with the rising prevalence of AIDS. In this scenario, AIDS would not add significantly to all the other reasons for uncertainty or doubt already endemic to the continent.
Most of the evidence produced so far on the impact of the epidemic at the household level has been of extremely variable quality and often anecdotal in nature. In one of the few field studies of the economic effects of AIDS in Africa, Barnett and Blaikie (1992) describe the results of their fieldwork undertaken in 1989 on a sample of 129 households in the Rakai district of Uganda. The authors were able to provide many rich anecdotes of how the 20 households in their sample might have been affected either directly or indirectly by AIDS, including a reduction in the production of food crops, a gradual depletion of household assets, a withdrawal of children from school, and an increase in household malnutrition. However, they were unable to show that the epidemic had affected producer-consumer ratios in these households, or indeed that any of the supposed effects of AIDS were suffered more frequently or to a greater degree by the 20 AIDS-affected households than by other households. Furthermore, there was no discernible impact on total agricultural production in the Rakai district. The authors (Barnett and Blaikie, 1992:102) conclude that:
… by 1989/90, AIDS had not yet drawn adaptive responses in production and consumption on a scale that dwarfed the many other adaptations households make all the time in response to other rapid processes of socioeconomic change. However, we believe that in certain localized areas AIDS is beginning to be the major determinant of socioeconomic change.
The ability of a household to cope with an AIDS illness and death is clearly a function of many factors, including the socioeconomic characteristics of the household, the economic role of the person with AIDS within the household (particularly how his/her illness affects household income), the household's access to alternative sources of income or support, the level of social and material support available to the household, and so forth. It is analytically convenient to divide the costs to the household of incurring a case of AIDS into three components: (1) direct costs associated with medical expenses; (2) indirect costs to the household directly afflicted with AIDS in terms of forgone earnings; and (3) indirect costs to other households, associated with contributing to funeral expenses or caring for orphaned children (Ainsworth and Rwegarulira, 1992).
Because AIDS manifests itself in a series of other diseases, the direct costs incurred by people with AIDS in seeking medical attention prior to their death can be considerable. The average cost of health care per HIV-seropositive patient admitted to Mama Yemo Hospital in Kinshasa, Zaire, was US $170, compared with US $110 per HIV-seronegative patient (Hassig et al., 1990). The direct costs of medical treatment of AIDS in Tanzania have been estimated at between US $104 and US $631 per person (Over et al., 1988). More recent estimates from