SBIR Phase I: Adaptive E-Triage in Emergency Medicine

与目前的护理标准相比，这项小企业创新研究（SBIR）第一阶段项目的更广泛影响/商业潜力是通过改善患者就诊时的风险分层（分流），推动更安全、更具成本效益的急诊科护理途径。电子分诊解决了急诊科拥挤的危机（美国每年有1.36亿人次就诊），这对患者的健康结果产生了不利影响，并导致了美国安全网的财政不可持续状态。电子分诊方法支持新的急诊科业务模式，将急症患者和非急症患者的服务流程分开。为了缓解急诊科拥挤，需要新的流模式：(1)为真正需要急诊护理的患者保留稀缺的急诊科资源；(2)防止非急诊患者不必要的等待和昂贵的资源过度利用。它通过使用本地ED电子健康记录（EHR）数据来根据关键事件的风险和疾病的严重程度对患者进行科学的风险分层。电子分诊在缓解拥挤、提高急诊科运营绩效和改善向急诊科患者提供的医疗保健价值方面具有商业机会。拟议的项目将在支持增长的商业模式下，将电子分诊转变为可扩展的商业平台。拟议的项目将产生一个规模庞大的商业化电子分诊决策支持平台，目前正在多个急诊科进行试点。电子分诊部署了数据科学方法和灵活的信息技术架构的新颖组合，支持不同ED客户的可用性。该工具依赖于先进的机器学习方法、利用用户反馈的机制以及灵活且可与EHR系统互操作的软件技术。它还必须安全地传输和存储患者数据，并具有计算效率，以适应快节奏的ED环境。电子分诊使急诊患者在就诊时能够根据关键事件的风险和疾病的严重程度，使用常见的本地收集的急诊数据，快速进行数据驱动的预后。与严重依赖提供者主观判断的美国分诊实践标准相比，基于回顾性和前瞻性评估的证据，电子分诊证明了对高风险和低风险患者的更好识别。电子分诊在设计上具有颠覆性，它支持新的急诊科运营模式，将急症患者和非急症患者的服务流分开，以减轻急诊科拥挤的负担。

The broader impact/commercial potential of this Small Business Innovation Research (SBIR) Phase 1 project is to drive safer and more cost-effective emergency department care pathways by improved risk stratification at patient presentation (triage) compared to the current standards of care. E-triage addresses the ED crowding crisis (136 million visits in US annually) that adversely affects patients' health outcomes and has led to a state of financial unsustainability in America's safety net. E-triage's approach supports new ED operational models to separate service streams for acutely ill and non-urgent patients. New streaming models are needed to mitigate ED crowding by: (1) conserving scarce ED resources for patients truly in need of emergency care, and (2) preventing unnecessary waiting and costly resource over-utilization for non-urgent patients. It does this by using local ED electronic health record (EHR) data to scientifically risk-stratify patients based on risk of critical events and severity of illness. E-triage meets a commercial opportunity to mitigate crowding, enhance ED operational performance, and improve the value of healthcare delivered to ED patients. The proposed project will transition E-triage to a scalable and commercially available platform under a business model that supports growth. The proposed project will yield a scaled and commercially available e-triage decision support platform that is currently being piloted in multiple emergency departments (EDs). E-triage deploys a novel combination of data-science methods and flexible information technology architecture that supports usability by diverse ED customers. The tool relies on advancements in machine learning methods, mechanisms to harness user feedback, and software technology that is flexible and interoperable with EHR systems. It must also securely transmit and store patient data and be computationally efficient to accommodate fast-paced ED environments. E-triage enables rapid data-driven prognostication of ED patients at presentation based on risk of critical events and severity of illness using common locally collected ED data. Compared to US triage practice standards, which relies heavily on provider subjective judgment, e-triage demonstrates improved identification of high- and low-risk patients based on evidence from retrospective and prospective evaluation. E-triage is disruptive in its design to support new ED operational models that separate service streams for acutely ill and non-urgent patients toward reducing the burden of ED crowding.