristics of ''ethnicity'' and ''race'' as though interchangeable concepts. This is exemplified in their Tables I and II: the former lists the subcategories White, Black, and Other under ''Ethnicity,'' while the latter lists the identical three categories as ''Race.'' Moreover, their Table IV, which considers ALL in Hispanics, lists two ethnic subcategories, White and Black.
Confusing ''ethnicity'' and ''race'' can conceal epidemiologic differences of value for elucidating disease causation. ''Race'' implies a mainly biological, albeit multidimensional paradigm (e.g., ''a social category based on the identification of (a) a physical marker transmitted through reproduction and (b) individual, group, and cultural attributes associated with that marker''[1]), while ethnicity encapsulates cultural, behavioral, and environ-