6. Discussion
Despite abundant and detailed hospital admission and discharge data available, only a very limited number of studies are related to the HSA delineation. The goal of this study is to develop an efficient spatial algorithm to produce as many eligible selfcontained HSA units as possible within a given area. The TTHA meets this criterion by producing an apparently larger number of HSAs than traditional methods. The skeleton of the algorithm is primarily composed of iterative spatial computation and spatial neighbor searching, which have been well supported and facilitated by Geographic Information Systems (GIS) and could be easily automated. In addition, a select review of literature on the delineation of HSAs and other types of service areas is presented, which serves as an important step forward to enable the adaptation of more traditional classical methods to healthcare facility catchment area delineation. An eligible self-contained HSA unit was defined as an area with LI 0.5, which followed the same setting of LI in previous studies for an easy comparison. When the threshold of LI changes from 50% to a higher level, such as 70% or 80%, delineation was expected to change as well; however, although more research is warranted, on the basis of the nature of the algorithm and results from this study, it can be reasonably hypothesized that the TTHA would still outperform the traditional method regardless of threshold.