On the Modifiable Areal Unit Problem (MAUP) in healthcare accessibility measurement via the two-step floating catchment area (2SFCA) method

The widely used two-step floating catchment area (2SFCA) method for healthcare accessibility measurement might be biased due to the Modifiable Areal Unit Problem (MAUP), which remains unknown. This study calculates and compares healthcare accessibility via the Gaussian-based 2SFCA method at various spatial scales to identify the scale effects. We also compare various spatial aggregation approaches at the township scale to examine the zoning effects, including geometric centroids, population centroids, government residences and a hybrid population-weighted average travel time approach. Taking Lhasa city as the study area, the grid-based travel time estimation method is adopted to reflect the influence of the physical environment. The results reveal significant scale effects in the application of the 2SFCA method. Using population centroids to represent the spatial distribution of population within townships can yield more accurate accessibility results, while the accessibility measured based on geometric centroids might be considerably biased. The population-weighted average travel time approach, which aggregates grid-scale travel times to the township scale, also ensures high accuracy of accessibility measurement. The population-weighted average travel time approach and the population centroid approach are more effective in mitigating potential scale effects on accessibility results with limited population data. These findings are valuable for researchers to apply accessibility methods properly and support reliable policy implications based on accessibility results.