Scholarlink

<p>Postoperative complications remain a&#xa0;substantial burden worldwide, and people with diabetes—representing 15–25% of surgical patients—may particularly benefit from more precise perioperative risk assessment. Artificial intelligence (AI) is gaining importance as perioperative care involves complex and large volumes of data. AI can assist in all phases of care: preoperatively through efficient patient education and risk stratification from routine data; intraoperatively by analyzing dynamic time-series (e.g., arterial pressure waveforms) to predict events like hypotension or hypoxemia; and postoperatively through automated monitoring and outcome prediction. For patients with diabetes, future applications may include the use of continuous glucose monitoring (CGM) or automated insulin delivery (AID) data to detect individual patterns or identify dysglycemic risk, although evidence remains limited. Despite promising developments, AI models require careful validation due to risks of bias, limited interpretability, and data-privacy concerns. With appropriate safeguards, AI may help make perioperative diabetes care more targeted, safer, and personalized.</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

Perioperatives Management von Menschen mit Diabetes mellitus: Datenanalyse und künstliche Intelligenz in der Betreuung

  • 期刊论文

  • 发表日期：

    2026年1月8日

查看全文
  • Dominic Ehrmann

摘要

Postoperative complications remain a substantial burden worldwide, and people with diabetes—representing 15–25% of surgical patients—may particularly benefit from more precise perioperative risk assessment. Artificial intelligence (AI) is gaining importance as perioperative care involves complex and large volumes of data. AI can assist in all phases of care: preoperatively through efficient patient education and risk stratification from routine data; intraoperatively by analyzing dynamic time-series (e.g., arterial pressure waveforms) to predict events like hypotension or hypoxemia; and postoperatively through automated monitoring and outcome prediction. For patients with diabetes, future applications may include the use of continuous glucose monitoring (CGM) or automated insulin delivery (AID) data to detect individual patterns or identify dysglycemic risk, although evidence remains limited. Despite promising developments, AI models require careful validation due to risks of bias, limited interpretability, and data-privacy concerns. With appropriate safeguards, AI may help make perioperative diabetes care more targeted, safer, and personalized.