Development and Validation of a Machine Learning Algorithm Using Clinical Pages to Predict Imminent Clinical Deterioration.
Development and Validation of a Machine Learning Algorithm Using Clinical Pages to Predict Imminent Clinical Deterioration.
复制标题
DOI：
10.1007/s11606-023-08349-3
复制
发表时间：
2024-01
影响因子：
5.7
通讯作者：
Wright, Adam
中科院分区：
文献类型：
--
作者：
Steitz, Bryan D. D.;McCoy, Allison B. B.;Reese, Thomas J. J.;Liu, Siru;Weavind, Liza;Shipley, Kipp;Russo, Elise;Wright, Adam
早期发现住院患者的临床恶化是患者安全和护理质量的临床优先事项。目前用于识别这些患者的自动化方法在识别迫在眉睫的事件方面表现不佳。开发一种机器学习算法，使用临床团队成员之间发送的寻呼消息来预测即将出现的临床恶化。我们使用长短期记忆机器学习模型对临床页面的内容和频率进行了一项大型观察性研究。我们包括2018年1月1日至2020年12月31日期间在范德比尔特大学医学中心住院的所有患者，其中至少包括一条发给医生的页面消息。排除标准包括接受姑息治疗的患者，有计划的重症监护住院的患者，以及住院时间最长的前2%的患者。模型分类性能，以识别院内心脏骤停、转入重症监护或在接下来的3、6和12小时内快速反应激活。我们将模型性能与三个常见的预警评分进行了比较：修正预警评分、国家预警评分和史诗恶化指数。共有87,783名患者(平均年龄54.0[18.8]岁；女性45,835[52.2%])经历了136,778次住院治疗。6214名住院患者经历了恶化事件。机器学习模型准确地识别了事件发生前3小时内62%的恶化事件和12小时内47%的事件。在每个时间范围内，该模型都超过了最佳预警分数的表现，包括6小时的接收器操作特征曲线下面积(0.856比0.781)、6小时的灵敏度(0.590比0.505)、6小时的特异度(0.900比0.878)和6小时的F分数(0.291比0.220)。将机器学习应用于临床页面的内容和频率，提高了对即将恶化的预测。使用临床页面监控患者的敏锐度支持改进对即将恶化的检测，而不需要更改临床工作流程或护理文档。网上版载有补充材料，可在10.1007/s11606-023-08349-3查阅。
Early detection of clinical deterioration among hospitalized patients is a clinical priority for patient safety and quality of care. Current automated approaches for identifying these patients perform poorly at identifying imminent events. Develop a machine learning algorithm using pager messages sent between clinical team members to predict imminent clinical deterioration. We conducted a large observational study using long short-term memory machine learning models on the content and frequency of clinical pages. We included all hospitalizations between January 1, 2018 and December 31, 2020 at Vanderbilt University Medical Center that included at least one page message to physicians. Exclusion criteria included patients receiving palliative care, hospitalizations with a planned intensive care stay, and hospitalizations in the top 2% longest length of stay. Model classification performance to identify in-hospital cardiac arrest, transfer to intensive care, or Rapid Response activation in the next 3-, 6-, and 12-hours. We compared model performance against three common early warning scores: Modified Early Warning Score, National Early Warning Score, and the Epic Deterioration Index. There were 87,783 patients (mean [SD] age 54.0 [18.8] years; 45,835 [52.2%] women) who experienced 136,778 hospitalizations. 6214 hospitalized patients experienced a deterioration event. The machine learning model accurately identified 62% of deterioration events within 3-hours prior to the event and 47% of events within 12-hours. Across each time horizon, the model surpassed performance of the best early warning score including area under the receiver operating characteristic curve at 6-hours (0.856 vs. 0.781), sensitivity at 6-hours (0.590 vs. 0.505), specificity at 6-hours (0.900 vs. 0.878), and F-score at 6-hours (0.291 vs. 0.220). Machine learning applied to the content and frequency of clinical pages improves prediction of imminent deterioration. Using clinical pages to monitor patient acuity supports improved detection of imminent deterioration without requiring changes to clinical workflow or nursing documentation. The online version contains supplementary material available at 10.1007/s11606-023-08349-3.
登录
查看更多内容
DOI：
10.1093/intqhc/mzx149
发表时间：
2017-12-01
期刊：
International journal for quality in health care : journal of the International Society for Quality in Health Care
影响因子：
--
作者：
Chua WL;See MTA;Legio-Quigley H;Jones D;Tee A;Liaw SY
通讯作者：
Liaw SY
影响因子：
3.7
作者：
Henriksen, Daniel Pilsgaard;Brabrand, Mikkel;Lassen, Annmarie Touborg
通讯作者：
Lassen, Annmarie Touborg
影响因子：
2.2
作者：
Collins, Sarah A.;Vawdrey, David K.
通讯作者：
Vawdrey, David K.
DOI：
10.1093/jamia/ocab111
发表时间：
2021-07-16
影响因子：
6.4
作者：
Fu, Li-Heng;Knaplund, Chris;Rossetti, Sarah Collins
通讯作者：
Rossetti, Sarah Collins
影响因子：
3.9
作者：
Kia, Arash;Timsina, Prem;Levin, Matthew A.
通讯作者：
Levin, Matthew A.