An investigation into the extent of uncertainty surrounding estimates of the impact of HIV/AIDS in South Africa
Journal Article
2007
Permanent link to this Item
Authors
Journal Title
South African Journal of Science
Link to Journal
Journal ISSN
Volume Title
Publisher
Academy of Science of South Africa
Publisher
University of Cape Town
Department
Faculty
Series
Abstract
HIV / AIDS statistics have been the source of much controversy in South Africa, but often the extent of uncertainty around these estimates is ignored. There is need for an assessment of the range of uncertainty around often-quoted HIV / AIDS statistics. This analysis determines ranges of uncertainty around the inputs and outputs of the ASSA2002 AIDS and Demographic model of the South African HIV / AIDS epidemic, using a generalized likelihood uncertainty estimation approach. A sample of 500 parameter combinations was drawn by weighting randomly generated parameter combinations by likelihood functions defined on the basis of four South African HIV / AIDS data sets. The estimated number of HIV infections in mid-2005 was 5.1 million (95% prediction interval: 4.2-6.0 million), equivalent to an HIV prevalence rate of 11.1% (9.1-13.1%). Between mid-2004 and mid-2005, the estimated number of new HIV infections was 490 000 (370 000-590 000) and the estimated number of AIDS deaths was 320 000 (270 000-380 000). The posterior mean HIV survival time was estimated to be 11.5 years (95% credibility interval: 10.0-12.9 years), longer than estimated for elsewhere in the developing world. This analysis confirms that South Africa is experiencing a severe HIV / AIDS epidemic, and suggests that HIV / AIDS epidemiology in the country probably differs from that elsewhere in Africa.
Description
Reference:
Johnson, L. F., Dorrington, R. E., & Matthews, A. P. (2007). An investigation into the extent of uncertainty surrounding estimates of the impact of HIV/AIDS in South Africa. South African Journal of Science, 103(3 & 4), 135-140.