Presentation (CTI)

OBJECTIVES: In 2025, artificial intelligence (AI) integration in Chinese healthcare reached a structural inflection point. This study aims to evaluate the real-world value proposition of AI among Chinese physicians. Specifically, it seeks to identify the applications physicians perceive as "most helpful" and conduct deep-dive case studies on these prioritized tools to understand how they transform clinical decision-making, workflow efficiency, and patient outcomes from an episodic to a longitudinal care model.
METHODS: Data were drawn from the 2025 Digital Life Physician (DLP) cross-sectional survey (N=5,000). Physicians who utilized AI were asked a specific open-ended question to identify the single AI application most beneficial to their clinical work, detailing the disease area and usage scenario. Based on these responses, a mixed-methods approach was applied: quantitative frequency analysis identified top-performing application archetypes, followed by targeted secondary desk research to validate the clinical evidence, operational mechanics, and outcome metrics of the highest-ranked use cases.
RESULTS: The analysis revealed six dominant AI archetypes shifting from experimental tools to core infrastructure. In chronic disease management, AI-driven physician-patient collaborative platforms improved hypertension control rates from 45% to 70% by automating follow-ups. Disease-specific Large Language Models (LLMs) for rare diseases reduced diagnostic timelines from four years to under four weeks through explainable reasoning chains. In imaging, functional assessment tools (e.g., CT-derived fractional flow reserve) and stroke decision support systems compressed interpretation times significantly (e.g., from 30 to 3 minutes) and extended therapeutic windows from 4.5 to 24 hours, enabling standardized care across tiered hospital systems.
CONCLUSIONS: Chinese physicians are actively transitioning AI from a diagnostic aid to a continuous care enabler. The physician-prioritized cases demonstrate that high-value AI applications are those embedded in workflows that standardize decisions and improve longitudinal outcomes. This necessitates HEOR value frameworks that capture long-term disease management benefits and system-wide equity improvements rather than solely immediate procedural efficiency.