Modeling Cyber Attack Impacts on Patient Outcomes

抽象概括 在过去的25年里，医疗保健经历了重大的数字化转型， 并且几乎完全依赖技术来提供临床护理。尽管科技发展如此迅速 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）并没有跟上技术进步的步伐和无处不在。 对医疗保健的网络攻击频率和严重程度都在增加，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泄露的例子。 此外，绝大多数新兴的医疗保健网络安全文献都侧重于保护 患者健康数据，而忽视了网络攻击对患者安全和临床结果造成的风险。长 长期目标是了解网络攻击对患者结局的负面影响，包括发病率和 mortality.本申请的总体目标是确定哪些临床工作流程、医疗器械， 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 当完整性和可用性网络攻击被恶意黑客使用时。核心假设是， 对医疗保健网络攻击的数据驱动模型可以识别流程和临床工作流程， 容易对患者结果产生负面影响。这个项目的理由是，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， 网络安全措施仍是未知数。获得这些知识将改变医疗保健 安全范式既包括对网络安全风险的更全面的理解， 考虑网络攻击对患者安全和结果的影响。该项目有两个具体目标：（1） 开发患者数据完整性受损的医疗保健网络攻击模型;以及（2） 开发医疗保健网络攻击模型，其中关键技术系统的可用性受到影响。的 第一个目标将利用微观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观模拟，但将 模拟医院中遭受可用性网络攻击（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模拟中风患者的护理， 心肌梗塞和败血症在这项应用中提出的研究是创新的，因为它是第一个 已知尝试正式模拟网络攻击对患者结局的影响。拟议研究 是重要的，因为它预计将提供一个强大的理论基础，以证明进一步的临床研究， 网络攻击对患者结局的影响，包括对真实的患者人群的实证研究。此外，本发明还 准确和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， 需要妥善保护数字基础设施免受恶意黑客攻击，最大限度地降低患者安全风险。

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.