Original Coverage & Source Attribution: bioengineer.org
A national analysis of public procurement data across Chinese public hospitals between 2020 and 2024 has found that geographic convergence in the adoption of medical artificial intelligence masks a deepening concentration of diagnostic hardware within top-tier tertiary hospitals. Researchers Panpan Jiang and Lisha Cheng tracked the spatial and institutional diffusion of capital-intensive, AI-enabled diagnostic and screening devices across the country to assess whether real-world deployments aligned with equitable healthcare access goals.
Deploying artificial intelligence in clinical medicine often encounters substantial operational hurdles after algorithm validation, particularly concerning how capital-intensive hardware is allocated across public health systems. To evaluate this post-validation deployment phase, the authors investigated procurement records across Chinese public healthcare facilities over the five-year study period. Their empirical investigation focused specifically on the dual dynamics of geographic distribution and institutional tier stratification.
Methodologically, the study applied dual-dimensional Theil-T indices, spatial agglomeration metrics, and structural location quotients to measure the spread and concentration of AI-enabled diagnostic and screening devices. By combining these analytical frameworks, the research team sought to untangle two coexisting yet contradictory trajectories: macro-level spatial equalization across broad regional territories and micro-level institutional concentration across differing facility tiers.
At the macro-geographic level, total national device volume expanded dramatically over the observation period, rising 37.25-fold from an initial baseline of 8 units in 2020 to 298 units in 2024. This rapid deployment drove a sustained convergence across procurement-based spatial equity indices. The population-based Theil-T index dropped from 0.425 to 0.047, while the hospital-based Theil-T index decreased from 0.591 to 0.070, completely eliminating zero-adoption administrative regions by 2022.
Despite this substantial geographic equalization, severe institutional stratification persisted across hospital hierarchies throughout the same period. Even when the investigators strictly restricted the analysis to healthcare facilities possessing demonstrated capital acquisition capacity, top-tier tertiary hospitals disproportionately accumulated diagnostic equipment, reaching a structural location quotient of 2.154 in 2024. In contrast, secondary hospitals experienced persistent relative deprivation, with their location quotient falling from 0.361 in 2020 to 0.320 by 2024.
A nested decomposition of total inequality demonstrated that the within-region, tier-based component increased from 47.3 percent to 84.5 percent of total inequality between 2020 and 2024. This structural shift confirmed that the apparent convergence observed across geographic boundaries increasingly concealed deep disparities between hospital tiers situated within the very same regions.
The authors concluded that tracking regional equity metrics alone understates the magnitude of institutional inequality in the diffusion of medical artificial intelligence. Because tier-based differences within regions have become the dominant source of remaining disparity, the researchers emphasized that policy frameworks and deployment monitoring must directly incorporate hospital tier classifications rather than focusing exclusively on aggregate device counts or regional territory distributions.
Subject of Research: Institutional and geographic distribution of medical artificial intelligence devices in Chinese public hospitals
Article Title: Geographic convergence masks institutional stratification in the scale-up of medical AI in China
Article References: Jiang, P., & Cheng, L. (2026). Geographic convergence masks institutional stratification in the scale-up of medical AI in China. BMC Health Services Research. https://doi.org/10.1186/s12913-026-16020-w
DOI: 10.1186/s12913-026-16020-w
Keywords: artificial intelligence, medical AI, healthcare equity, diagnostic devices, public procurement, hospital stratification, tertiary hospitals, secondary hospitals, health services research, Theil-T index, spatial agglomeration, location quotients, health technology assessment, capital acquisition, technology diffusion, China healthcare
