The “DeepSeek effect” and the adoption–integration gap of generative artificial intelligence in clinical practice: a national online convenience cross-sectional survey of academic critical care physicians in China - Report - MDSpire

The “DeepSeek effect” and the adoption–integration gap of generative artificial intelligence in clinical practice: a national online convenience cross-sectional survey of academic critical care physicians in China

  • By

  • Yuankai Zhou

  • Wanrong Xiong

  • Huiying Liu

  • Qianlin Wang

  • Hongbo Luo

  • Yingying Yang

  • Shengjun Liu

  • Yun Long

  • June 23, 2026

  • 0 min

Share

Clinical Report: The Impact of the DeepSeek Effect on GAI in Critical Care

Overview

The introduction of DeepSeek significantly increased self-reported GAI utilization among Chinese academic critical care physicians, rising from 64.7% to 94.1%. However, there was no corresponding increase in formal GAI training or self-assessed professional competence.

Background

The integration of generative artificial intelligence (GAI) into healthcare is crucial for enhancing clinical decision-making, particularly in critical care settings where accuracy is vital. Despite the potential benefits, a gap exists between the adoption of GAI tools and their effective integration into clinical practice.

Data Highlights

MetricPre-DeepSeekPost-DeepSeekP-value
GAI Utilization64.7%94.1%< 0.001
Formal GAI Training13.2%13.7%0.84
Trained Physicians Completing Training< 30%N/AN/A

Key Findings

  • Self-reported GAI utilization increased from 64.7% to 94.1% after DeepSeek's launch.
  • Formal GAI training participation remained low at 13.2% pre-DeepSeek and 13.7% post-DeepSeek.
  • Less than 30% of trained physicians completed structured GAI training.
  • Completion of structured training was linked to improved self-reported competence (AOR = 22.2).
  • Physicians emphasized the need for critical integration skills over basic technical proficiency (OR = 16.3).

Clinical Implications

Addressing the gap between GAI adoption and effective integration is essential.

Conclusion

The DeepSeek effect has led to increased GAI adoption among critical care physicians, yet the lack of formal training and integration strategies remains a challenge.

Related Resources & Content

  1. Frontiers in Medicine, 2026 -- An evaluation of DeepSeek and healthcare professionals’ Q&A capabilities in improving patient-family satisfaction in the ICU
  2. npj Digital Medicine, 2026 -- Evaluating the Real-World Application of Deep Learning Technologies in Healthcare: A Systematic Review Based on Implementation Science Principles
  3. Journal of Medical Internet Research (JMIR), 2026 -- Applications of DeepSeek in Medicine: Bibliometric Analysis and Scoping Review
  4. Critical Care (Springer), 2026 -- Understanding Generative AI's Influence on Perceptions of Racial and Gender Diversity in Critical Care Medicine
  5. WHO, 2025 -- Ethics and governance of artificial intelligence for health: Guidance on large multi-modal models
  6. An LLM chatbot to facilitate primary-to-specialist care transitions: a randomized controlled trial
  7. Human–large language model collaboration in clinical medicine: a systematic review and meta-analysis
  8. Ethics and governance of artificial intelligence for health: Guidance on large multi-modal models

Original Source(s)

Related Content