Improving tuberculosis detection and prevention Insights for case-finding programs, diagnostic, and vaccine evaluation

Tuberculosis remains the leading infectious cause of death worldwide, causing an estimated 10 million illnesses and 1.3 million deaths each year. Current tools and strategies are not reducing incidence rapidly enough to meet the End TB Strategy targets of a 90% decline in incidence and a 95% reduction in mortality by 2035. This thesis applies epidemiological and modelling methods to three domains of tuberculosis control (case finding, diagnostics/prognostics, and vaccine evaluation) to identify where greater impact can be achieved.
An evaluation of 47 years of data from a mass chest radiography program in Cape Town shows that active case finding was associated with subsequent declines in tuberculosis notifications, including among unscreened young children, consistent with reduced transmission. Over a century of notifications in the city reveals strong seasonality, greatest in children, pointing towards transmission during winter as a key driver and informing when to screen for optimal effect. A transmission model calibrated to India, Vietnam, and the Philippines shows that a near point-of-care, high specificity screening assay could reduce tuberculosis mortality as much as a high sensitivity test, at lower cost. In young children, a quantitative QuantiFERON value above 4.0 IU/mL, ten times the standard threshold, identified those at very high risk of progression to disease, which could help target preventive therapy efforts. Finally, simulated vaccine trials show that “all-or-none” and “leaky” vaccines yield similar, unbiased efficacy estimates, and that trials in very-high-incidence populations enrolling both infected and uninfected individuals could be adequately powered with far fewer participants, potentially accelerating vaccine development. Together, these findings offer practical guidance for accelerating progression towards tuberculosis elimination.