Evaluating the Effects of Non-Financial Health System Policies on Health Care Access and Quality in Low- and Middle- Income Countries

This thesis consists of four independent chapters contributing to the empirical evidence on health care access and quality in low- and middle-income countries. The first two chapters examine issues of health care access and accessibility in Malawi, while the last two chapters evaluate topics on the quality of health care in South Africa, specifically examining policies which may act to increase the quality of health care supplied. The unifying theme of these chapters is that these system-wide policies do not lend themselves to evaluation through random assignment to treatment. In circumstances where experimental designs are not practicable or pragmatic, alternative methods of evaluation must be pursued. Therefore, all chapters exploit observational data to evaluate the respective health-system level policies examined. Chapter 1 examines distance as a determinant of obstetric health care utilisation in Malawi. Specifically, the study explores how the relationship between distance and health care utilisation may change across levels of distance. When combined with information on geographic population distributions this provides valuable evidence for health infrastructure planning. Additionally, attempts to address possible endogeneity of distance are made and whether the effect of distance differs across sub-groups. The results illustrate that distance continues to be a barrier to obstetric health care utilisation. Chapter 2 evaluates the effect of the availability of maternity waiting homes, a policy devised to overcome the distance barrier women face in accessing obstetric health care services in Malawi. Time and space variation in the construction of maternity waiting homes at health facilities are exploited, estimating various difference-in-difference specifications. The findings suggest minimal effects of maternity waiting homes on maternal health care utilisation or child health outcomes. Chapter 3 analyses the effect of a quality improvement programme implemented in primary health care facilities in South Africa. Specifically, the chapter explores whether the programme exacerbated pre-existing differences in facility quality. The findings suggest the quality improvement programme improved quality across all facilities, but may have increased variation in quality in the short run. Continuing the exploration of means to improve quality, Chapter 4 investigates whether facilities’ quality responds to neighbouring peer facilities. Strategic interactions between health facilities are modelled using both a spatial econometric framework and instrumental variable approach. Despite the absence of material incentives, the results indicate facilities do respond to quality changes among their peer facilities.