Ethnicity has overtaken race in medical science: MEDLINE-based comparison of trends in the USA and the rest of the world, 1965-2005.

Ethnicity and race are among the most commonly used epidemiological variables, closely following age, sex and social class. Relative increase in the use of the term ‘ethnicity’ rather than ‘race’ in the health literature from 1966 to 2000 has been observed. These terms describe two distinct, overlapping concepts and, arguably, ethnicity is preferable to race. There are limited genetic differences between racial groups, undermining the traditional use of race as an indicator of biological difference between populations. The broader concept of ethnicity emphasizing cultural differences helps to determine aetiology, tackle inequalities, assess need, make public health plans and direct resource allocation. In Europe, race has been largely superseded by ethnicity. Since Afshari and Bhopal’s 2002 article, the compound terms race/ethnicity or race–ethnicity have emerged, particularly in North America. The significance of this is unclear but it may herald a switch from race to ethnicity there, as in Europe. This article presents an update of our earlier work and tests a prediction implied there—that ethnicity would continue its comparatively rapid growth. It also separates North America and the rest of the world. We used the methods already reported. In January 2008, PUBMED/MEDLINE-listed articles for 2001–05 were studied and searched in all fields, publication types and languages. We counted the number of appearances of Ethnic groups, Ethnicity, Race, Racial stocks, Racism and Prejudice. Some variables relevant to ethnicity and race (heart, depression, social and income) were also selected as a reference to the general growth in publications—findings available on request to the authors. Based on publication dates, findings were categorized in 5-year periods. Previously, the number of articles in each category in 1995–2000 was divided by the number in 1966–70 to yield a publication growth ratio. For the updated search, the number of articles during 2001–05 were used divided by the number of articles during 1966–70 (based on the new search, yielding different numbers for earlier years for reasons briefly mentioned below and discussed in a full report available from the authors). We examined race/ethnicity and race–ethnicity for the first time. Finally, data for the USA and the rest of the world were separated (combined data are available in the full report). The ratio of articles on race in relation to ethnicity was calculated to assess the difference in the USA and the rest of the world (Figure 1). Table 1 shows that the number of appearances of Ethnic groups, Ethnicity, Race and Racial stocks were still considerably higher than Racism and Prejudice, but the difference reduced over time (in relative terms). For the four reference search terms (Heart, Depression, Social and Income) the growth ratios ranged from 3.0 to 13.1, which is similar to our 2002 report. Assuming this range as the background pattern, articles on race–ethnicity, race/ ethnicity, racism (Europe) and prejudice (Europe) increased relatively fast. Ethnicity grew faster than race and racial stocks. Race/ethnicity and race– ethnicity were virtually unused until 1986, when race–ethnicity began its rapid growth, with race/ ethnicity following 10 years later. Table 1 shows that ethnicity overtook race in 1991–95 in the USA. The same switch occurred in the rest of the world in 1976–80. Their ratios are shown in Figure 1.