Economic analysis of scaling up HIV prevention interventions in Southern India

This thesis explores the impact of scale on the costs of HIV prevention interventions and the contractual relationships between governments and NGOs in two Southern Indian states. Evidence on the resources required and organisational structures for scaling up health interventions efficiently in low income countries is scarce. The Indian National AIDS Control Programme (NACP) is tackling the largest HIV epidemic in the world. Between 1998 and 2004 the NACP, through state level AIDS programmes, scaled up contracting with NGOs delivering priority targeted HIV prevention projects for high risk populations by increasing the number of NGO projects. Standardised data on production costs and qualitative data on transaction costs were collected and analysed from a sample of these NGO projects in two Indian states. Two data sets are used to explore the impact of increasing coverage on costs: economic cost data from 17 case study interventions; and expenditure data from 82 interventions contracted by one of the state programmes. An econometric cost function examines the influence of coverage on costs. Factors affecting the transaction costs of HIV prevention programmes supporting large numbers of contracts with NGOs are identified from document review and semi-structured interviews. Across the interventions there are variations in coverage, and total and average costs (median cost per person reached is US$ 19.21 (range: US$ 10-51 )). Coverage explains over 50% of the variation in total costs. The cost function estimates confirm a non-linear relationship between average cost and coverage. The two states use different contracting models: direct contracting; and contracting a management agency for NGO recruitment, monitoring and evaluation and technical support. The management agency reduces opportunity costs of leakage and poor quality and copes better with large scale contracting than direct contracting. More efficient scaling up could be achieved by increasing the number of interventions delivered by an NGO rather than recruiting new NGOs. Through better understanding costs and transaction costs, the thesis demonstrates that the economies of scale vary with coverage and that governance must adapt with increasing scale of operation.