Abstract
South Africa has some of the most genetically diverse and the most genetically admixed human population groups on the planet. This is due to South Africa’s peculiar history, both social and biological. Nevertheless, people in the country are divided into four ‘population groups’ by which official agencies mean ‘demographic groups’ designed to correspond to Apartheid racial categories. These categories are “Black African,” “Coloured,” “Indian/Asian,” and “White”. Although there may be strong normative reasons for us to continue to use racial classifications in monitoring and reporting on issues of equity and health inequalities, here I argue that the use of racial classifications in health is such a blunt and imprecise instrument as to be dangerously misleading except when assessing the health consequences of racism. The racial or demographic population groups to which people in South Africa are sorted are so internally diverse as to constitute many distinct biological population groups whose variation is relevant to therapeutic interventions in different ways. These biological differences between race groups and most especially within race groups is important for the (lack of) predictive ability of race in medicine despite the broad correlation of health outcomes with race in South Africa. Each race group also comprises of different ethnic and cultural groups whose differences in environmental exposures have independent health consequences. This makes race of quite limited use in properly incorporating in analysis other social determinants of health that are not necessarily tied to racism. I argue that both the internal biological variation of race groups in South Africa and the differences in their environmental exposures makes the use of racial classifications inappropriate in clinical and biomedical settings. The only exception to this rule is the use of racial categories in tracking progress on equity and the elimination of racial inequalities.