Developing spatial models of health service access and utilisation to define health equity in Kenya

Noor, Abdisalan Mohamed (2005). Developing spatial models of health service access and utilisation to define health equity in Kenya. PhD thesis The Open University.



Background: Distance is important in access to health care, in turn a key measure of attainment of Millennium Development Goals. The aim of this thesis was to develop spatial models of access and utilisation of government health services in Kenya.

Methods: High-resolution spatial data on health services, population, transportation, elevation, rivers and gazetted areas have been developed for four study districts. Four theoretical spatial access and utilisation models, based on different distance definitions, were then developed for each district. The assumptions of commonly used access models were assessed using data from health facility-based surveys. High-resolution household data were used to adjust the models for actual use. A test of model-fit was carried out and the 'best-fit' model identified. The potential of scaling-up the best-fit model to the national-level was explored.

Results: Six kilometres was the threshold within which most patients used government health services for fever treatment. Higher-order facilities had larger patients draw. Adjusting the models for competition between facilities increased the mean distances to health services. The model incorporating the transport network and physical barriers to movement, adjusted for competition was found to be the `best-fit' model. The Euclidean model estimated that 82% of the population in the districts lived within commonly used target of 5 km of government health services; 78% when adjusted for competition; while the best-fit model further reduced the estimate to 63%. The models could not be scaled-up to national level due to paucity of appropriate data at this level.

Conclusions: Adjustment of models for competition improves their predictive accuracies. The Euclidean model commonly used to measure access estimates 19% (6 million) more people nationwide than the best-fit model to have access to government health. This has major implications for measurement of health development goals. To redress the situation, more research needs to be done in defining spatial access better by including all the key spatial and aspatial parameters. Ultimately, the use of the best access model at the national level requires the development of more and higher-resolution spatial and empirical data.

Viewing alternatives

Download history


Public Attention

Altmetrics from Altmetric

Number of Citations

Citations from Dimensions

Item Actions