Scholarlink

Background <p>Generative language models (large language models, LLM) such as GPT‑4 are increasingly entering medical communication. Particularly in documentation-intensive fields such as cardiac surgery, they offer the potential to reduce workload, standardize texts and improve interdisciplinary communication. The targeted design of inputs, i.e., prompt engineering, emerges as a&#xa0;new key competence.</p> Objective <p>This study explores the potential, fields of application and challenges of prompt engineering in cardiac surgery. The aim is to develop a&#xa0;framework for the effective and responsible integration of artificial intelligence (AI) models into documentation, patient communication and decision support.</p> Method <p>A&#xa0;narrative review of current studies (2021–2025) on LLMs and prompt engineering in medical contexts was conducted, complemented by observations from workshops, simulations and clinical scenarios. Ethical, regulatory and safety aspects were also considered.</p> Results <p>The use of AI can reduce the documentation time in cardiac surgery by up to 90% without compromising quality. Optimized prompts resulted in more complete operation reports, clearer patient information documents and more structured discharge letters. Risks such as hallucinations, bias, and sensitivity were mitigated through precise input structures.</p> Conclusion <p>Prompt engineering is evolving into a&#xa0;core medical competence. Carefully designed prompts enable more efficient and higher quality communication processes in cardiac surgery; however, successful implementation requires data protection-compliant IT infrastructures, certified prompt libraries and structured training programs. Applied responsibly, AI can thus contribute to patient-centered, evidence-based care.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Next Generation Herzchirurgie: Prompt Engineering

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年3月9日

查看全文
  • Rauf Safarov,
  • J. Wippermann,
  • M. Wacker

摘要

Background

Generative language models (large language models, LLM) such as GPT‑4 are increasingly entering medical communication. Particularly in documentation-intensive fields such as cardiac surgery, they offer the potential to reduce workload, standardize texts and improve interdisciplinary communication. The targeted design of inputs, i.e., prompt engineering, emerges as a new key competence.

Objective

This study explores the potential, fields of application and challenges of prompt engineering in cardiac surgery. The aim is to develop a framework for the effective and responsible integration of artificial intelligence (AI) models into documentation, patient communication and decision support.

Method

A narrative review of current studies (2021–2025) on LLMs and prompt engineering in medical contexts was conducted, complemented by observations from workshops, simulations and clinical scenarios. Ethical, regulatory and safety aspects were also considered.

Results

The use of AI can reduce the documentation time in cardiac surgery by up to 90% without compromising quality. Optimized prompts resulted in more complete operation reports, clearer patient information documents and more structured discharge letters. Risks such as hallucinations, bias, and sensitivity were mitigated through precise input structures.

Conclusion

Prompt engineering is evolving into a core medical competence. Carefully designed prompts enable more efficient and higher quality communication processes in cardiac surgery; however, successful implementation requires data protection-compliant IT infrastructures, certified prompt libraries and structured training programs. Applied responsibly, AI can thus contribute to patient-centered, evidence-based care.