Financial Risk Protection, Decomposition and Inequality Analysis of Household Out-of-Pocket Health Payments

Abstract

This research examines equity trends in financing health care through out-of-pocket payments (OOP) using South African Income and Expenditure Surveys for the periods 1995, 2000, 2005-06 and 2010-11. South Africa is interesting to examine for a variety of reasons. In 1994, South Africa removed user charges at public health facilities (clinics) for children aged below six years, pregnant and nursing mothers and the elderly (as long as they were not covered by any medical aid scheme) with the aim of increasing access to public health care facilities. The policy was extended to the entire population in 1996. These initiatives, even though they were targeted at promoting access, were also an effort on the part of policy makers to cushion households against the financial costs associated with the consumption of medical care – something that is likely to influence the distribution of household OOP. Whether, this indeed has been the case remains relatively unknown. Within the scope of the investigation, this thesis tries to answer three broad questions: (i) What is the incidence of catastrophic health care expenditures (CHE) arising from OOP health care financing in South Africa from 1995 to 2011? (ii) What are the factors influencing the incidence of CHE among male and female headed households? and (iii) Who pays for health care in South Africa? In investigating the incidence of catastrophic health expenditure, the research has employed two approaches, which are: the financial burden approach and the income approach – the income approach is derived from the equity measures of public finance where progressivity is the main concern, while the financial burden approach argues that the burden should be equally distributed across all households (see Carrin et al., 2009). Both approaches relate health payments incurred by households to households’ capacity (ability) to pay and not to households’ risks of illness, albeit with different definitions of the capacity (ability) to pay. The research has found that in 1995, around 0.03 percent of households incurred health expenses that are likely to force them to cut back on consumption of other basic needs, while for the years 2000, 2005-06 and 2010-11, the incidence is 0.06 percent, 0.09 percent and 0.07 percent, respectively. Given such a low incidence of CHE, the research evaluated the utilisation of health care facilities by households when confronted with illness. This was only done for the year 1995, as it is only year in which data was collected on the illness status of each household member, whether or not they consulted when ill and where they consulted. The results suggest that a negligible percentage of households did not seek treatment when ill. Of those who consulted, it was found that a relatively higher percentage sought treatment in public health care facilities (0.21 percent) than in private facilities (0.13 percent). Having established the incidence of CHE, the second analysis examined the factors associated with CHE and then decomposed the difference between male-headed and female-headed households to establish whether the gap between the two groups had widened or narrowed. The results suggest that the gender gap in the incidence of CHE narrowed by 0.4 percent between 1995 and 2010-11. This reduction in the gender gap is attributable to education, access to piped water and residing in urban areas. Across the different surveys (as well as over the entire time period) education, having access to piped water and residing in urban areas narrowed the gender gap. These results are consistent with existing evidence documenting the important role played by access to basic amenities, such as water and sanitation, as well as human capital (education), in explaining gendered inequalities in health care. Finally, the research examined the distribution of health payments relative to income, focusing on who incurs OOP for their health care needs to establish OOP concentration and quantify its magnitude. The levels of concentation were compared over time, and decomposed to see if it was possible to attribute changes in social determinants of health to the level of concentration in OOP payments for health care. In general, health care payments are concentrated among non-poor households, suggesting that there is progressivity in health care financing, at least as it pertains to OOP. Such results are corroborated by the corresponding concentration indices. When the analysis occurs across the 15-year time period from 1995 to 2010-11, the research finds that changing inequalities across age groups, racial groups, education (particularly completion of secondary education), well-being quintiles and type of toilet used, as well as water source for drinking, explained changes in OOP concentration. It was also found that changing elasticities with respect to OOP payments also play a crucial role in explaining differences over time. Overall, most of the changes in OOP payment inequality are attributable to inequality in the social determinants.