Measuring access to health facilities: does spatial resolution matter?

Background: Access to health facilities is essential to the well-being for all kinds of people worldwide. Even a relatively small decrease in the distance to travel to health facilities can significantly reduce child mortality rates. The accessibility to healthcare can be assessed by the estimation of travel time to the nearest health facilities. This kind of geographic estimation is mainly based on building the cost surface of travel. With the availability of finer resolution land cover data, the creation of the cost surface at more accurate levels comes into existence along with more detailed travel time estimation. However, the finer resolutions also come with computing and storage problems. Methods: This study reclassified land cover data of four different resolutions (20m, 100m, 300m and 500m), and used Python software to model the cost surfaces of the data. We used the least-cost path algorithm to estimate the travel time across these cost surfaces. Then the results were validated and compared with the previous 1km resolution model to see the differences. Results: On average, the coarser cost surfaces (resolutions at 300m, 500 and 1km) have lower travel time estimation results (which may be overestimating access to health facilities). The estimation results of finer spatial resolution (20m and 100m) cost surfaces are quite similar. Conclusions: Finer resolution ones may provide more accurate results, while the coarser ones may underestimate the travel time for certain areas. The resolution of 100m is a possible balance point for the conflict between accuracy and computing flexibility