Socio-economic Determinants of Tuberculosis and of Risk Taking Behaviour

This thesis is separated into two distinct parts. The first of these investigates the extent to which tuberculosis morbidity and mortality in the twentieth century were determined by the conditions in which people worked and lived and how institutional responses to the tuberculosis problem shaped individual chances of survival locally, nationally and internationally. We collate several new datasets to investigate socio-economic determinants of tuberculosis disease. Firstly, we use newly sourced data to investigate to what extent, why and to what effect different areas in interwar Britain pursued different policy agendas and how these affected tuberculosis outcomes, both through qualitative discussion and through the application of statistical methods. We argue that where local authorities identified tuberculosis as a socio-economic phenomenon, outcomes were favourable. Secondly, we identify a tripartite divide between developing, newly industrialising and developed countries after the introduction of antimicrobial drugs. We explain why peoples in some parts of the world did not benefit from the promises of prevention and cure enjoyed those in the West. Thirdly, we construct new a data set to investigate the cross country determinants of tuberculosis morbidity and mortality and to observe this tripartite divide, with special focus on the nutritional composition of diet, in the post war period. We find a tuberculosis Kuznets relationship and that adequate nutrition and living conditions most strongly predict tuberculosis outcomes. Finally, we ask whether, in the current era of drug resistant tuberculosis, nutrition science can be utilised to prevent the development of active symptoms if drugs cannot cure the disease. In the second part, individual level data sourced from the World Health Organisation is used to investigate whether background risks to health influence the decision to smoke. A modified double hurdle model of the decision to smoke reveals that background health risks may increase smoking, contrary to theoretical prediction, but that the effect is moderate.