In Africa the literature specifically linking the state, class and the allocation of health resources is sparse, and the evidential base for health research is inadequate and difficult to interpret. This paper looks at some of the ways in which state, class and health may be related in southern Africa. The region provides useful comparisons because of the starkness of the relationships between class and race and disease patterns and health care in much of the sub-continent; the different types of state and class structure within southern Africa; and the changes in ideology and to some extent health practice which came with the political independence of some of its component parts. Using both historical and contemporary data, it pinpoints the importance of analysing the specific and changing form of the state in the different countries of the region, in order to understand the social determinants of disease and the allocation of health resources, and looks at the significance of class, race, ethnicity and gender in the incidence of health and the state's response. It highlights the specific colonial legacies, continuing imperial linkages and location of countries in the international division of labour which inhibit changes in health care. Within the region, the migrant labour system and South Africa's aggressive policies of destabilisation create particular problems for weak states and for individuals within them attempting to implement more progressive health care programmes. The paper also argues that the ideological role played by health care has to be understood, and shows the diverse uses to which it is put across the region. The paper concludes that while the position of the state in the international and regional economy, its specific form and the nature of its class relations are predictors in some sense of health and health care, a variety of micro-level political and social decisions and mediations have also to be taken into account. While most of the countries of the region are in some sense part of the 'periphery', and a product of colonialism, these labels are insufficient to explain the differences between them in terms of disease patterns and health care systems. The specificities of internal social dynamics, local class ethnic and gender struggles and political conflicts are also crucial.