Economic evaluation of a community-based intervention : measuring the value to society

Community-based participatory interventions such as those promoting health-related behaviour change present many challenges to the measurement of benefits in economic evaluation. Such interventions can influence social as well as physical well-being, and entail a multitude of outcomes both directly for intervention participants as well as indirectly for other members of the community. Furthermore, they may not achieve immediate changes in health status but their interactive and participatory nature could mean that non-health benefits are significant. The aim of this thesis is to measure the benefits of a community-based participatory intervention in rural Nepal and draw lessons more broadly for how such interventions can be valued within economic evaluation. A contingent valuation survey of a women's group intervention designed to improve maternal and newborn health was carried out in rural Nepal. Members of eleven women's groups were interviewed along with a sample of female non-members and males from the same communities. Monetary and non-monetary measures were used to elicit preferences and respondents were asked which aspects ofthe intervention they were willing to pay for: health outcomes, non-health outcomes or both. Focus group discussions were used to both inform the design of a locally relevant survey tool, to maximise content validity, and as a means of supplementing survey data with group-based discourse. Construct validity was assessed by testing for the association between willingness-to-pay and indicators of demographic and socio-economic status. Willingness-to-pay values were aggregated and combined with costs as a cost-benefit analysis. These results were compared to those of a cost-effectiveness analysis. A sensitivity analysis was conducted to assess the impact of different assumptions on total aggregate willingness-to-pay and the net benefit of the intervention. The response rate was high in all stakeholder groups indicating that the survey was well understood and acceptable to respondents. The study found that non-health benefits were valued by over 80% of respondents and thus their omission would lead to the undervaluation of such programmes. There was no significant difference between the willingness-to-pay of women's group members and non-members, suggesting that the programme generates positive externalities. Focus groups helped to improve the content validity of the survey and to achieve high response rates by enabling questions in the survey to be framed in a manner more relevant to the community. They also gave insight into the valuation context, helped to interpret values derived from the survey and highlighted the importance of trust in the payment vehicle. This thesis also shed light on some of the empirical challenges that are faced when attempting to extrapolate sample values to a larger population and deciding whose values to include.