Concordance between physician supply and inpatient utilization across Japanese prefectures: a cross-sectional ecological analysis

Equitable distribution of the physician workforce is central to health equity. In universal-coverage systems, whether physician supply is aligned with realized inpatient utilization across regions is debated. We examined the concordance between prefectural physician supply and inpatient utilization in Japan and identified prefectures that diverge from the overall pattern. We conducted a cross-sectional ecological analysis of all 47 prefectures using publicly available government data for 2020. The exposure was physician density (physicians per 100,000 population) and the outcome was the inpatient discharge rate (discharges per 100,000 population, summed across the standardized major diagnostic categories). Associations were assessed using Spearman’s rank and Pearson correlations, population-quartile-stratified correlations, Spearman partial correlation adjusting for population aging, and multivariable linear regression adjusting for the percentage of the population aged ≥ 65 years, population size, and population density. Prefectures were cross-classified around national medians into four quadrants. Reporting followed STROBE. Physician density (mean 286.0, SD 51.0 per 100,000) and inpatient discharges (mean 8,556.8, SD 921.3 per 100,000) each showed substantial geographic variation across prefectures. Physician supply was strongly and positively correlated with inpatient utilization (Spearman ρ = 0.75, p < 0.001; Pearson r = 0.73). The association persisted after adjustment for population aging (partial ρ = 0.77) and in multivariable regression (adjusted β = 12.4 additional discharges per 100,000 per one-unit increase in physician density; 95% CI 8.8–16.0; model R² = 0.69). Positive associations were observed within every population-size stratum (ρ = 0.50–0.85). Forty-three of 47 prefectures fell within concordant quadrants (22 high-supply/high-utilization; 21 low-supply/low-utilization); only four diverged—two high-utilization/low-supply (Akita, Yamaguchi) and two low-utilization/high-supply (Tokyo, Miyazaki). In Japan’s universal-coverage system, physician supply is broadly concordant with regional inpatient utilization. Divergence is confined to a small number of prefectures, which may warrant targeted attention rather than indicating pervasive structural inequity. Not applicable.