How many people have really died of HIV/AIDS in South Africa?

HIV/AIDS in South Africa: how many people died from the disease between 1997 and 2010?

Bradshaw D, Msemburi W, Dorrington R, Pillay-van Wyk V, Laubscher R, Groenewald P, team SN. AIDS. 2015 Oct 27. [Epub ahead of print]

Objectives: Empirical estimates of the number of HIV/AIDS deaths are important for planning, budgeting, and calibrating models. However, there is an extensive misattribution of HIV/AIDS as an underlying cause-of-death. This study estimates the true numbers of AIDS deaths from South African vital statistics between 1997 and 2010.

Methods: Individual-level cause-of-death data were grouped according to a local burden of disease list and source causes (i.e. causes under which AIDS deaths are misclassified) that recorded a rapid increase. After adjusting for completeness of registration, mortality rate of the source causes, by age and sex, was regressed on lagged HIV prevalence to estimate the rate of increase correlated with HIV. Background trends in the source-cause mortality rates were estimated from the trend experienced among 75-84 year olds.

Results: Of 214 causes considered, 19 were identified as potential sources for cause misattribution. High proportions of deaths from tuberculosis, lower respiratory infections (mostly pneumonia), diarrhoeal diseases, and ill-defined natural causes were estimated to be HIV-related, with only 7% of the estimated AIDS deaths being recorded as HIV. Estimated HIV/AIDS deaths increased rapidly, then reversed after 2006, totalling 2.8 million deaths over the whole period. The number was lower than model estimates from UNAIDS and the Global Burden of Disease Study.

Conclusion: Empirically based estimates confirm the considerable loss of life from HIV/AIDS and should be used for calibrating models of the AIDS epidemic which generally appear too low for infants but too high for other ages. Doctors are urged to specify HIV on death notifications to provide reliable cause-of-death statistics.

Abstract access  [1]

Editor’s notesIn many countries, the true number of HIV-associated deaths is significantly under-reported in national vital registration data making it difficult to monitor the epidemic trends from this source. This study describes new estimates of HIV-associated mortality based on empirical vital registration data which aimed to provide accurate estimates of the numbers of HIV-associated deaths in South Africa. The study estimates that, from 1997-2010, 2.86 million deaths in South Africa were due to HIV – over one-third of all deaths. However, relatively few deaths, 7%, were registered as HIV-associated. At the peak of the epidemic in 2006 the vital registration derived estimates show lower trends than other models. All models estimated a decline in the number of HIV-associated deaths post-2008, a finding which is consistent with the extensive roll-out of antiretroviral therapy in South Africa, and with trends reported from verbal autopsy data for all deaths in rural South African demographic surveillance sites. This paper highlights the importance of reporting accurate causes for HIV-associated deaths in the death registration process - however, without de-stigmatisation of HIV, this is going to be difficult to achieve.

Africa [9]
South Africa [10]
  • [11]