Healthcare service structural recovery potential across Chinese provinces, 2009–2024: measurement, obstacle diagnosis, and improvement pathways

Public health emergencies have exposed uneven structural recovery potential across China’s regional healthcare systems. This study measures healthcare service structural recovery potential in 31 provincial-level regions from 2009 to 2024 and diagnoses its internal weaknesses. We work within the "recovery" dimension of the “absorb–adapt–recover–transform” framework, build an indicator system of 7 dimensions and 10 indicators, weight them with the CRITIC method (cross-validated against three alternatives), and score structural recovery potential with TOPSIS. An obstacle degree model identifies the internal bottlenecks. The national mean rose from 0.4294 in 2009 to 0.6380 in 2024, an east-high, west-low gradient in which only the east–west gap is statistically significant. Graded-road density, basic medical insurance coverage, and daily inpatient bed-days per physician are the three largest obstacles (18.05%, 15.51%, 15.23%); the first is a support weakness, the other two healthcare-service weaknesses. The weakness structure varies across regions, provinces, and time. The resource-connectivity obstacle declines from west to east (21.29% to 15.51%). In 2024, graded-road density and physician bed-days are the leading obstacles in 19 and 12 provinces respectively. The insurance-coverage obstacle falls from 23.95% to 8.57%. The bottleneck thus shifts from expanding insurance coverage to reducing physician workload. The weighting method changes the diagnosis. The entropy method inflates the obstacle degree of highly dispersed indicators; the claim that resource connectivity is the single dominant weakness holds only under the entropy and coefficient-of-variation specifications, while the other two methods produce a more balanced structure. The paper closes with three improvement pathways: consolidate insurance coverage, reduce physician workload, and strengthen resource connectivity.