Multicentre validation of a sepsis prediction algorithm using only vital sign data in the emergency department, general ward and ICU.

Multicentre validation of a sepsis prediction algorithm using only vital sign data in the emergency department, general ward and ICU.
复制标题

DOI:
10.1136/bmjopen-2017-017833
复制
发表时间:
2018-01-26
期刊:
影响因子:
2.9
通讯作者:
Das R
Das R
中科院分区:
医学3区
文献类型:
--
作者:
Mao Q;Jay M;Hoffman JL;Calvert J;Barton C;Shimabukuro D;Shieh L;Chettipally U;Fletcher G;Kerem Y;Zhou Y;Das R

文献摘要

参考文献

被引文献

相似文献

我们验证了一种基于机器学习的脓毒症预测算法(InSight),用于检测和预测三种脓毒症相关的金标准,仅使用六种生命体征。我们评估了对缺失数据的鲁棒性,使用迁移学习对特定于站点的数据进行定制,以及对新设置的通用性。一种具有梯度树提升的机器学习算法。用于预测的特征是从六个生命体征测量值及其随时间的变化的组合中创建的。来自加州大学弗朗西斯科分校(UCSF)医学中心(弗朗西斯科,加州,美国)的混合病房回顾性数据集作为主要来源,来自Beth Israel Deaconess医学中心(波士顿,马萨诸塞州,美国)的重症监护室数据集作为迁移学习来源,以及四个额外机构的数据集,以评估普遍性。684443总遭遇,与90353遭遇从2011年6月到2016年3月在加州大学旧金山分校。没有。检测和预测脓毒症、严重脓毒症和脓毒性休克的受试者工作特征(AUROC)曲线下面积。对于脓毒症和严重脓毒症的检测,InSight的AUROC曲线分别为0.92(95% CI 0.90至0.93)和0.87(95% CI 0.86至0.88)。在发病前4小时,InSight预测脓毒性休克的AUROC为0.96(95% CI 0.94至0.98),严重脓毒症的AUROC为0.85(95% CI 0.79至0.91)。InSight在识别和预测脓毒症、严重脓毒症和脓毒性休克方面优于现有的脓毒症评分系统。这是第一个仅使用生命体征输入就超过0.90 AUROC的败血症筛查系统。InSight对缺失数据具有强大的鲁棒性,可以使用一小部分研究中心数据定制新的医院数据,并在所有机构中保持强大的区分力。
We validate a machine learning-based sepsis-prediction algorithm (InSight) for the detection and prediction of three sepsis-related gold standards, using only six vital signs. We evaluate robustness to missing data, customisation to site-specific data using transfer learning and generalisability to new settings. A machine-learning algorithm with gradient tree boosting. Features for prediction were created from combinations of six vital sign measurements and their changes over time. A mixed-ward retrospective dataset from the University of California, San Francisco (UCSF) Medical Center (San Francisco, California, USA) as the primary source, an intensive care unit dataset from the Beth Israel Deaconess Medical Center (Boston, Massachusetts, USA) as a transfer-learning source and four additional institutions’ datasets to evaluate generalisability. 684 443 total encounters, with 90 353 encounters from June 2011 to March 2016 at UCSF. None. Area under the receiver operating characteristic (AUROC) curve for detection and prediction of sepsis, severe sepsis and septic shock. For detection of sepsis and severe sepsis, InSight achieves an AUROC curve of 0.92 (95% CI 0.90 to 0.93) and 0.87 (95% CI 0.86 to 0.88), respectively. Four hours before onset, InSight predicts septic shock with an AUROC of 0.96 (95% CI 0.94 to 0.98) and severe sepsis with an AUROC of 0.85 (95% CI 0.79 to 0.91). InSight outperforms existing sepsis scoring systems in identifying and predicting sepsis, severe sepsis and septic shock. This is the first sepsis screening system to exceed an AUROC of 0.90 using only vital sign inputs. InSight is robust to missing data, can be customised to novel hospital data using a small fraction of site data and retains strong discrimination across all institutions.
DOI: 10.1371/journal.pone.0174708
发表时间: 2017
期刊: PloS one
影响因子: 3.7
作者:
Horng S;Sontag DA;Halpern Y;Jernite Y;Shapiro NI;Nathanson LA
通讯作者: Nathanson LA
DOI: 10.1097/ccm.0000000000000026
发表时间: 2014-03
影响因子: 8.8
作者:
Stevenson EK;Rubenstein AR;Radin GT;Wiener RS;Walkey AJ
通讯作者: Walkey AJ
DOI: 10.2196/medinform.5909
发表时间: 2016-09-30
影响因子: 3.2
作者:
Desautels T;Calvert J;Hoffman J;Jay M;Kerem Y;Shieh L;Shimabukuro D;Chettipally U;Feldman MD;Barton C;Wales DJ;Das R
通讯作者: Das R
DOI: 10.1126/scitranslmed.aab3719
发表时间: 2015-08-05
影响因子: 17.1
作者:
Henry, Katharine E.;Hager, David N.;Saria, Suchi
通讯作者: Saria, Suchi
DOI: 10.1056/nejmoa010307
发表时间: 2001-11-08
影响因子: 158.5
作者:
Rivers, E;Nguyen, B;Tomlanovich, M
通讯作者: Tomlanovich, M