Modelling Spatial Accessibility for Precision Targeting of School-based Interventions: a Case Study of Kilifi County

Abstract

School-based interventions are an effective way to improve health outcomes for school aged children (SAC). However, the effectiveness of these interventions depends on the children’s spatial accessibility to schools. Through modelling of spatial access and accurate identification of the coverage of SAC within and without access to schools at certain travel time thresholds, precise targeting can be done to maximize on allocation of limited resources. Travel time to public primary schools (PPS), as a measure of spatial accessibility, was computed using the AccessMod software alpha version 5.8. The locations of public primary schools were overlaid on top of a friction surface that takes into consideration roads, land cover, and physical barriers including rivers, parks and reserves using the least cost path algorithm. The cost given to each cell is the travelling time to cross the cell, which is determined through the travelling speed attributed to the landcover of the cell. Accessibility raster surfaces based on travel time were generated at both 100m and 1km spatial resolution to match the resolution of the available population raster dataset obtained from WorldPop. To quantify the proportion of SAC within the catchment of the nearest PPS at a recommended threshold of 24 mins, population were intersected with the resulting travel time estimates to extract spatial accessibility coverages of SAC. Spatial access to schools varied across the county ranging from minimum to 804 and 775 minutes at 100m and 1km spatial resolution, respectively. When considering the 24-minute travel time threshold, the population of school-aged children covered at 1km was 302,256 (72.53%) in comparison to 293,031 (70.31%) at 100m spatial resolution. The school catchment areas (SCAs) generated at 1km spatial resolution overestimates the size of SCA by 1.17 km2 by average and number of SAC by an average of 70. There were heterogeneities in coverage of SAC within 24 mins travel time threshold at subcounty level. While 4 subcounties (Rabai, Malindi, Kilifi North and Kilifi South) located in the eastern central and southern regions of the county had more than 70% of the SAC covered, 2 sub-counties (Magarini and Ganze) in the western region of the county had less than 55% coverage. The results of the study can help to inform precise targeting of schoolbased interventions to the SAC and maximize on limited resources.