文献速递复苏、休克与脓毒症创伤与烧伤重症综合管理与镇静镇痛

通过马尔可夫决策过程实现标准化ICU创伤复苏

Toward Standardized ICU Trauma Resuscitation via Markov Decision Processes.
PubMed 文献 · Shock · 2025-09-12
一句话结论在成人创伤ICU患者中,基于马尔可夫决策过程训练的强化学习策略与临床实践总体一致,但在去甲肾上腺素和血管加压素使用上存在显著差异,离线评估提示学习策略表现不及行为策略。

本研究为回顾性观察研究,纳入2012至2019年间入住一级创伤中心ICU、乳酸≥2 mmol/L的成人创伤患者,将复苏定义为ICU入科48小时内最后一次静脉液体推注(≥250 mL)或血管活性药物给药后1小时。研究构建可解释的马尔可夫决策过程模型,训练强化学习智能体推荐静脉液体和血管活性药物,并与实际临床实践比较。共纳入4305例受试者,其中3027例(70%)达到复苏终点(中位复苏时间22小时)。学习策略与临床实践总体一致,但推荐去甲肾上腺素更少(4%±6% vs 12%,P<0.001)、血管加压素更多(13%±12% vs 4%,P<0.001),推荐静脉液体≥250 mL的比例为19%±12%,观察实践中为13%。离线评估估计学习策略表现不及行为策略,与先前脓毒症领域的发现一致。

要点

  • 共纳入4305例成人创伤ICU患者,3027例(70%)达到复苏终点,中位复苏时间22小时。
  • 学习策略推荐去甲肾上腺素的比例为4%±6%,显著低于观察实践的12%(P<0.001)。
  • 学习策略推荐血管加压素的比例为13%±12%,显著高于观察实践的4%(P<0.001)。
  • 学习策略推荐静脉液体≥250 mL的比例为19%±12%,观察实践中为13%。
  • 离线评估估计学习策略表现不及行为策略,与先前脓毒症领域的发现一致。

适用人群与场景

  • 2012至2019年间入住高容量一级学术创伤中心ICU的成人创伤患者,且乳酸≥2 mmol/L。

诊断标准与关键阈值

  • 复苏终点定义为ICU入科48小时内最后一次静脉液体推注(≥250 mL)或血管活性药物给药后1小时。

研究设计与关键结果

  • 研究类型为回顾性观察研究,纳入4305例受试者,其中3027例(70%)达到复苏终点,中位复苏时间22小时。
  • 学习策略与临床实践总体一致,但推荐去甲肾上腺素更少(4%±6% vs 12%,P<0.001),推荐血管加压素更多(13%±12% vs 4%,P<0.001)。
  • 学习策略推荐静脉液体≥250 mL的比例为19%±12%,观察实践中为13%。
  • 离线评估估计学习策略表现不及行为策略,与先前脓毒症领域的发现一致。

证据类型、等级与局限

  • 证据类型为回顾性观察研究。
  • 局限:原文未报告局限性。

对床旁实践的改变

  • 本研究提出的马尔可夫决策过程框架可用于创伤复苏的回顾性强化学习建模,学习策略与临床医生行为一致但暴露了血管活性药物使用的变异。
  • 作者认为这些发现验证了所提出马尔可夫决策过程在创伤救治中的相关性,展示了在该背景下应用强化学习的可行性,并为未来建模、评估和决策支持奠定了基础。

提炼说明

  • 本条依据论文摘要整理;摘要未涉及「处置流程与药物剂量」「红旗征与不能漏的情况」「指南或共识出处」,故未列出这些栏目;全文请点「阅读原文」。
英文原文(摘要)· 点击展开

OBJECTIVE: Guideline-based recommendations for posthemostasis resuscitation in trauma patients remain limited. This study aimed to define an interpretable Markov Decision Process (MDP) to model intensive care unit (ICU) trauma resuscitation and evaluate its potential to support clinical decision-making using a previously established framework from sepsis care.

DESIGN: Retrospective observational study.

SETTING: High-volume, level I academic trauma center.

PATIENTS: Adult trauma patients admitted to the ICU between 2012 and 2019 with lactate ≥2 mmol/L.

INTERVENTIONS: Resuscitation was framed as a sequential decision-making problem using an MDP and focused on intravenous fluid (IVF) and vasopressor (norepinephrine and vasopressin) administration. Resuscitation was defined as concluding 1 hour after the last IVF bolus (≥250 mL) or vasopressor administration within the first 48 hours of ICU admission. A reinforcement learning (RL) agent was trained to recommend IVF and vasopressor treatments. Learned policy actions were compared to observed clinical practice using McNemar's test, and model performance was evaluated retrospectively.

RESULTS: A total of 4,305 subjects were included; 3,027 (70%) met the resuscitation endpoint (median resuscitation time: 22 hours). The learned policy aligned with clinical practice but differed significantly in vasopressor use, recommending norepinephrine less often (4% ± 6% vs . 12%, P < 0.001) and vasopressin more frequently (13% ± 12% vs . 4%, P < 0.001). IVF ≥250 mL was recommended in 19% ± 12% of decisions versus 13% in observed practice. Offline evaluation estimated the learned policy underperformed the behavior policy, consistent with prior findings in sepsis.

CONCLUSION: This study presents a clinically grounded MDP framework for trauma resuscitation that enables RL to be applied retrospectively in critical care. The learned policy mirrored trends previously observed in sepsis, aligning with clinician behavior while exposing variation in vasopressor use. These findings validate the relevance of the proposed MDP in trauma care, demonstrate the feasibility of applying RL in this context, and establish a foundation for future modeling, evaluation, and decision support.

来源:Shock,查看原文。版权归原作者所有。

参考来源

  1. Preuett LD, Murphy BE, Beni C, et al. Toward Standardized ICU Trauma Resuscitation via Markov Decision Processes. Shock. 2025. PMID: 41166072. doi:10.1097/SHK.0000000000002707

供执业医师抢救与诊疗参考,不替代临床判断;剂量请结合体重、肝肾功能和本院药品规格复核,以指南原文为准。