Neither a behavioral intervention alone nor the same intervention coupled with syndromic management of sexually transmitted infection (STI) was effective in lowering the incidence of HIV in a rural area of Uganda between 1994 and 2000.¹ The prevention strategies had little success in changing risky behavior, but they were associated with reduced incidence of herpes simplex virus type 2 (HSV-2) and syphilis, and with a lowered prevalence of gonorrhea.

In a six-year trial, 18 communities were randomly assigned to receive the single- or two-pronged intervention or, for comparison, routine government health and community development services. The behavioral intervention involved information, education and communication activities meant to increase participants' HIV risk-related knowledge and skills; STI diagnosis and treatment services were delivered by government and private health care workers who were specially trained in syndromic management of STIs. As part of the trial, social marketing of male condoms and voluntary HIV counseling and testing were provided in all 18 communities.

At the start of the trial and at two subsequent points, the researchers conducted a census of selected villages close to the government health facility in each community and then administered laboratory tests and survey questionnaires to residents of those villages. They obtained information about STIs and sexual behavior from men and women who were aged 25 or older, or who were younger but were married or sexually experienced. The median length of follow-up was 3.6 years per person.

The three groups of communities (i.e., the six with the single-pronged intervention, the six with the double-pronged approach and the six controls) were quite similar at baseline. Each had an adult population of roughly 7,000, slightly more than half of whom were women and were aged 20-54. The vast majority of residents of the study communities were sexually experienced, but fewer than one in five had ever used condoms; about two-thirds had had two or more sexual partners. Some 7-10% had had a genital ulcer in the last year, and slightly higher proportions had had one previously.

At baseline, the prevalence of HIV in each group of communities was 9-10%. During the course of follow-up, HIV infections occurred at a rate of 0.7 per 100 person-years at risk. After adjustment for age, sex and baseline HIV prevalence, the incidence of HIV in the communities receiving preventive services was statistically indistinguishable from the incidence in the control communities.

In each group of communities, 28% of 13-29-year-olds had HSV-2 at baseline. During follow-up, new infections occurred at a rate of 2.3 per 100 person-years at risk in communities that received just the behavioral intervention and 3.5-3.6 per 100 in the other two groups of communities. Analyses adjusting for age and sex confirmed that communities with the behavioral intervention had a significantly lower HSV-2 incidence than control communities (rate ratio, 0.7), but that communities with the two-pronged intervention fared no better than controls.

By contrast, the incidence of other STIs was reduced only in communities where both behavioral interventions and STI services were in place. At baseline, about one in four people in each group of communities had serological evidence of past syphilis infection, and slightly more than one in 10 had active syphilis. Subsequently, communities with a two-pronged prevention approach had a marginally lower overall incidence of syphilis than comparison communities (2.1 vs. 2.9 per 100 person-years at risk; rate ratio, 0.8) and a significantly lower incidence of high-titer syphilis, which is related to greater infectivity (0.3 vs. 0.6; rate ratio, 0.6). Data on gonorrhea were limited, but the available evidence indicates a considerably lower prevalence in communities served by both types of interventions than in control communities (0.5% vs. 1.2%; prevalence ratio, 0.3).

Between baseline and the final survey, in each group of communities, the proportion of respondents who had had any casual partners in the previous year decreased, and the proportions reporting ever having used condoms and having used condoms with their last casual partner increased. In all three survey rounds, reports of genital ulcers and vaginal discharge were most common in communities with a two-pronged intervention in place. Adjusted analyses confirmed that in the periods covered by the second and third surveys combined, the prevalence of condom use with the last casual partner and of recent vaginal discharge was elevated in communities with a two-pronged prevention strategy; condom use with the last casual partner also was marginally elevated in communities with a behavioral intervention.

The researchers offer a number of possible explanations for their findings. Most important, they note that substantial shifts toward safer sexual behavior, unrelated to the interventions being assessed, were occurring during the trial period; these changes could have come about simply as a response to the HIV epidemic or as a result of HIV education messages from the government and the media.

Despite the interventions' ineffectiveness in this trial, the researchers contend that similar ones "could work in other regions with a rising incidence of HIV-1 and less health education available." The authors of an editorial accompanying the study echo that conclusion and suggest that "this was the right trial done at the wrong time."² They note that whereas behavioral interventions undertaken too early in an HIV epidemic may have little effect because people do not feel threatened enough to modify their behaviors, those implemented too late may have little impact because people who are motivated to change their behavior likely are already doing so. Nevertheless, they conclude that "the development, piloting, and full-scale evaluation of promising interventions is essential to the success of HIV prevention and should receive greater priority."--D. Hollander

REFERENCES

1. Kamali A et al., Syndromic management of sexually-transmitted infections and behaviour change interventions on transmission of HIV-1 in rural Uganda: a community randomised trial, Lancet, 2003, 361(9358):645-652.

2. Stephenson JM and Cowan FM, Evaluating interventions for HIV prevention in Africa, Lancet, 2003, 361(9358):633-634.