There is strong evidence from developing countries to support the hypothesis that physical accessibility of health services, particularly absolute distance from clinics, is a major determinant of health service utilization and health status. In South Africa, such evidence is very limited and as a result the relationship between absolute distance and health service utilization and health status is not fully understood. As an attempt to understand this relationship, a household survey of mothers with children aged 12-23 months was conducted in a rural district of KwaZulu-Natal province, South Africa. Maternal and child health (MCH) service utilisation and MCH status patterns were then compared at different absolute distances from PHC clinics. The find ings reveal that the study population is characterised by impoverished living conditions (86%), high functional illiteracy (67%), high fertility and unemployment rates . In comparison with other studies conducted previously in the same population, MCH service utilization rates are high. Based on mean distances of homesteads from PHC clinics in the entire study area before the Clinic Upgrading and Building Programme it has been concluded that the physical accessibility of fixed PHC clinics, when compared with the WHO recommendations, was suboptimal. When this assessment is based on clinic usage patterns, it is found that clinic usage decreased from 86.4% at 0-5 km to 79% at 6-10 km with a dramatic decrease to 37.8% at distances beyond 10 km. This decrease in usage at distances above 5 km translates into a considerable reduction in effective coverage of the target population by PHC clin ic services if it is considered that above 50% of the population live greater than 5 km from these clinics. An assessment of the effect of distance of homesteads from PHC clinics on specific MCH service utilization and MCH status has found very few or no significant differences between mothers and children living at 0-5 km, 6-10 km or > 10 km from these clinics. This observation is consistent even after adjustment for the effects of potential confounding. The fact that distance from clin ics has little or no effect on the indicators of MCH service utilization and MCH status is counter-intuitive. A few explanations can be provided. These include the fact that only 50% of the population, even in one of the most rural parts of South Africa access clinics on foot. Since the traditional assumption has been that this distance effect is a function of straight-line walking distances between homesteads and clinics, Euclidian distances alone may be a poor explanatory variable for health service utilization. Furthermore, if the hypothesis is valid that health status is a function of service utilization, it may also be a poor explanatory variable for health status of community members who are reliant on these services. Secondly, based on data from other sources, there is evidence that there have been steady declines in both mortality and fertility rates in the study population over the past 10-20 years suggesting that client communities are already benefiting quite substantially from health services in general and from MCH services in particular in spite of residual distance barriers. In other words, this distance effect on service utilization and health status may be more evident in populations with much higher background infant, child and maternal mortality rates. Thirdly, it is also possible that distance effect still exists, but that methodological limitations prevented this study from showing this effect. For instance, the fact that people use mobile clinics for some MCH services may have confounded the effect of distance from fixed clinics. It is also possible that people use different facilities for different services even though they are further away, and the assumption that all facilities have equal attraction for clients and that the only determinant of use is distance may be flawed. For example, it is evident from this and from other studies in South Africa that whereas most clients use fixed clinics for vaccinations, deliveries are now increasingly conducted at hospitals. Other methodological issues include the fact that certain health outcomes such as stunting are not an exclusive reflection of health service inputs, but are a function of social and economic determinants. Based on these findings, a number of recommendations are made. / Thesis (Ph.D.)-University of Natal, Durban, 2004.