Achieving high-quality care for tuberculosis (TB) remains one of the most important obstacles to ending the global epidemic of TB, the world's leading infectious cause of death. Care cascades have been used to show how well each stage of care is implemented from screening to diagnosis and treatment, and to provide a useful framework for assessing multiple dimensions of quality, including efficiency, effectiveness, and timeliness. However, a major barrier to improving TB case finding, treatment initiation, and cure is a lack of capacity for health systems to collect and utilize routine data to identify determinants of high- and low-quality care and targets for improvement interventions. This training proposal addresses these challenges and identifies methodological gaps for me to address through three scientific aims and three training aims. First, we will measure the accuracy of routine TB surveillance data in Uganda as compared to a reference standard and determine if aggregated routine data or high-fidelity sampling of routine individual-patient data can provide the best operational measure of the quality of TB care in Uganda. Next, we will use the best approach to evaluate the impact of Xpert MTB/RIF, a novel rapid, ultra-sensitive diagnostic test, on case finding and treatment initiation. Finally, we will develop a mathematical model of the potential impact of different quality improvement interventions on the TB care cascade in Uganda. Three training aims are well-matched to these scientific aims and will provide coursework and mentored research experiences to allow me to refine and develop mastery in advanced biostatistics, mathematical modeling, and in how to collaborate with policymakers to solve real-world analytical problems. Our findings will help program decision makers use surveillance data not only to monitor TB, but also to address key gaps in quality care. The resulting inferences will yield insights about the accuracy of surveillance data, measuring and improving the quality of care, and operationalizing these elements for quality improvement ? that go beyond Uganda and have relevance in a wide range of high TB burden, low- income settings. This project will draw upon established research collaborations at the Uganda Tuberculosis Implementation Research Consortium (U-TIRC) between investigators at Yale University, Makerere University, and the Uganda National Tuberculosis and Leprosy Programme (NTLP) who have provided access to the existing data that will be used in this proposal. The proposed research and the collaborative, interdisciplinary training environment at Yale and U-TIRC will allow the applicant to develop expertise in quantitative methods for infectious disease epidemiology. After completing the fellowship, the applicant will be well-positioned to seek career opportunities as an independent researcher working collaboratively between academic, government, and other public health partners.
Low-quality care for tuberculosis (TB) in high-burden settings is a major obstacle to progress in decreasing incidence and increasing cure of this deadly disease. A better understanding of how to use routine TB surveillance data to drive quality improvement is timely and necessary. Through the proposed aims, we will use a care cascade approach to describe quality of TB care and its determinants, measure the impact of a new diagnostic tool, and model the effects of potential quality improvement interventions in the high-burden setting of Uganda.
