The Body: The Complete HIV/AIDS Resource
Follow Us Follow Us on Facebook Follow Us on Twitter 
Professionals >> Visit The Body PROThe Body en Espanol
  • Email Email
  • Printable Single-Page Print-Friendly
  • Glossary Glossary

Uganda Study Found That Death Reduced HIV Prevalence; Did the Public Take Home the Wrong Message?

April 8, 2005

Summary: Uganda has had a remarkable decline in HIV prevalence, and the question of what caused this decline is controversial. An intensive study of the Rakai region of Uganda from 1994 - 2003 found that much of the decreased prevalence resulted from death of people with HIV. But the incidence of new HIV infections was low throughout this study and did not change greatly, suggesting that the real cause of the success was a large reduction in new infections before the study began. The early data presented at the February 2005 Retroviruses conference also showed increasing use of condoms, and some backsliding on reducing the number of sexual partners. But neither change was big enough to greatly affect the incidence of new infections, at least in the aggregate data across the 50 villages studied. In summary, the big reduction in HIV prevalence occurred because of changes that happened before this study, not those measured within it. Therefore the new information does not contradict reduction in the number of sexual partners as a major cause of Uganda's success.

A major study reported at the Retroviruses conference (Boston, February 22-25, 2005) found that death of people with HIV accounted for most of the large decrease in the prevalence of HIV infection in the population, in a small area of Uganda that was intensively studied from 1994-2003.1 During that period, in the 50 villages studied in the Rakai district, the percentage of men with more than one sexual partner somewhat increased, condom use greatly increased but to well under 100%, and the incidence of new HIV infections remained relatively stable.

The researchers noted, "In summary, declines in Rakai HIV prevalence in the past decade are associated primarily with C and D"1 -- meaning Condom use and Death, extending the "ABC" alphabet of Abstain, Be faithful, or use Condoms, the talking point that represents a much more comprehensive social and national mobilization to prevent HIV in Uganda.

The key point to keep in mind in understanding this study is that throughout the entire study period, the incidence of new HIV infections was relatively stable, mostly under 1.5% per year. But before the study began, the incidence had been far higher, at least in some parts of Uganda (there are apparently no comparable data for the same villages). So the big reduction in new cases had already occurred before the Rakai study started.

This observation also explains the seemingly puzzling finding in the Rakai study that while HIV prevalence went down greatly in adults, it hardly changed in adolescents (it started low and remained low).1 Those who were adolescents during the study were not old enough to have been sexually active in the time of high transmission years before. So adolescents did not have the excess of deaths over new infections that led to the decline in prevalence overall.


The great decline in the percentage of people with HIV, reported by the Rakai study, happened because the number of deaths was far larger than the number of people newly infected. The many deaths in the study period were a dark shadow of an earlier time when the HIV incidence (new-infection rate) was much higher. The key question, then, is what brought the incidence down, before the study began?

It seems unlikely that it was condoms, since at the beginning of the Rakai study, fewer than 10% of the adults interviewed reported using them with their most recent non-marital partner (this percentage increased rapidly during the study). Unless condoms had been used previously and then abandoned, this low percentage could not have accounted for the big change in HIV prevalence.

Could the reduced rate of new infections be explained by the natural history of the epidemic? This seems unlikely because most other countries at a similar point in the epidemic had continuing increases in HIV incidence and prevalence, not the decline that Uganda had.

So behavior change in the period before the study may be the most likely explanation for the reduction in the incidence of new infections then. During the study, the percentage of men age 15-49 reporting only one sexual partner in the last year remained greater than the percentage reporting two or more (this percentage increased somewhat during that time). Meanwhile, condom use greatly increased, giving men an additional opportunity to protect themselves and their partners -- but not enough to show statistically as a further overall reduction of HIV incidence, perhaps because the number of sexual partners also increased, and the median age of first sex declined in the study, especially among men.

What will likely be done now is to break down the data by villages, which probably differ in just when condoms were introduced, and just when the changes toward higher-risk behaviors occurred. Then a statistical analysis could better separate the effect of condom use, vs. number of sexual partners, vs. age of first sexual activity, on the incidence of new HIV infections in each village. Also, the database could be analyzed by each individual -- when did he or she report condom use, report multiple partners, or become HIV positive? This analysis was not presented at the meeting, probably because it had not been done by the time of this "late breaker" report of new information not available in time for the usual process of submission to the conference. This early report had overall rates for the entire study, not broken down by villages -- and the decrease in new infections due to condom use, and increase due to sexual risks, may have partly balanced each other. Once the Rakai study is fully analyzed, it may provide some of the best information available on what works and what does not.

Meanwhile, this study found some backsliding in the "B" part of "ABC" (reducing the number of sexual partner), in the Rakai district; clearly the uptake of condom use was more successful. But the results presented at the Retroviruses conference do not contradict the possibility that an earlier large reduction in the number of sexual partners (in previous years, so it was not included in the data recently reported) was largely responsible for the decline in HIV prevalence that was seen during this study.

Incidentally, we see no blame; the researchers accurately reported their results, and the press correctly reported the researchers. But in the rush of late-breaker presentation and next-day reporting, no one made sure that the message received was accurate. It would be tragic if public misunderstanding led to less emphasis on any effective means of reducing the spread of HIV.

Note: This analysis is a work in progress. AIDS Treatment News is open to considering different views, making corrections if necessary, and perhaps publishing a followup article. Contact us at


  1. MJ Wawer, R Gray, D Serwadda, and others. Declines in HIV prevalence in Uganda: Not as simple as ABC. 12th Conference on Retroviruses and Opportunistic Infections. Boston, February 22-25, 2005 [Abstract #27LB, available at].

ISSN # 1052-4207

Copyright 2005 by John S. James. Permission granted for noncommercial reproduction, provided that our address and phone number are included if more than short quotations are used.

  • Email Email
  • Printable Single-Page Print-Friendly
  • Glossary Glossary

This article was provided by AIDS Treatment News. It is a part of the publication AIDS Treatment News.