University of California, San Francisco Logo

University of California, San Francisco | About UCSF | Search UCSF | UCSF Medical Center

Steep declines in population-level AIDS mortality following the introduction of antiretroviral therapy in Addis Ababa, Ethiopia
Global Health Sciences Literature Digest
Published September 7, 2009
Journal Article

Reniers G, Araya T, Davey G, et al: Steep declines in population-level AIDS mortality following the introduction of antiretroviral therapy in Addis Ababa, Ethiopia. AIDS 2009 Feb 20;23(4):511-8.


To estimate trends in adult AIDS mortality and the effects of antiretroviral therapy (ART) in Ethiopia using data from burial sites and verbal autopsies

Study Design

Surveillance of burials at all cemeteries in Addis Ababa, using lay reports of causes of death, was done from 2001 through 2007 to estimate AIDS mortality among persons aged 20-64 years. Reports of cause of death from verbal autopsies in 2001, 2003, and 2007 were used as a reference standard


Cemeteries in Addis Ababa, Ethiopia. Surveillance covered cemeteries associated with all religions


Participants were people between ages 20 and 64 years who died between 2001 and 2007 and were buried in Addis Ababa, Ethiopia


Cemetery clerks collected information on date of burial and demographics and a lay report of cause of death from relatives or close friends. Verbal autopsies were conducted on a random sample of registered burials by trained community health workers who visited the household of the deceased two to four weeks after the death. Causes of death were verified through physician review: two physicians independently assigned an underlying cause of death using ICD-10 coding. Deaths were classified as AIDS or non-AIDS, and tuberculosis (TB) and AIDS were grouped together because of difficulties in distinguishing these as distinct causes of death. Results of verbal autopsies from 2001 and 2007 also were used to validate the estimation method. The projected number of deaths using the Spectrum method was compared to the observed numbers from burial surveillance. Positive-predictive values (PPVs), sensitivity, and specificity for predicting AIDS-related deaths were calculated using the lay reports as compared to verbal autopsies with physician review. The authors used these PPVs and the results from the Spectrum model to estimate the number of averted AIDS deaths as a result of ART roll-out.


Deaths explicitly labeled HIV/AIDS by lay report had a PPV of 89%. Deaths labeled herpes zoster, diarrhea, uterine cancer, mental problems, and emaciation had a PPV of 71%. Lung disease and cough had a PPV 86%, and 40% of all deaths were labeled as such by the lay reporter.

The frequency of most lay-reported AIDS-related deaths declined over time, an overall 50% reduction from 2001 to 2007. However, deaths explicitly attributed to HIV/AIDS by lay report actually increased during this time period. The overall decline in AIDS-related deaths was concentrated in the last two years of the period and was higher in women than in men (38.2% vs. 42.9%). Based on these estimations, the cumulative number of averted adult AIDS deaths between 2003 and 2007 exceeded 18,400.


The authors conclude that AIDS mortality has been reduced by more than half after five years of the ART program in Addis Ababa, Ethiopia. Once ART began being provided for free in 2005, the rate of decline of AIDS-related mortality increased and became balanced between genders.

Quality Rating

Not applicable-this was a modeling study

In Context

The authors note that the increased decline in the later years of the study period (2005-2007) may be due to the free provision of ART.(1) Before treatment was free and a small co-pay was charged, the decline was 15.8%; after free treatment the decline was 40.6%. However, the authors also note that other factors, such as increased HIV testing and increased ART advocacy, may have contributed to the increasing rate of decline. The authors also note that many deaths are occurring just before or shortly after ART initiation and that patients are not showing up for treatment until it is too late.(2,3) Further study is needed to find how best to target these groups.

Programmatic Implications

Because very few of the countries with the highest HIV prevalence and highest AIDS-related mortality rates have vital registration systems, to understand the impact that ART roll-out has had on mortality at the population level, innovative methods for reporting cause of death are needed. Burial surveillance is one such method that is logistically simple and inexpensive and can fill in gaps where verbal autopsy reports are unavailable. Very few low-income countries, however, have burial practices similar to Ethiopia's that would allow for this type of study to occur.


  1. Ivers LC, Kendrick D, Doucette K. Efficacy of antiretroviral therapy programs in resource-poor settings: a meta-analysis of the published literature. Clin Infect Dis 2005 Jul 15;41(2):217-24.
  2. Braitstein P, Brinkhof MW, Dabis F, et al. Mortality of HIV-1-infected patients in the first year of antiretroviral therapy: comparison between low-income and high-income countries. Lancet 2006 Mar 11;367(9513):817-24.
  3. Lawn SD, Myer L, Orrell C, Bekker LG, Wood R. Early mortality among adults accessing a community-based antiretroviral service in South Africa: Implications for programme design. AIDS 2005 Dec 2;19(18):2141-8.