Spatiotemporal evolution of innovation collaboration networks in China’s AI medical device industry: implications for public health governance

Abstract

Artificial intelligence (AI) medical devices have become an important technological foundation for enhancing healthcare system resilience and promoting the equitable allocation of medical innovation resources. However, limited attention has been paid to the distribution of upstream innovation resources and the spatial organization of this industry in the wake of the pandemic. Using AI medical device patent collaboration data from the Yangtze River Delta during 2018-2025, this study constructs local and external collaborative innovation networks and applies social network analysis and the Geodetector to examine their spatial evolution and associated factors across different phases of the COVID-19 pandemic. The results show that all innovation networks expanded continuously, particularly the external collaborative network. Nevertheless, cross-regional collaboration remained concentrated in a limited number of core cities, resulting in a persistent core-periphery structure. Regional GDP, retail market size, higher education resources, science expenditure, and financial support consistently exhibited strong explanatory power for collaboration intensity, while GDP growth and openness became increasingly associated with network differentiation following the COVID-19 shock. Moreover, the explanatory power of most variables was stronger for external collaboration than for local collaboration, suggesting that cross-regional knowledge exchange relies more heavily on comprehensive innovation capacity. This study advances the understanding of the spatial organization of AI medical innovation and provides empirical evidence for promoting a more balanced allocation of cross-regional medical R&D resources.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 6 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.