Tuberculosis (TB) remains a major global health problem. The World Health Organization (WHO) has set ambitious targets to eliminate TB by 2035. To achieve this, dramatic reductions in TB incidence must be achieved worldwide. Strategies to achieve these targets include the early detection and treatment of all patients with TB, and preventative treatment to reduce the burden of infection sustaining this epidemic. Intensified research and innovation to develop new methods and approaches are crucial to achieve these elimination targets. The research in this thesis encompasses novel strategies that directly align with the plan for TB elimination set by the WHO. The two main themes of this research revolve around active case finding for TB and diagnostic and prognostic biomarkers for TB infection and disease. The research setting is a large cluster randomised controlled trial (RCT) of community-wide active case finding for TB in an endemic region – Ca Mau Province, Vietnam. This study uses a simple automated polymerase chain reaction test, Xpert MTB/RIF, performed on spontaneously expectorated sputum as the primary screening tool. This test has been endorsed by the WHO and is in widespread use as a diagnostic tool in symptomatic individuals in many TB endemic countries. However, it’s use as a screening tool in the community is novel. As part of the research in this thesis, I assessed the performance of Xpert in the setting of community-wide active case finding. Due to the large number of Xpert tests performed, I was able to precisely estimate the positive predictive value and specificity of this test, and to show that both these are substantially higher than previously estimated (in a Cochrane review and by WHO). I also explore novel strategies to improve the feasibility and cost effectiveness of large scale sputum testing in the setting of TB screening. These strategies are: using sputum quality as a predictor of the presence or absence of Mycobacterium tuberculosis in sputum and pooling sputum samples from two or more individuals to reduce the number of Xpert tests required. My studies show that sputum quality assessments cannot be used to exclude samples from testing. Sputum pooling, on the other hand, could be used to reduce the overall number of samples needed to test without substantially affecting the sensitivity of Xpert. In this research, community-wide active case finding for TB in a large population was conducted annually for four consecutive years. Therefore, we were able to accurately determine the incidence of TB in this region. This is the first study that has directly determined TB incidence on a large scale. I have shown that the incidence of TB is substantially higher than estimated, using inferential methods and imprecise data, by WHO. Lastly, I conducted a case control study to evaluate potential biomarkers for TB using transcriptomic analysis (RNA sequencing) performed on whole blood samples. This was nested within this large RCT. This study identified concise gene signatures that were able to accurately distinguish between early TB disease, TB infection, and healthy controls. Incorporating these gene signatures into a simple point of care test could revolutionise the way we screen individuals for TB and also, detect early disease prior to the onset of symptoms.