The use of 'racial' categories in contemporary South African health research. A survey of articles published in the South African Medical Journal between 1992 and 1996.

OBJECTIVE: In the light of growing concern about the clinical, social and political impact of 'racial' categorisation in health research, this survey aimed to re-evaluate the current use of 'racial' categories in articles published by the South African Medical Journal. SURVEY
DESIGN: Any categories that might have been used or interpreted as measures of genetically determined 'racial' differences (including 'racial', ethnic and sociopolitical 'population group' categories) were included in this survey of 668 articles describing South African health research published during the past 5 years. By classifying the research contained in each of these articles it was possible to assess the importance of 'racial' categorisation in study design. The explanations given for any 'racial' differences in health were then used to evaluate the impact of 'racial' categorisation on the perception that innate characteristics were responsible.
RESULTS: Three hundred and seventeen (47.5%) articles mentioned one or more 'racial' categories, 292 (43.7%) of which used 'racial' categories to describe the subjects they studied. The commonest generic labels used for these categories were "race" and "population group", while the commonest descriptive labels referred to traditional 'racial' characteristics such as phenotype, nationality and geographical origin. Only 15 (5.1%) articles fully defined the categories and labels they used, and many more used different generic and descriptive labels interchangeably. The use of 'racial' categories was highest among genetic (73.9%), descriptive (55.3%) and quasi-experimental studies (38.2%), although most used these categories simply to describe the subjects they examined. Of those 162 (24.3%) articles that discussed 'racial' differences in health, only 120 (18.0%) contained an explanation, and 60 of these suggested that inherent genetic or behavioural factors were responsible. Nine articles contained derogatory statements that could be interpreted as prejudiced or racist.
CONCLUSION: The use of 'racial' categorisation remains widespread in South African health research. By using generic and descriptive labels from traditional 'racial' taxonomies, many of these articles reinforce the perception that distinct human 'races' exist. Although most of the articles failed to explain any of the 'racial' differences in health they observed, it is likely that these will be interpreted as evidence of innate genetic or behavioural differences, like those suggested by the few articles that offered explanations. The continued use of 'racial' categorisation in health research might be inevitable, particularly for examining the impact of social forces, such as apartheid and other forms of racism, that use 'racial' categories to create unequal access to health and health care. However, any studies that use 'racial' categories should be careful to avoid legitimising the biological concept of 'race', misidentifying the causes of 'racial' disparities in health and reinforcing 'racial' prejudice.