Modeling Cyber Attack Impacts on Patient Outcomes

摘要摘要 在过去的25年里，医疗保健经历了重大的数字化转型，导致 而且几乎完全依赖技术来提供临床护理。尽管技术发展如此之快 部署，保护这些系统免受恶意黑客等对手的攻击(这种做法 构成网络安全的学科)已经跟不上技术进步的步伐和无处不在。 针对医疗保健的网络攻击在频率和严重性上一直在增加，导致许多公众 临床护理受损、收入损失和受保护的健康信息被泄露的例子。 此外，绝大多数新生的医疗保健网络安全文献都侧重于保护 患者健康数据，忽略网络攻击对患者安全和临床结果构成的风险。《长河》 学期目标是了解网络攻击对患者预后的负面影响，包括发病率和 死亡率。此应用程序的总体目标是确定哪些临床工作流程、医疗设备 软件系统和其他数字化医院基础设施对患者的潜在危害最大 当恶意黑客使用完整性和可用性进行网络攻击时。中心假设是 针对医疗保健的网络攻击的数据驱动模型最能识别流程和临床工作流 容易对患者的预后产生负面影响。这个项目的基本原理是，它的模型将有助于 创建医疗保健网络安全知识的基础，如果没有这些知识，需要增加的目标 网络安全措施仍将是未知的。这些知识的获得将改变医疗保健 安全范式既包括对网络安全风险的更全面的理解，也包括 考虑网络攻击对患者安全和结果的影响。本项目有两个具体目标：(1) 开发患者数据完整性已受损的医疗保健网络攻击模型；以及(2) 开发关键技术系统可用性受到影响的医疗保健网络攻击模型。这个 First Aim将利用微模拟来模拟医院中遭受完整性网络攻击的患者护理 恶意修改诊断和治疗数据。第二个目标也将利用微模拟，但将 模拟医院中正在经历网络攻击(如勒索软件)的患者的护理 某些技术系统无法运行。这两个目标都将为中风患者的护理树立榜样， 心肌梗塞和败血症。这一应用中提出的研究具有创新性，因为它是第一个 已知的试图对网络攻击对患者预后的影响进行正式建模。拟议的研究 具有重要意义，因为它有望为进一步的临床研究提供坚实的理论基础。 网络攻击对患者结局的影响，包括对真实患者群体的实证研究。另外， 准确、可用的医疗保健网络攻击模型可以为利益相关者提供关键信息 需要适当地保护数字基础设施免受恶意黑客的攻击，将患者安全的风险降至最低。

ABSTRACT SUMMARY Over the last 25 years, healthcare has undergone significant digital transformation resulting in an increasing and near total dependence on technology to deliver clinical care. Despite this rapid acceleration of technology deployment, the protection of these systems from adversaries such as malicious hackers (a practice which constitutes the discipline of cybersecurity) has not matched the pace and ubiquity of technological advances. Cyber attacks on healthcare have been increasing in frequency and severity, resulting in many public examples of compromised clinical care, lost revenue, and breaches of protected health information. Furthermore, a vast majority of the nascent healthcare cybersecurity literature focuses on the protection of patient health data, and ignores the risks cyber attacks pose to patient safety and clinical outcomes. The long term goal is to understand the negative impacts of cyber attacks on patient outcomes including morbidity and mortality. The overall objective of this application is to identify which clinical workflows, medical devices, software systems, and other digitized hospital infrastructure present the greatest potential harm to patients when Integrity and Availability cyber attacks are used by malicious hackers. The central hypothesis is that data-driven models of cyber attacks on healthcare can identify processes and clinical workflows most vulnerable to negatively impacting patient outcomes. The rationale for this project is that its models will help create a foundational base of healthcare cybersecurity knowledge, without which targets in need of increased cybersecurity measures will remain unknown. The acquisition of this knowledge will change the healthcare security paradigm to include both a more holistic understanding of cybersecurity risks but also one that considers the patient safety and outcome impacts of cyber attacks. This project has two specific aims: (1) Develop healthcare cyber attack models where the integrity of patient data has been compromised; and (2) Develop healthcare cyber attack models where the availability of critical technical systems are impacted. The first aim will utilize microsimulation to model patient care in a hospital undergoing integrity cyber attacks that maliciously modify diagnostic and therapeutic data. The second aim will also utilize microsimulation but will model the care of patients in hospitals undergoing availability cyber attacks such as Ransomware which render certain technical systems inoperable. Both aims will model the care of patients presenting with stroke, myocardial infarction, and sepsis. The proposed research in this application is innovative, because it is the first known attempt to formally model the impacts cyber attacks have on patient outcomes. The proposed research is significant because it is expected to provide a strong theoretical foundation to justify further clinical studies of cyber attack patient outcome impacts, including empirical studies on real patient populations. Additionally, accurate and usable models of healthcare cyber attacks can give stakeholders the critical information they need to properly defend digital infrastructure from malicious hackers, minimizing risk to patient safety.