Grosman-Rimon Liza, Rivlin Leon, Spataro Rosa, Zhu Zhiqiang, Casey Jane, Tory Susan, Solanki Jhanvi, Wegier Pete
Humber River Health, Toronto, Canada.
University of Toronto, Institute of Health Policy, Management and Evaluation, Toronto, Canada.
Digit Health. 2024 Apr 26;10:20552076241250255. doi: 10.1177/20552076241250255. eCollection 2024 Jan-Dec.
Sepsis alerts based on laboratory and vital sign criteria were found insufficient to improve patient outcomes. While most early sepsis alerts were implemented into smaller scale operating systems, a centralized new approach may provide more benefits, overcoming alert fatigue, improving deployment of staff and resources, and optimizing the overall management of sepsis. The objective of the study was to assess mortality and length of stay (LOS) trends in emergency department (ED) patients, following the implementation of a centralized and automated sepsis alert system.
The automated sepsis alert system was implemented in 2021 as part of a hospital-wide command and control center. Administrative data from the years 2018 to 2021 were collected. Data included ED visits, in-hospital mortality, triage levels, LOS, and the Canadian Triage and Acuity Scale (CTAS).
Mortality rate for patients classified as CTAS I triage level was the lowest in 2021, after the implementation of the automated sepsis alert system, compared to 2020, 2019, and 2018 ( < 0.001). The Kaplan-Meier survival curve revealed that for patients classified as CTAS I triage level, the probability of survival was the highest in 2021, after implementation of the sepsis alert algorithm, compared to previous years (Log Rank, Mantel-Cox, χ²=29.742, < 0.001). No significant differences in survival rate were observed for other triage levels.
Implementing an automated sepsis alert system as part of a command center operation significantly improves mortality rate associated with LOS in the ED for patients in the highest triage level. These findings suggest that a centralized early sepsis alert system has the potential to improve patient outcomes.
基于实验室检查和生命体征标准的脓毒症警报被发现不足以改善患者预后。虽然大多数早期脓毒症警报是在较小规模的操作系统中实施的，但一种集中式的新方法可能会带来更多益处，克服警报疲劳，改善人员和资源的调配，并优化脓毒症的整体管理。本研究的目的是评估在实施集中式自动化脓毒症警报系统后，急诊科（ED）患者的死亡率和住院时间（LOS）趋势。
自动化脓毒症警报系统于2021年作为全院指挥与控制中心的一部分实施。收集了2018年至2021年的行政数据。数据包括急诊就诊、院内死亡率、分诊级别、住院时间和加拿大分诊与 acuity 量表（CTAS）。
与2020年、2019年和2018年相比，在实施自动化脓毒症警报系统后的2021年，分类为CTAS I级分诊级别的患者死亡率最低（<0.001）。Kaplan-Meier生存曲线显示，对于分类为CTAS I级分诊级别的患者，在实施脓毒症警报算法后的2021年，生存概率高于前几年（对数秩检验，Mantel-Cox检验，χ²=29.742，<0.001）。其他分诊级别未观察到生存率的显著差异。
作为指挥中心运作的一部分实施自动化脓毒症警报系统，可显著提高急诊科最高分诊级别患者与住院时间相关的死亡率。这些发现表明，集中式早期脓毒症警报系统有改善患者预后的潜力。
