Abstracto
- Medical devices are becoming more complex than ever, as do the networks they pertain to. The current trends in MedTech manufacturing complicate the work of a systematical and comprehensive RM process. At present, manufacturers implement many different, but exclusively document-based RM approaches. The work described in this thesis focuses on conceiving and validating a model-based risk management system that enables RM operators to overcome the endemic deficits of the document-based approaches. This shall be achieved by the formalization of RM steps, the role-based separation of procedures in computation and human action and by providing an RM system that enables an iterative RM process during the entire product lifecycle for all stakeholders. The research approach adopted mainly comprises an extensive study of relevant literature, reasoning and an implementation of applied research, carried out in a case study. A systematical and comprehensive RM for medical devices can be accomplished with a MBR concept. The iterative system design separates the operational and computational procedures in the MBR Core from the actions of the experts and stakeholders. A universal API processes all changes to and all documents generated from the MBR core. The sequential use of human expertise and computational rigor allows for the integration of document-based RM methods and techniques that are broadly accepted in the industry. A main factor for comprehensive RM results is the computerization of the identification of critical characteristics. The elements of the physical product are tagged with approved industry classifications. A novelty in product modeling is the utilization of an own block class for interactions instead of relational elements; it has proven to be functional and valuable in implementation. The software implementation of the system is shown on a demonstrator level. The validation of MBR in a case study applying two RM methods on two similar complex medical device systems, advanced prototypes of automated stem cell platforms, has shown the potential to drastically reduce the deficits endemic to document-based approaches. The augmentation of established RM techniques with legacy information can improve identification of critical characteristics. However, the case study also showed that the disposition of the panelists is key to the success of the concept. While proficient users of tools in MBSE explored the full potential in the utility tests, panelists with a basic to intermediate knowledge of MBSE, showed strong reservations against accepting “advise” from the augmented graphical model. In total, the model-based RM approach can be a significant contribution to the improvement of RM processes for complex medical devices and, in general, to the dissemination of risk-based thinking throughout all lifecycle stages, provided that all stakeholders engage in an openminded and interdisciplinary process.