机器学习算法预测严重脓毒症与脓毒性休克:开发、实施及对临床实践的影响
在一家三级教学医院系统的所有非ICU住院患者中,开发并验证了一种基于电子病历数据的随机森林分类器用于预测严重脓毒症和脓毒性休克。算法推导队列n=162,212(2011年7月至2014年6月),验证队列n=10,448(2015年10月至12月)。算法敏感性26%、特异性98%、阳性预测值29%、阳性似然比13。以静默与弹窗提醒对比(静默n=22,280;提醒n=32,184),提醒使乳酸检测和静脉补液有小幅统计学显著增加,死亡率、出院去向或ICU转入无显著差异,但ICU转入时间缩短。
要点
- 算法敏感性26%、特异性98%、阳性预测值29%、阳性似然比13,用于预测非ICU住院患者严重脓毒症/脓毒性休克。
- 弹窗提醒组较静默组乳酸检测和静脉补液有小幅统计学显著增加。
- 弹窗提醒组与静默组死亡率、出院去向、ICU转入无显著差异,但ICU转入时间缩短。
- 算法训练要求:严重脓毒症或脓毒性休克ICD-9编码、血培养阳性,且乳酸>2.2 mmol/L或收缩压<90 mm Hg。
适用人群与场景
- 所有非ICU住院患者;算法推导队列n=162,212(2011年7月至2014年6月),验证队列n=10,448(2015年10月至12月),静默与提醒对比队列为静默n=22,280、提醒n=32,184(2016年1月至2017年2月)。
- 算法训练纳入患者需具备严重脓毒症或脓毒性休克的ICD-9编码、住院期间血培养阳性,且乳酸>2.2 mmol/L或收缩压<90 mm Hg。
诊断标准与关键阈值
- 算法训练标签定义:严重脓毒症或脓毒性休克ICD-9编码 + 血培养阳性 + 乳酸>2.2 mmol/L或收缩压<90 mm Hg。
研究设计与关键结果
- 研究类型:回顾性队列用于算法推导与验证,以及前后对照影响评估;地点为宾夕法尼亚州费城一家三级教学医院系统。
- 算法性能:敏感性26%、特异性98%、阳性预测值29%、阳性似然比13。
- 静默期n=22,280 vs 提醒期n=32,184:提醒使乳酸检测和静脉补液有小幅统计学显著增加。
- 提醒期与静默期相比,死亡率、出院去向或ICU转入无显著差异,但ICU转入时间缩短。
证据类型、等级与局限
- 证据类型:回顾性队列算法推导与验证,以及前后对照影响评估。
- 局限:原文未报告局限性。
对床旁实践的改变
- 该算法以低敏感性、高特异性预测严重脓毒症和脓毒性休克的即将发生;算法生成的预测提醒小幅影响临床指标(乳酸检测和静脉补液增加),未显著改变死亡率、出院去向或ICU转入,但缩短了ICU转入时间。
- 作者认为下一步包括描述临床对该工具的感知,并优化算法设计与交付。
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
OBJECTIVES: Develop and implement a machine learning algorithm to predict severe sepsis and septic shock and evaluate the impact on clinical practice and patient outcomes.
DESIGN: Retrospective cohort for algorithm derivation and validation, pre-post impact evaluation.
SETTING: Tertiary teaching hospital system in Philadelphia, PA.
PATIENTS: All non-ICU admissions; algorithm derivation July 2011 to June 2014 (n = 162,212); algorithm validation October to December 2015 (n = 10,448); silent versus alert comparison January 2016 to February 2017 (silent n = 22,280; alert n = 32,184).
INTERVENTIONS: A random-forest classifier, derived and validated using electronic health record data, was deployed both silently and later with an alert to notify clinical teams of sepsis prediction.
MEASUREMENT AND MAIN RESULT: Patients identified for training the algorithm were required to have International Classification of Diseases, 9th Edition codes for severe sepsis or septic shock and a positive blood culture during their hospital encounter with either a lactate greater than 2.2 mmol/L or a systolic blood pressure less than 90 mm Hg. The algorithm demonstrated a sensitivity of 26% and specificity of 98%, with a positive predictive value of 29% and positive likelihood ratio of 13. The alert resulted in a small statistically significant increase in lactate testing and IV fluid administration. There was no significant difference in mortality, discharge disposition, or transfer to ICU, although there was a reduction in time-to-ICU transfer.
CONCLUSIONS: Our machine learning algorithm can predict, with low sensitivity but high specificity, the impending occurrence of severe sepsis and septic shock. Algorithm-generated predictive alerts modestly impacted clinical measures. Next steps include describing clinical perception of this tool and optimizing algorithm design and delivery.
参考来源
- Giannini HM, Ginestra JC, Chivers C, et al. A Machine Learning Algorithm to Predict Severe Sepsis and Septic Shock: Development, Implementation, and Impact on Clinical Practice. Crit Care Med. 2019. PMID: 31389839. doi:10.1097/CCM.0000000000003891
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。