Structuring Assurance Cases to Facilitate Effective Certification of Medical Devices

软件是一种非常好的使能技术，现在市场上几乎没有不以某种方式依赖软件的新设备/技术。随着软件和设备变得越来越复杂，越来越依赖软件的安全功能与功能功能进一步交织在一起，造成严重灾难的机会也急剧增加。在医疗领域尤其如此，在医疗领域，为了挽救生命和/或提高生活质量，有巨大的压力将新设备推向市场，但如果这些设备使用不安全，可能会造成巨大的伤害。 软件的监管制度通常依赖于检查是否使用了经批准的开发过程，然后推断该软件不会导致器械故障。可以预见的是，这并不奏效。多年来，安全案例在英国的多个监管领域都是强制性的。安全案例迫使制造商提供与制造产品相关的证据。我们开始看到保证/安全案例在北美的监管制度中发挥更突出的作用。美国FDA在监管医疗器械方面采取的方法尤其如此。保证案例基于促进对器械有效性和安全性的满意置信度的假设。这些假设涉及尚未解决的基本问题。我打算对标准进行逆向工程，以发现其中包含的隐含保证案例。我还计划比较哲学和形式方法的论证方法，以帮助创建一种方法，通过这种方法可以评估支持论点。我的研究重点是确定：i）如何构建和评估有助于系统整体可靠性和安全性的索赔和子索赔; ii）如何在满足索赔的背景下评估论点的合理性; iii）如何将不同类型的联合收割机证据组合起来，以提高索赔得到满足的信心;以及iv）如何使用规定的结构来组织保证案例，这将有助于认证机构能够积累评估保证案例的经验。

Software is an enabling technology so good that there is now almost no new device/technology on the market that does not depend on software in some way. As software and devices become increasingly complex and safety features, relying more and more on software, get further intertwined with functional features, the chance of creating serious disasters also dramatically increases. This is especially true in the medical domain where there is tremendous pressure to get new devices onto the market in the hope of saving lives and/or improving quality of life, but where those same devices can cause immense harm if they are not safe for use. Regulatory regimes for software usually rely on checking that an approved development process was used, and then infer from that the software will not contribute to a device failure. Predictably this has not worked well. Safety Cases have been mandatory in several regulatory domains in the UK for many years. Safety cases force manufacturers to provide evidence related to the manufactured product. We are starting to see assurance/safety cases assuming a more prominent role in regulatory regimes in North America. This is especially true in the approach taken by the US FDA in the way they regulate medical devices. Assurance cases are predicated on assumptions that promote satisfactory confidence in the efficacy and safety of the device. These assumptions relate to fundamental issues that have not yet been addressed. I intend to reverse engineer standards to discover the implicit assurance cases embodied therein. I also plan to compare approaches to argumentation from philosophy and formal methods to help create a means by which supporting arguments can be evaluated. My research focuses on determining: i) how to construct and evaluate claims and sub-claims which contribute to the overall dependability and safety of the system; ii) how to evaluate the soundness of arguments in the context of satisfying claims; iii) how to combine evidence of different types to contribute to the confidence that a claim is satisfied; and iv) how to organize the assurance case using a prescribed structure, which will facilitate certifiers being able to build experience in evaluating assurance cases.