From local improvement to spatial diffusion: healthcare digital transformation and resource allocation efficiency in China

Introduction The inefficient and spatially uneven allocation of healthcare resources remains a major public health governance challenge in China. Healthcare digital transformation may improve the use of existing resources; however, its interprovincial spillovers, regional and temporal heterogeneity, and implications for efficiency convergence remain insufficiently understood. This study examines whether healthcare digital transformation is associated with higher healthcare resource allocation efficiency, whether this association extends across provincial boundaries, how it varies across regions and over time, and whether digitalization is associated with faster interprovincial efficiency convergence. Methods We constructed a balanced panel of 30 Chinese provincial-level units from 2014 to 2023. Healthcare resource allocation efficiency was measured using a super-efficiency non-radial directional distance function with undesirable outputs, and healthcare digital transformation was measured using an entropy-weighted TOPSIS index. A two-way fixed-effects spatial Durbin model with three alternative spatial weight matrices was used to estimate direct, indirect, and total associations. Regional subgroup analysis and geographically and temporally weighted regression assessed spatial and temporal heterogeneity, while spatial sigma- and beta-convergence models examined changes in interprovincial disparities. All estimates are interpreted as associational. Results Healthcare digital transformation was positively associated with resource allocation efficiency under all three spatial weight matrices. Under the primary contiguity matrix, the indirect effect (0.0521) was approximately twice the direct effect (0.0258), yielding a total effect of 0.0779; this ordering was preserved under the alternative matrices. Regional point estimates suggested stronger spillovers in Eastern China and more localized associations in Western China, although no pairwise regional contrast remained significant after Holm adjustment. The mean local digitalization coefficient increased from 0.0095 in 2014 to 0.0658 in 2023, while the national coefficient of variation declined from 0.352 to 0.218. Estimated convergence half-lives declined from 13.0 years in the absolute model to 8.3 years after structural controls, 8.2 years after adding Dig , and 7.5 years in the full spatial digitalization model. Discussion Healthcare digital transformation was associated with higher resource allocation efficiency, with a substantial cross-provincial component and clear spatiotemporal heterogeneity. Digitalization was also associated with faster conditional convergence, although structural factors explained much of the acceleration. These findings highlight the importance of interprovincial coordination, interoperability, and capacity building, while the observational design precludes causal interpretation of specific policy mechanisms.