Factors driving provider adoption of the TREWS machine learning-based early warning system and its effects on sepsis treatment timing,Nature Medicine

当前位置： X-MOL 学术 › Nat. Med. › 论文详情

Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)

Factors driving provider adoption of the TREWS machine learning-based early warning system and its effects on sepsis treatment timing
Nature Medicine ( IF 58.7 ) Pub Date : 2022-07-21 , DOI: 10.1038/s41591-022-01895-z
Katharine E Henry _{1,

2} , Roy Adams _{2,

3} , Cassandra Parent ₄ , Hossein Soleimani ₅ , Anirudh Sridharan ₆ , Lauren Johnson ₇ , David N Hager ₈ , Sara E Cosgrove ₈ , Andrew Markowski ₉ , Eili Y Klein ₁₀ , Edward S Chen ₈ , Mustapha O Saheed ₁₀ , Maureen Henley ₇ , Sheila Miranda ₁₁ , Katrina Houston ₇ , Robert C Linton ₆ , Anushree R Ahluwalia ₇ , Albert W Wu _{8,

12,

13,

14} , Suchi Saria _{1,

2,

8,

12,

15}

Affiliation

Department of Computer Science, Johns Hopkins University, Baltimore, MD, USA.
Malone Center for Engineering in Healthcare, Johns Hopkins University, Baltimore, MD, USA.
Department of Psychiatry and Behavioral Science, Johns Hopkins School of Medicine, Baltimore, MD, USA.
Department of Biomedical Engineering, Johns Hopkins University, Baltimore, MD, USA.
Health Informatics, University of California, San Francisco, San Francisco, CA, USA.
Howard County General Hospital, Columbia, MD, USA.
Department of Quality Improvement, The Johns Hopkins Hospital, Baltimore, MD, USA.
Department of Medicine, Johns Hopkins School of Medicine, Baltimore, MD, USA.
Suburban Hospital, Bethesda, MD, USA.
Department of Emergency Medicine, Johns Hopkins School of Medicine, Baltimore, MD, USA.
Department of Medicine, The Johns Hopkins Hospital, Baltimore, MD, USA.
Department of Health Policy and Management, Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, USA.
Department of Epidemiology, Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, USA.
Department of International Health, Johns Hopkins Bloomberg School of Public Health, Baltimore, MD, USA.
Bayesian Health, New York, NY, USA.

Machine learning-based clinical decision support tools for sepsis create opportunities to identify at-risk patients and initiate treatments at early time points, which is critical for improving sepsis outcomes. In view of the increasing use of such systems, better understanding of how they are adopted and used by healthcare providers is needed. Here, we analyzed provider interactions with a sepsis early detection tool (Targeted Real-time Early Warning System), which was deployed at five hospitals over a 2-year period. Among 9,805 retrospectively identified sepsis cases, the early detection tool achieved high sensitivity (82% of sepsis cases were identified) and a high rate of adoption: 89% of all alerts by the system were evaluated by a physician or advanced practice provider and 38% of evaluated alerts were confirmed by a provider. Adjusting for patient presentation and severity, patients with sepsis whose alert was confirmed by a provider within 3 h had a 1.85-h (95% CI 1.66–2.00) reduction in median time to first antibiotic order compared to patients with sepsis whose alert was either dismissed, confirmed more than 3 h after the alert or never addressed in the system. Finally, we found that emergency department providers and providers who had previous interactions with an alert were more likely to interact with alerts, as well as to confirm alerts on retrospectively identified patients with sepsis. Beyond efforts to improve the performance of early warning systems, efforts to improve adoption are essential to their clinical impact and should focus on understanding providers’ knowledge of, experience with and attitudes toward such systems.

中文翻译：

推动供应商采用基于 TREWS 机器学习的早期预警系统的因素及其对脓毒症治疗时机的影响

基于机器学习的脓毒症临床决策支持工具创造了识别高危患者并在早期时间点开始治疗的机会，这对于改善脓毒症结果至关重要。鉴于此类系统的使用越来越多，需要更好地了解医疗保健提供者如何采用和使用它们。在这里，我们分析了提供者与脓毒症早期检测工具（有针对性的实时早期预警系统）的交互，该工具在 2 年期间部署在五家医院。在 9,805 个回顾性确定的败血症病例中，早期检测工具实现了高灵敏度（82% 的败血症病例被识别）和高采用率：系统发出的所有警报中有 89% 由医生或高级实践提供者评估，38%评估的警报由供应商确认。根据患者的表现和严重程度进行调整，与警示为解除、在警报后 3 小时以上确认或从未在系统中解决。最后，我们发现急诊科提供者和之前与警报有过互动的提供者更有可能与警报互动，以及确认关于回顾性确定的脓毒症患者的警报。除了努力提高早期预警系统的性能外，提高采用率的努力对其临床影响至关重要，并且应侧重于了解提供者对此类系统的知识、经验和态度。与警报解除或确认时间超过 3 小时的脓毒症患者相比，供应商在 3 小时内确认其警报的脓毒症患者首次使用抗生素的中位时间缩短了 1.85 小时（95% CI 1.66–2.00）警报后或从未在系统中处理过。最后，我们发现急诊科提供者和之前与警报有过互动的提供者更有可能与警报互动，以及确认关于回顾性确定的脓毒症患者的警报。除了努力提高早期预警系统的性能外，提高采用率的努力对其临床影响至关重要，并且应侧重于了解提供者对此类系统的知识、经验和态度。与警报解除或确认时间超过 3 小时的脓毒症患者相比，供应商在 3 小时内确认其警报的脓毒症患者首次使用抗生素的中位时间缩短了 1.85 小时（95% CI 1.66–2.00）警报后或从未在系统中处理过。最后，我们发现急诊科提供者和之前与警报有过互动的提供者更有可能与警报互动，以及确认关于回顾性确定的脓毒症患者的警报。除了努力提高早期预警系统的性能外，提高采用率的努力对其临床影响至关重要，并且应侧重于了解提供者对此类系统的知识、经验和态度。与警报被解除、警报后 3 小时以上确认或系统中从未处理过的脓毒症患者相比，首次使用抗生素的中位时间缩短 85 小时（95% CI 1.66–2.00）。最后，我们发现急诊科提供者和之前与警报有过互动的提供者更有可能与警报互动，以及确认关于回顾性确定的脓毒症患者的警报。除了努力提高早期预警系统的性能外，提高采用率的努力对其临床影响至关重要，并且应侧重于了解提供者对此类系统的知识、经验和态度。与警报被解除、警报后 3 小时以上确认或系统中从未处理过的脓毒症患者相比，首次使用抗生素的中位时间缩短 85 小时（95% CI 1.66–2.00）。最后，我们发现急诊科提供者和之前与警报有过互动的提供者更有可能与警报互动，以及确认关于回顾性确定的脓毒症患者的警报。除了努力提高早期预警系统的性能外，提高采用率的努力对其临床影响至关重要，并且应侧重于了解提供者对此类系统的知识、经验和态度。在警报发出后超过 3 小时得到确认，或者从未在系统中得到解决。最后，我们发现急诊科提供者和之前与警报有过互动的提供者更有可能与警报互动，以及确认关于回顾性确定的脓毒症患者的警报。除了努力提高早期预警系统的性能外，提高采用率的努力对其临床影响至关重要，并且应侧重于了解提供者对此类系统的知识、经验和态度。在警报发出后超过 3 小时得到确认，或者从未在系统中得到解决。最后，我们发现急诊科提供者和之前与警报有过互动的提供者更有可能与警报互动，以及确认关于回顾性确定的脓毒症患者的警报。除了努力提高早期预警系统的性能外，提高采用率的努力对其临床影响至关重要，并且应侧重于了解提供者对此类系统的知识、经验和态度。

更新日期：2022-07-21

点击分享查看原文

点击收藏

阅读更多本刊最新论文本刊介绍/投稿指南11