Cost-Effectiveness of HIV Messages Varies By Medium in Benin
In a 2009 cost-effectiveness analysis of various behavior change communication (BCC) methods directed at the prevention of HIV transmission among youths, truck drivers and sex workers in Benin, the most cost-effective methods of promoting consistent condom use were disseminating safer-sex messages via magazines, radio broadcasts and public outreach events (approximately US$22, $25 and $31, respectively, per consistent user).1 Individuals with high exposure to messages through each of these three methods were more likely than those with little or no exposure to report consistent condom use (odds ratios, 1.4–1.9). Two other methods—billboards and peer education—not only had high per-person dissemination costs, but were not associated with increased condom use among exposed respondents (and hence were not cost-effective) in this setting.
To address the general lack of economic evaluations of BCC methods in interventions to prevent HIV transmission, this study assessed the costs of five methods and their cost-effectiveness at promoting consistent condom use. Understanding the effectiveness of interventions is particularly needed in low- and middle-income countries, as well as among populations with low HIV awareness and high levels of risky sexual behavior. Although the HIV prevalence in Benin is relatively low (2%), in the last decade the proportion of adults who are infected has increased 12-fold. In 2007–2009, as part of the country’s National Framework to Fight HIV/AIDS, Population Services International implemented interventions to promote safer sexual behavior and consistent condom use.
An analysis estimated the economic costs of implementing the programs; implementation expenditures were collected for 2009, and start-up and capital costs incurred in earlier years were annualized and converted to 2009 prices. To assess the cost-effectiveness of methods, 5,451 randomly selected individuals at risk of HIV infection—youths (aged 15–24), truck drivers (15 or older) and sex workers (aged 15–29)—were asked if they had heard messages about how to prevent HIV in the last 12 months, and if so, where they had heard them; depending on how many times respondents had been exposed to each type of communication, they were classified into low- (≤1), medium- (2–3) and high-exposure (≥4) groups. Respondents were then asked whether they had consistently used condoms with recent partners (for sex workers, with all partners in the last week; for truck drivers, with all casual partners and sex workers in the last month; and for youth, with all casual partners in the last year). For each medium, the impact of BCC messages was estimated by calculating the odds that someone with high exposure reported consistent condom use relative to someone with little or no exposure; cost-effectiveness ratios were then calculated per high-exposure individual reporting consistent condom use.
The economic cost for all BCC interventions was nearly US$800,000 in 2009. The costliest intervention was peer education (one-to-one and small-group discussions), followed closely by radio broadcasts (30- second messages and talk shows) and public outreach events (e.g., theatrical sketches and condom demonstrations); the least costly interventions were messages conveyed via youth-oriented magazines and billboards. Overall, personnel expenses accounted for 64% of the annual cost of the interventions, while recurrent contracted services and supplies made up 8% and 7%, respectively; the remaining costs were capital and operation expenditures. The numbers of individuals reached by each method over the year varied widely: About 2,600 viewed billboards, 5,000 participated in peer education, 8,000 were exposed to messages in magazines, 43,000 heard radio broadcasts and 82,000 attended public outreach events. The cost to reach each person also varied by method, from just over US$2 for public outreach and $4.50 for radio to $18 for magazines, $25 for billboards and $39 for peer education.
In an analysis assessing the likelihood that exposure to a given method was associated with reporting consistent condom use, individuals with high exposure to HIV messages in magazines were more likely than those with low exposure to have used condoms consistently (odds ratio, 1.9); similarly, individuals classified as having high exposure to radio broadcasts or public outreach events had increased odds of reporting consistent use compared with their low-exposure counterparts (1.4 and 1.5, respectively). No associations were found for exposure to billboards or peer education. Magazines were found to be the most cost-effective approach; they cost about US$22 per consistent condom user, compared with $25 and $31 per individual exposed to radio broadcasts or public outreach, respectively. Sensitivity analyses for a variety of variables and assumptions showed no change in relative cost-effectiveness across methods.
The researchers identified several limitations of their study: the inability to determine causal relationships between intervention methods and consistent condom use; possible respondent or social desirability bias because of reliance on self-reports of condom use; and the calculation of combined cost-effectiveness estimates for all three targeted population groups. Yet the investigators believe their findings provide policymakers and program funders with “new information regarding the comparative costs and cost- effectiveness of several BCC methods [that] can help to guide intervention choices particularly in contexts with high levels of risk behaviors and low levels of HIV/AIDS awareness.” In implementing any particular intervention, they recommend that policymakers and program planners develop context-specific outreach efforts, and that future research employ “a quasi-experimental design to evaluate cost-effectiveness by specific target groups in combination with other interventions.”—J. Thomas
1. Hsu J et al., Comparative costs and cost-effectiveness of behavioural interventions as part of HIV prevention strategies, Health Policy and Planning, 2013, 28(1):20–29.