Exploring spatial distribution of occupancy rates in elderly care facilities and its influencing factors: A study based on Beijing, China

Amid global aging, elderly care facilities (ECFs) providing long-term care commonly face the challenge of low occupancy rates (ORs). Existing OR studies lack a systematic integration of internal and external factors, geospatial perspectives, and typological analysis. To support efficient ECF allocation, this study took Beijing, a deeply aging city, as a case to explore the spatial distribution of OR and its influencing factors from both overall and typological perspectives. Based on multi-source data from 546 ECFs, geospatial techniques like Moran’s I quantified OR distribution patterns. Correlation and regression analyses identified influencing factors and their relationships, which were further interpreted through word frequency and content analysis of coded user evaluations. Results showed a clustered urban–rural divide (Moran’s I = 0.189, p < 0.01) with high and stable OR in central urban areas (76.48% ± 16.52%) and low and volatile OR in suburbs (70.81% ± 21.87%). Spatial patterns varied by facility scale, presenting as core‑clustered, near‑city random, and north‑south divergent. Endogenous capacity had the strongest influence on OR distribution. Star rating was the top positive factor for overall-, small-, and large-scaled ECFs (mean | β | = 4.140), while metro stations most strongly affected medium‑scaled ECFs (| β | = 1.935). Social development factors such as GDP per capita and residential price showed no significant association. User evaluations highlighted service level (32.29%) and built quality (24.11%), underscoring the dominance of star rating and illustrating the decision logic that prioritizes service quality. Based on these findings, differentiated strategies for ECFs of different scales were proposed while strengthening endogenous capacity. This study provides a replicable empirical basis for precise planning and efficient resource allocation of ECFs in similar cities.