Using national Hospital Bed Function Reports for fiscal year 2024, the authors identified facilities reporting reimbursable intensive and high care beds and aggregated counts to Japan's 330 secondary medical areas (SMAs). In 2024 there were 20,631 reported critical care beds across 1,044 facilities, subdivided into 7,114 ICU-type and 13,517 HCU-type beds. Among SMAs that had any ICU/HCU beds, zone-level totals ranged from 3 to 688 beds, a 229-fold difference. The 90th/10th percentile ratio of per-capita bed density was 3.6, illustrating substantial variation in local supply density.
The authors emphasize the compositional shift toward intermediate critical care: total capacity rose by +27.9% over the observation period, but this increase was driven primarily by HCU beds (+41.6%), whereas ICU-type beds grew by only +8.0%.
MeshScope-Region integrated a road-network accessibility layer computed with the Open Source Routing Machine (OSRM). Travel-time calculations were performed from 176,962 populated 1-km census grid cells to the nearest facilities reporting ICU or HCU beds. Results for ICU access in fiscal year 2024 showed that 90.1% of the population lived within a 30-minute drive of an ICU facility, 97.8% within 60 minutes, and only 0.8% beyond 90 minutes.
Although 140 of the 330 SMAs had no ICU facility within their administrative boundaries, geographic proximity reduced isolation: 84.7% of residents in those SMAs could reach an ICU facility in an adjacent SMA within 60 minutes of driving. This finding underscores that facility absence at the SMA level does not equate directly to inaccessibility for most residents when road-network travel is considered.
The platform anchored the clustering framework on fiscal year 2024 and projected earlier years into the same multidimensional space to permit longitudinal comparison. Supply-structure types were classified using k-means clustering (k = 6) in an eight-dimensional principal component analysis (PCA) space derived from 2024 features. Earlier-year data (2016–2023) were then projected into that PCA space to assign supply-structure types consistently across years.
This approach allowed the authors to track whether SMAs changed structural type over nine years and to quantify transitions in supply patterns while maintaining a consistent reference frame.
Longitudinal results revealed marked persistence in SMA-level supply structures. Sixty-three percent of SMAs (208 of 330) retained the same structural type across all nine fiscal years analyzed. Adjacent-year rank correlations for a supply-vulnerability index were high (0.887–0.924), indicating stability from year to year; the correlation between 2016 and 2024 was lower but still moderate (rho = 0.711).
The number of SMAs with zero ICU beds remained essentially stationary over the study window, fluctuating only between 133 and 141 SMAs. This constancy indicates a stable subset of roughly 140 SMAs that have lacked ICU beds for nearly a decade, reflecting regional rigidity in the distribution of intensive care capacity.
The authors report changes in inequality measured by the Gini coefficient. Calculated on total ICU/HCU beds, the Gini coefficient declined from 0.384 to 0.262 over the nine years, suggesting an apparent equalization of aggregate critical care capacity. However, when the metric was computed on ICU-type beds alone for fiscal year 2024, the coefficient remained high at 0.365, indicating persistent unequal distribution specifically for ICU-level care.
The composition of capacity growth is central to interpretation: most of the national increase in critical care capacity derived from HCU-type beds, meaning that intermediate care expansion accounts for much of the observed equalization in total beds, while true ICU availability remained comparatively less redistributed.
MeshScope-Region was built on publicly available administrative data, specifically the Hospital Bed Function Reports (byosho-kino-hokoku) covering fiscal years 2016–2024. ICU and HCU beds were identified from notified reimbursement categories within those reports and aggregated to SMAs based on the 2025 classification (330 areas).
Geospatial accessibility used OSRM to compute drive times from 1-km census grid cells with population to facilities that reported ICU or HCU beds. The longitudinal supply-structure analysis used an 8-dimensional PCA derived from fiscal year 2024 variables and k-means clustering (k = 6) to define structural types; earlier years were projected into the same PCA space for consistent classification.
The authors note that ethical review was not required under Japanese guidelines because only publicly available administrative data on institutions were used.
Key conclusions from this SMA-level analysis are: (1) Japan's critical care supply structure shows regional rigidity, with a persistent set of roughly 140 SMAs lacking ICU beds across nearly a decade; (2) road-network accessibility substantially mitigates the practical effects of SMA-level absence for most residents, with the majority reachable within 60 minutes by car; and (3) recent national capacity growth and much of the apparent equalization have been driven largely by expansion of HCU (intermediate care) rather than ICU beds.
MeshScope-Region provides a reproducible, standing evidence base at the geographic unit used for Japanese medical planning cycles, enabling cross-sectional, accessibility, and longitudinal evaluations. The source data, routing methods, and clustering approach are described; specific algorithmic or parameter details beyond those reported here were not provided in the source preprint.
Limitations reported or inherent to the described approach include reliance on administrative reimbursement categories to identify bed types and the fact that this work is a preprint that has not undergone peer review. The platform’s measures of accessibility are based on road-network drive times and do not account for non-transport factors such as bed occupancy, clinical capability differences among facilities, emergency medical service capacity, or temporal variations in traffic.
Overall, the analysis highlights the differing trajectories of ICU versus intermediate critical care expansion in Japan and provides geographically specific evidence that can inform regional planning discussions.