This thesis aims to gain understanding of current methods used to forecast the costs at scale for a new technology in low- and middle-income countries (LMIC) and to propose improved methods for cost predictions at scale, using community-based HIV self-testing (HIVST) kits distribution programmes in southern and western Africa as a case study. Following the review of ~8,000 studies through seven databases on quantitative analyses of costs for informing the scale-up of an intervention in LMIC, I propose a framework to guide the decision process of fitting cost functions by study objective. I then conduct costing studies for implementing community-based HIVST distribution models in southern and western African regions. I also explore potential efficiency gains arising from the addition of HIVST to HIV testing services (HTS) and from continuous programme development in Lesotho, and the importance of distinguishing between full and incremental HIVST costs for country financial planning. I deepen our understanding of the variation of HIVST costs at scale up by fitting an empirical econometric costs function using data from Malawi, Zambia, Zimbabwe, South Africa, and I test it against the observed HIVST scale-up in Lesotho to inform on its external validity. In western Africa, I use an accounting cost function to quantify the expected returns on investment of adding HIVST to civil society organisation-led HTS programmes in Côte d’Ivoire, Senegal and Mali. This research provides insights into the economic considerations for integrating and scaling up the community-based HIVST distribution programmes in southern and western Africa. These findings inform costing studies design for data collection and analysis, encourage the use of cost functions that are the most relevant to the policymaker research questions, ultimately to guide the scale-up of the most promising health interventions in LMIC.