Work package 8: Analysis of cost and determinants of sustainability

Overall objective

  • To provide a comprehensive incremental costing of all interventions in WP6 and a cost effectiveness analysis for FSWs, for policy information and exploration of financial sustainability

Specific objectives

  • To estimate total and incremental costs of the programme from the provider perspective
  • To estimate programme cost effectiveness using programme effect among FSWs (as measured in WP7)


The project included a comprehensive costing of all interventions mentioned in WP6 and a cost effectiveness analysis of impact on the vulnerable FSW population.  This analysis will inform policy and enable exploration of the sustainability of the interventions. The total and incremental costs of the programme was estimated from the provider perspective, using data from the situation analysis (WP3) and project accounts. WP3 included a detailed exploration of baseline costs of existing service delivery. This exploratory baseline phase asked; What services are currently being offered? By whom? At what cost? This enabled us to establish what was being added by the intervention packages (WP6) and at what cost. Baseline measures of cost was thus retrospective, while the interventions were costed prospectively.  Prospective costs were collected at each study site using a standardised tool that differentiated between research and delivery costs, allowed for inclusion of economic as well as accounting costs, and enabled sensitivity analysis on key variables. We also standardised, as far as possible, processes for cost collection and categorisation to facilitate comparisons between study sites. Program effect was measured in WP7 using data from health facility statistics and cross-sectional surveys conducted among FSWs.  Measures of programme effect for use in cost-effectiveness analysis included change in the overall utilisation of services and change in FSW behaviour, agency and empowerment, and self-reported STI/RTI and HIV infection.  All assumptions in the economic models were explicit and subject to sensitivity analyses to test their impact on findings.