Rapidly evolving technological progress and automation in the field of medical devices and systems (e.g. in the field of ambient assisted living or in clinical context) not only lead to an enhancement in efficiency and effectiveness concerning therapeutic results but also to a change of the Human-Machine-Interaction characteristics. Increasing functionality and complexity of new technical medical equipment can implicate deficiencies in the use process, bringing along high potential for hazardous human-induced failure implicating higher risk for the patient and the OR-team. Studies confirmed the important role of human factors in safety aspects of Human-Machine-Interaction in clinical environment.
A risk analysis of the human-machine interaction (MMI) (e.g. within the framework of the usability engineering process according to IEC 60601-1-6 or IEC 62366) is in many cases carried out too late or not sufficiently. One reason for this is that currently used risk analysis tools (e.g. FMEA, FBA) and formal-analytical usability evaluation methods are limited with regard to their modeling structure and therefore often cannot be used for detailed mapping and analysis of complex human-machine interaction. In particular, the estimation and evaluation of human-induced errors are problematic.
In this context, the Chair of Medical Engineering of the RWTH Aachen University has developed a new method/software tool in order to support designers as well as risk assessors. The mAIXuse risk analysing tool uses formal, normative models to predict user-, interaction- and system-behaviour. With the help of specific task categories and the integration of temporal relations, the use-process-state can be modelled, even with complex (graphical) user interfaces. On the basis of different human error taxonomies, a systematic failure analysis is associated, in order to identify potential errors in human information processing and conflicts in the use of mental resources. The method can be used from early developmental stages up to validation process with an existing prototype.
The methodology and the corresponding software tool mAIXuse are based on a twofolded model-based approach. Within the initial task analysis, the usability investigator gains a systematic overview of the high-level operations of the user, the system and the interactions which are required to achieve a specific objective within the task fulfilment. As part of the graphical notation, it is possible to describe different task categories and types as well as various attributes and objects. Within the mAIXuse modelling, there are five task categories. Human-system-interaction tasks are subdivided into the classical information processing sequence “Perception”, “Cognition” and "Action" in combination with an analysis of human-human-interactions (“Human-Human-Interaction“).
To detect potential errors early in the user interface design process a table of potential human errors (classification according to its external appearance) is given. Furthermore, potential causes of the failures are defined for the cognitive information processing and the perception process (information acquisition). A possible cause of error e.g. lies within the cognitive information processing. For this purpose, a failure checklist has been derived on the basis of the cognitive modelling system called the Generic Error Modelling System. The Generic Error Modeling System (GEMS) is a framework for understanding error cause types and designing error prevention strategies. Additionally, on the basis of relevant standards (e.g. 9241-12) a checklist for the analysis of potential error causes in the acquisition and perception of information has been developed.
The mAIXuse method/tool has been investigated in comparison to classic risk analysis methods concerning efficiency, effectiveness and usability aspects in the framework of the risk analysis with an existing planning and navigation system for hip resurfacing in orthopaedic surgery. This tool is currently being developed at the Chair of Medical Engineering. The mAIXuse tool clearly outperforms the Process-FMEA (Failure Mode and Effects Analysis) in terms of effectiveness and efficiency (objective evaluation criteria) and amongst others intuitiveness, learnability and user satisfaction (subjective evaluation criteria).
In 2010 Dipl.-Ing. Armin Janß as one of two awardees has been decorated with the Walter Masing Prize (endowed with 10.000€) of the German Society of Quality (DGQ) for the development of the mAIXuse method/tool.
The mAIXuse tool has been successfully applied in research and development with different industrial partners (e.g. Aesculap AG & Co. KG, Biotronik GmbH & Co. KG., BEGER DESIGN, SurgiTAIX AG). In the framework of our services in the field of risk management and usability engineering we offer seminars and workshops as well as usability evaluation within the development process of enterprises.
Please send us a note if you are interested in further informations or if you want to apply the mAIXuse method/tool on your products.