Medicine (RMH) - Theses
Permanent URI for this collection
Search Results
1 - 1 of 1
-
ItemMathematical modelling for programmatic responses to tuberculosis in the Asia-Pacific( 2016)Despite being a treatable disease, tuberculosis (TB) has not yet been controlled globally, and the burden in Australia’s region remains huge, with two Asia-Pacific countries, India and China, constituting one third of cases worldwide. Moreover, several highly endemic regional hotspots exist and multidrug-resistant TB (MDR-TB) threatens to derail control efforts. The epidemic in the Asia-Pacific region differs markedly from other parts of the world, as transmission is not primarily driven by HIV-coinfection as it is in sub-Saharan Africa, nor by transmission in conjugate settings (such as prisons and hospitals) as in the former Soviet states. Mathematical modelling can help to understand the reasons for our failure to achieve control and to better direct programmatic resources. This thesis first presents the construction of a dynamic ten-compartment model to simulate TB transmission in highly endemic regions of the Asia-Pacific and describes its general characteristics. Findings include the importance of reinfection during late latency, the contribution of community transmission to MDR-TB burden, the importance of properly addressing MDR-TB despite a possible fitness cost and the need to model partially protective vaccines. Next, strategies for TB control were modelled in Western Province, Papua New Guinea, a region characterised by high burden, high proportions of MDR-TB and poor-quality programmatic data. After calibrating to local conditions, the model was used to simulate five programmatic responses to TB and Bayesian inference was employed to explore the uncertainty range around plausible outcomes. The model was then used as the basis for participation in an international collaboration to consider whether the post-2015 Sustainable Development Goals (SDGs) for TB are achievable with current tools. Eleven leading global modelling groups were invited to participate in the multi-modelling exercise to simulate six “ambitious but feasible” interventions for India, China and South Africa. The previously developed model was elaborated to incorporate smear-negative and extrapulmonary TB, initial default and misdiagnosis of MDR-TB, and was then calibrated to each country to capture local TB dynamics before applying interventions. Key conclusions include the small impacts of improved care quality and molecular diagnostics but greater improvements resulting from expansion of access to care, and that active case finding may significantly reduce disease burden. As most targets were not met under the modelled scenarios, future technological advances, such as new treatments and vaccines, are likely to be required to achieve the ambitious rates of decline envisaged in the post-2015 agenda. To improve understanding of TB dynamics, this thesis explores the latent period between infection and active disease. Using Victorian TB Program data, individuals recently infected with M. tuberculosis were linked to subsequently notified active cases, and the resulting dataset was used to perform a survival analysis on the outcome of progression to active disease. I then imputed censorship to account for effective loss to follow-up through death, migration out of the surveillance region and preventive treatment. Results show the five-year risk of disease is 11-18%, several-fold higher than commonly accepted estimates. These revised estimates have important implications for programmatic responses, individual patients and structuring and parameterising compartmental models.