The decline in the prevalence of HIV that occurred in Uganda during the 1990s can be attributed to a reduction in the number of premarital and nonmarital partners and to increased use of condoms with these partners, according to a multifaceted analysis.1 Data from diverse sources, ranging from newspaper articles and donor reports to nationally representative surveys, suggest that Ugandans began changing their sexual behavior in the late 1980s, as the incidence of HIV peaked, and then began using condoms a few years later, accelerating the decline in transmission.

Because an understanding of the factors that led to the epidemic's decline in Uganda may provide guidance for anti-HIV programs elsewhere, several studies have examined the events in that country. Although these studies concluded that behavioral changes played an important role in the decline, they have not yielded a consensus regarding the nature and timing of these changes. To shed further light on the matter, the researcher who devised the current analysis sought evidence from numerous, diverse sources, including models of HIV incidence and prevalence in Uganda; newspaper articles and surveys about sexual behavior; reports on donated condom shipments; and documents from AIDS programs in Uganda.

The models were crucial because nationally representative data on trends in HIV incidence and prevalence in Uganda are not available for the first two decades of the epidemic. However, models of data from sources such as women seeking antenatal care or voluntary HIV tests suggest that the incidence of infection in Kampala (Uganda's capital) rose steadily for most of the 1980s, peaked around 1987, declined slowly for about five years, and then fell rapidly, returning to circa-1982 levels by 1993. Few data are available for other areas of Uganda, although models for several major Ugandan towns suggest that the incidence peaked a few years after it did in Kampala. Because of the time lag commonly seen between incidence and prevalence, the pattern for the latter is slightly different: Prevalence rose rapidly in Kampala from the early 1980s until 1987 and then increased at a slower pace until 1992; the prevalence then declined rapidly for several years before beginning a slower descent.

Survey data documenting sexual behavior around the time that incidence peaked is not available, but articles in Uganda's primary English-language newspaper indicate that at least some Ugandans were changing their behavior. For example, articles published in 1987 reported that visits to sex workers in a fishing village had greatly declined as visitors and residents became aware that HIV was sexually transmitted; other articles from the period noted that the epidemic had "ruined business" for sex workers in Kampala and elsewhere and that men were "staying home" more. Condom use was generally mentioned only briefly, or not at all, in these articles.

Data from two large surveys—the nationally representative Demographic and Health Surveys, conducted in 1988–1989 and 1995, and the subnationally representative Global Program on AIDS Surveys, conducted in 1989 and 1995—confirm that behavioral changes were taking place by the late 1980s and early 1990s. The 1989 Global Program on AIDS survey found that 9% of respondents had changed their sexual behavior (e.g., by avoiding casual or transactional sex, or being faithful to their spouse), and 60% had made or were intending to make other behavioral changes (e.g., avoiding people with AIDS). In the 1995 Demographic and Health Survey, more than half of never-married adults reported that they had delayed having sex, stopped having sex or limited their number of partners. Condom use remained relatively uncommon, as only 17% of never-married men, and 3% of never-married women, said they had begun to use condoms. Use was much higher, however, among certain subgroups and in certain contexts; for example, 62% of men in urban areas had used a condom the last time they had had sex with a casual partner. Moreover, comparisons between the 1995 surveys and their earlier counterparts revealed decreases in sexual risk behaviors. The proportion of women aged 15–54 who had had sex in the past year (regardless of marital status) declined from 82% to 75%, and the proportion of men who had had extramarital sex declined from 23% to 16%.

Data on condom shipments to Uganda (which manufactured no condoms of its own during the 1980s and 1990s) and documents from AIDS programs support the idea that condom use did not become common until after the incidence of HIV had already begun to fall. In 1989, about 15 million condoms were shipped to Uganda, but until then the number had never exceeded a few million per year. Shipments declined during the next few years, but reached 20 million in 1993 and generally exceeded that level for the rest of the decade. Moreover, documents reveal that the Ugandan government's reaction to the AIDS crisis, beginning in 1986, was to emphasize messages such as "be faithful" and "love carefully"; although condoms were sometimes mentioned, the public's belief in myths about condoms hindered widespread use. Not until the early 1990s, several years after behavioral changes had taken place, did the government, the media and religious groups make concerted efforts to promote condom use.

Although none of the evidence sources used in the analysis is ideal, the strengths of one often offset the limitations of another, the researcher notes. For example, demographic surveys may examine large, representative samples, but they cannot pinpoint the timing of behavioral changes; conversely, newspaper articles can identify the timing of events, but the persons involved may not be representative of the general population. Together, these sources suggest that the incidence of HIV peaked around 1987, at which time behavioral changes began to take hold, and the subsequent decline in incidence was accelerated when condom use became common. Considered in tandem with analyses of data from elsewhere in Sub-Saharan Africa, the findings suggest that "giving a strong emphasis to partner reduction while also encouraging condom use (and abstinence) is much more effective" in reducing HIV transmission than promoting condom use or abstinence by themselves, the researcher concludes.

—P. Doskoch


1. Kirby D, Changes in sexual behavior leading to the decline in the prevalence of HIV in Uganda: confirmation from multiple sources of evidence, Sexually Transmitted Infections, 2008, 84(Suppl. II):ii35–ii41.