Comparative costs and cost-effectiveness of behavioural interventions as part of HIV prevention strategies

Abstract

Behavioural interventions have been widely integrated in HIV/AIDS social marketing prevention strategies and are considered valuable in settings with high levels of risk behaviours and low levels of HIV/AIDS awareness. Despite their widespread application, there is a lack of economic evaluations comparing different behaviour change communication methods. This paper analyses the costs to increase awareness and the cost-effectiveness to influence behaviour change for five interventions in Benin. Cost and cost-effectiveness analyses used economic costs and primary effectiveness data drawn from surveys. Costs were collected for provider inputs required to implement the interventions in 2009 and analysed by 'person reached'. Cost-effectiveness was analysed by 'person reporting systematic condom use'. Sensitivity analyses were performed on all uncertain variables and major assumptions. Cost-per-person reached varies by method, with public outreach events the least costly (US$2.29) and billboards the most costly (US$25.07). Influence on reported behaviour was limited: only three of the five interventions were found to have a significant statistical correlation with reported condom use (i.e. magazines, radio broadcasts, public outreach events). Cost-effectiveness ratios per person reporting systematic condom use resulted in the following ranking: magazines, radio and public outreach events. Sensitivity analyses indicate rankings are insensitive to variation of key parameters although ratios must be interpreted with caution. This analysis suggests that while individual interventions are an attractive use of resources to raise awareness, this may not translate into a cost-effective impact on behaviour change. The study found that the extensive reach of public outreach events did not seem to influence behaviour change as cost-effectively when compared with magazines or radio broadcasts. Behavioural interventions are context-specific and their effectiveness influenced by a multitude of factors. Further analyses using a quasi-experimental design would be useful to programme implementers and policy makers as they face decisions regarding which HIV prevention activities to prioritize.