Although a five-country HIV and STI prevention intervention trial showed no overall reductions in STI acquisition or unprotected extramarital sex, a subanalysis of data from the China site suggests that the intervention's efficacy may have been masked by the inclusion of many participants with low-to-moderate behavioral risks.1 When outcomes in the community-based, randomized controlled trial were examined according to participants' STI history, those who had had an STI at baseline were less likely to have an STI at 24 months if they had been in the intervention rather than control group (odds ratio, 0.5). Similarly, the proportion of these participants who reported having had unprotected extramarital sex in the past three months dropped from 11% at baseline to 5% at 24 months in the intervention group, but did not decline in the control group.
The National Institute of Mental Health Collaborative HIV/STD Prevention Trial, conducted in 2002–2006 in China, India, Peru, Russia and Zimbabwe, sought to reduce HIV and STI risk by changing social norms for safer sex behaviors. Community leaders attended four training sessions in which they were taught skills for delivering theory-based HIV and STI prevention messages to their peers during casual conversations. During the training period, the community leaders practiced delivering their messages daily and gave weekly reports on the frequency of their conversations. After training, the leaders attended sessions bimonthly to support their messaging activities and to report on their ?efforts.
In China, food markets were identified as the venue for the trial, and migrant vendors were chosen to serve as the study population. Forty markets in a coastal city were selected, 20 of which were randomly assigned to receive the community leader intervention. Market vendors were eligible to participate if they were 18–49 years old and had engaged in unprotected sex in the past three months; half of the 3,912 market workers were assigned to the intervention group and half to the control group. About 20% of vendors were invited to serve as opinion leaders and underwent the training sessions; however, the intervention was not implemented in the control group until after the study's 24-month ?follow-up, so that researchers could assess the effectiveness of the intervention. Participants in both groups received STI diagnostic and treatment services, HIV and STI educational materials and presentations, access to condoms and regular monitoring.
Assessments were completed at baseline, 12 months and 24 months, and included an interview, a physical exam by a doctor, and STI testing of blood, urine and vaginal swabs. The primary biological outcome was any new STI (chlamydia, gonorrhea, syphilis, herpes, HIV or, among females, trichomonas), and the primary behavioral outcome was having had any unprotected extramarital sex in the previous three months. Differences between the intervention and control groups at baseline were tested using chi-square, t or Wilcoxon rank tests; the effectiveness of the intervention was examined using mixed-effects regression models.
More than half (55%) of the participants were female, and the mean age was 35 for men and 36 for women. Only 13% of the market workers had a high school education or higher, and 9% had no education. Ninety-seven percent of women and 87% of men were married. Eighty-seven percent of participants regularly earned money.
At baseline, about 7% of participants reported having had unprotected extramarital sex in the previous three months—12% of men and 2% of women. In addition, 20% had an STI—14% of men and 24% of women. During the trial, the intervention and control groups both had significant decreases in STIs and unprotected extramarital sex. At 24 months, women in the intervention group had a lower STI rate than those in the control group (6% vs. 8%), but levels of unprotected extramarital sex were similar in the two groups (<1%); neither outcome differed by intervention group among men.
Further analysis revealed that men, older participants and those who had not previously had an STI were less likely to acquire an STI by the 12- or 24-month follow-up than were women, younger participants and those who had had an STI at any time point, respectively. Next, researchers examined outcomes by STI history. In both the intervention and control groups, market workers who had not had an STI at baseline were unlikely to acquire one during the study period (<5% for each), and outcomes did not differ by intervention status. However, among high-risk participants—those who had had an STI at study entry—17% of those in the intervention group acquired a new STI by 24 months, compared with 30% of those in the control group. After adjustment for age and gender, the odds that a participant with an STI at baseline received a new STI diagnosis at the 24-month follow-up were significantly lower in the intervention group than in the control group (odds ratio, 0.5). Among women who had had an STI, those in the intervention group had reduced odds of acquiring any new STI (0.4) or any bacterial STI (0.5).
A similar trend was noted for unprotected extramarital sex. For the full sample, the decline in levels of unprotected extramarital sex was similar in the intervention and control groups. However, outcomes differed substantially by participants' STI history. Among those who had had no STIs at baseline, the percentage who reported having had unprotected extramarital sex in the past three months dropped in both the intervention group (from 6% to 4%) and the control group (from 7% to 4%). Among participants who had had an STI at baseline, the percentage reporting unprotected extramarital sex dropped from 11% to 5% in the intervention group, but did not significantly change among control group participants (7% at baseline and 6% at 24 months).
The researchers note that despite the negative findings of the larger study, their subanalysis shows that the intervention reduced STI incidence and behavioral risk among the community members at highest risk. Unlike most efficacy trials, which focus on high-risk individuals, the current study focused on high-risk communities; however, the large proportion of participants at low-to-moderate risk masked the intervention's effects. The researchers conclude by noting that although "community-level interventions are likely to be most relevant when all members of a population are in need of intervention," in general, significant intervention effects "can only be expected among the highest risk participants." They suggest that when evaluating the efficacy of interventions, future researchers consider the differing levels of risk within samples.—L. Melhado
1. Rotheram-Borus MJ et al., Reductions in sexually transmitted infections associated with popular opinion leaders in China in a randomized controlled trial, Sexually Transmitted Infections, 2011, 87(4):337–343.