From evidence to practice: the use of mathematical models to inform HIV programme planning and policy decision making

Abstract

From early in the HIV epidemic, mathematical models have been used to understand patterns of infection and the potential for spread and can be a valuable tool to help inform strategic decisions. This thesis aims to investigate the use of mathematical models to inform HIV programme planning and policy decision making. This is done by examining key mathematical models used for this purpose, generating recommendations to advance the utility of these models, and investigating their use in the policy environment. Quantitative and qualitative methods from epidemiology, political science and social science are used to provide an integrated global health perspective. Mathematical models are first used to investigate the long-term epidemiological implications of different policy decisions for HIV prevention and treatment in the countries most affected by HIV. Next, they are used at the national level in a country application to produce short-term projections of incidence within the population. The results from the second model are used to frame a discussion which arose at the international level regarding its use and formulates recommendations for improved use. Finally, a descriptive multi-case study investigation is conducted in Malawi and Zambia exploring the use of mathematical models in guiding national policy with respect to HIV interventions. A qualitative approach drawing on principles from grounded theory is used and a theoretical framework is developed to guide and provide structure for the investigations. This framework views research utilisation as a spectrum and considers a range of different types of use across this continuum. This chapter describes the use of modelling within the policy environment, the key stakeholders involved, and identifies the barriers, facilitators and conditions for use of modelling to inform programme planning and decision making. Taken together, this thesis progresses from global to local, taking modelling beyond the research arena and into the policy environment.