Using mathematical modelling to challenge accepted methods and paradigms of tuberculosis control and transmission
Document TypePhD thesis
Access StatusOpen Access
© 2018 Dr. Romain Frederic Corneille Ragonnet
Tuberculosis (TB) represents a major public health issue at the global level. Despite the availability of vaccines and treatments, TB still kills around 1.6 million persons each year due to a combination of unresolved challenges. Firstly, around 40% of diseased individuals are never identified and can therefore not be provided with adequate care. A second substantial challenge is the extremely high prevalence of latent tuberculosis infection (LTBI), which serves as a large reservoir of future disease that is difficult to control. Furthermore, the emergence of drug-resistant TB (DR-TB) has hampered the progress made by TB control in the last decades and required novel strategies to be adopted. Optimal approaches to address these challenges are hampered by substantial knowledge gaps. The lack of a comprehensive epidemiological understanding of TB has also resulted in today’s TB control relying heavily on strong assumptions or preconceived opinions, which are not necessarily supported by evidence. In this thesis, I used mathematical modelling to challenge several of these accepted paradigms. First, this thesis presents a simple model incorporating epidemiological and programmatic characteristics used to quantify the respective contributions of the different pathways leading to DR-TB at re-treatment around the world. This exercise identified failure to detect DR-TB at first presentation as the leading source of DR-TB at re-treatment. This challenges the accepted paradigm that DR-TB results mainly from poor treatment adherence during treatment of drug-susceptible patients. Important geographical heterogeneity was also observed in the results, and so a web-based interface was built to allow the model to be applied immediately to any epidemiological setting. Next, this thesis presents a novel exploration of the relationship between TB incidence and the effectiveness of preventive treatment (PT). Although it is widely accepted that using PT would be less efficient in high-burden settings, the exploration suggests that PT would yield optimal efficiency where TB incidence is as high as 700 new cases/100,000/year. To improve TB modelling methods, this thesis next presents an evaluation of the existing approaches used to simulate the transition from LTBI to active disease. This was done by comparing the reactivation dynamics produced by different model structures to those empirically observed in contacts of infectious TB patients. This exercise demonstrated that two latency compartments are needed to replicate the TB reactivation dynamics in a compartmental model. It further highlighted that the usual cut-off of two or five years used to distinguish late from early latency should be revised to a much shorter duration. Finally, a novel modelling approach combining country-specific social mixing data with time-variant programmatic parameters within a TB agent-based model is presented in this thesis. The newly built tool was used to detail the profile of Mycobacterium tuberculosis (M.tb) transmission and TB burden in the five highest TB burden countries (India, Indonesia, China, the Philippines and Pakistan). Findings include the unexpectedly high contribution of adolescents and young adults to M.tb transmission. This study also provides estimates of the age-specific size of the latent infection pool, along with the age-specific risk that this infection reservoir represents in terms of future disease.
Keywordstuberculosis; mathematical modelling; epidemiology; infectious diseases; preventive medicine
- Click on "Export Reference in RIS Format" and choose "open with... Endnote".
- Click on "Export Reference in RIS Format". Login to Refworks, go to References => Import References
- Medicine (RMH) - Theses