With the global scale-up of universal health coverage and a renewed interest in reducing poverty and vulnerability in the context of health, there have been substantial efforts towards including a representation of the economic impact of illness on patients and their households more formally in the evaluation of health care and development of health policy. Programmes increasingly need to track progress against disease-specific global targets, such as the End Tuberculosis (TB) Strategy which identifies a target of “no TB-affected family facing catastrophic cost due to TB” by 2020. There is also an increasing interest in understanding the impact of disease on impoverishment, and the broader dynamics of the interaction between health service delivery, health service expenditures and poverty as part of the economic evaluation of new technologies. This shift towards including the household perspective in health economics research requires consideration of whether the methods and data currently being used are appropriate. There are inconsistencies in current methodological approaches to estimate disease-specific costs from the patient perspective, and the applicability of methods for different study purposes is unclear. This thesis aims to improve the estimation of disease-specific patient, household, and catastrophic costs collected in the context of facility-level intervention-focused studies for different policy purposes. Focusing specifically on the case of TB in South Africa, this thesis evaluates existing methods for estimating catastrophic costs for TB and highlight methodological issues that researchers need to consider when collecting these costs. The research presented in this thesis confirms that people in South Africa continue to encounter catastrophic costs due to TB and provides some indications where investment from the South African government can reduce this burden. The thesis also identifies several important limitations in the current implementation of methods to estimate disease-specific catastrophic costs, and makes a strong argument in favour of a reference case on estimating patient, household, and catastrophic costs.