PROJECT SUMMARY Candidate: Dr. Karen du Preez has dedicated herself to a research career focused on improving health care for children in low and middle income countries (LMIC). Her overall career development goal is to become an independent clinician scientist with a comprehensive, interdisciplinary skill set and research network to efficiently answer high priority public health questions, and have the ability to translate research findings into meaningful policy and implementation changes to improve health care for children in South Africa, and globally. This overall goal is supported by three specific training aims: 1) To understand and apply specialized analytical skills to public health research, specifically disease estimation modelling techniques and spatial analytic methods; 2) To strengthen clinical, management and leadership expertise; and 3) To develop the ability to translate research into meaningful policy changes and implementation. Dr. du Preez has the full support of the Desmond Tutu TB Centre, an internationally recognized research center at Stellenbosch University, Cape Town, South Africa and has assembled a mentoring team highly experienced in all facets of pediatric TB research, and actively involved at both national and international level contributing to scientific steering committees and supporting pediatric TB program implementation. Research: The lack of reliable surveillance data for the disease burden and spectrum of pediatric tuberculosis (TB) is a challenge that limits our ability to appropriately manage the disease globally. TB meningitis (TBM) is a severe form of pediatric tuberculosis with very poor outcomes and often consequent lifelong neurological disability if not diagnosed and treated early. Very little childhood TBM surveillance data exist and the overall goal of this research is therefore to determine the burden, incidence and outcomes of childhood TBM at three levels: global, national (South Africa) and sub-national (Cape Town district), whilst identifying opportunities for prevention, earlier diagnosis and treatment. Mathematical modelling techniques will be used to estimate the global burden of childhood TBM disease and attributable mortality, by country and WHO region (Specific Aim 1). Routine TB surveillance data from South Africa, spanning 14 years (2004-2017), will be used to determine the spatiotemporal heterogeneity of the national disease burden and treatment outcomes of childhood TBM and identify population-level drivers of high burden locations and unfavorable outcomes (Specific Aim 2). Prospective, complementary surveillance strategies will identify all diagnosed and undiagnosed childhood TBM cases (<15 years) within the Cape Town health district over a 2-year period. All identified children will be eligible for enrolment in an observational cohort study that will collect data on diagnostic certainty, disease severity, comorbidities, outcomes and missed opportunities for both TB preventive therapy and earlier diagnosis, with the aim of informing earlier diagnosis (diagnostic algorithms) and targeted interventions for improved TBM prevention and care (Specific Aim 3).

PROJECT NARRATIVE Despite tuberculosis (TB) being a treatable and preventable disease, recent modeling work estimated that nearly 200,000 children <5 years died of TB during 2015, making it one of the top ten causes of death in children under 5 years globally. Unreliable and incomplete surveillance data for children with TB is a key challenge that currently limits our ability to appropriately manage children with this disease, including those suffering from severe forms such as TB meningitis (TBM) who experiences high mortality and often lifelong disability. This project will use interdisciplinary methods, including modelling and spatial analyses, to provide critical information on the burden, incidence and related mortality of TBM in children at a global, national and sub-national level in South Africa, whilst identifying opportunities to improve prevention and care through in- depth prospective surveillance.