Sean Hägen, Founding Principal and Director of Research & Synthesis, BlackHägen Design02.07.22
Due to the pandemic, research and development of medical devices has been dramatically impacted with regard to conforming to a usability engineering process informing design. Many usability evaluation methods requiring in-person protocols were suspended, while other methods that do not were utilized more than usual and enhanced. One method in particular saw a rapid evolution in its technology and application—virtual reality (VR). The technology has been driven by the high-volume gaming industry and, therefore, benefitted from rapid advancement in affordability and ease-of-use of the development tools. Although VR has been applied readily to training, it can also be valuable upstream in early device development. The methodology it can augment or even replace is configuration modeling that is intended to inform system architecture, especially relative to user interfaces (UI).
General Study Methodology
The predicate methodology involves designing study models that enable the user to evaluate—at a feature level—disparate user interface configurations prior to system engineering, design for manufacturability, and industrial design. This complements other system-level design inputs that impact requirements like manufacturability, service, sustainability, and power management by providing insights born from the user’s needs and context of use. These study models are not iterations of a design direction under development, rather, they are more of a mock-up or prop.
The methodology is intended to provide a user with the means to experience a set of alternative user interface configurations and demonstrate their preferred features. Each alternative model is intentionally configured to present UI features in a different way. For example, the overall UI is configured in a portrait or landscape orientation, handles are positioned at different attitudes, disposable sets are loaded from different approaches, etc. This approach enables study participants to isolate and determine what their preferred feature configurations are, and indicate why they are preferred in the context of use without choosing an overall design solution.
The downside of this methodology is each study model has to be engineered in CAD to provide the study participant with an appropriate level of interaction in order to experience the UI in a meaningful way. For example, features that open and close will need hinges engineered. Although the models are not going through the extensive process of designing for manufacturability, it still takes significant time to engineer such features. Furthermore, it then has to be fabricated and debugged. All this time and CAD data is effectively “throw away engineering,” because it is not representing the actual design direction yet.
The development time for a study model can be drastically reduced, however, by providing the study participant with a virtual means of evaluating the study models. Evaluating virtual models eliminates the need to engineer all the functional details and, of course, nothing needs to be fabricated. Instead, functions are assigned kinematic characteristics. So, if a display is to tilt and swivel, the extent of those kinematics is defined without designing the mechanisms.
Simulating the Contextual Environment
There is an upfront investment in creating the environment of use, although it is reusable. If the fabricated methodology is utilized, the study would require a high fidelity, simulated operating room (OR) to be utilized for context. In this example, VR study configurations for a surgical robot are to be evaluated and the contextual environment is an OR. All the people and equipment are critical for the realistic evaluation of the system. For instance, the OR environment should have surgical lights, monitors, patient table, booms, IV poles, an anesthesia cart, infusion pumps, etc. This level of contextual realism enables the following parameters to be assessed:
Using the surgical robot development example, study participants (potential users and designers) can experience how certain configurations and feature layouts are impacted by the context of use. A simple evaluation in CAD of the robot’s impact on the context of use is inadequate; even though it is 3D, it’s still static and not experiential. In Figure 1, it appears the robot integrates into the sterile field appropriately. However, this does not take into account the device’s kinematics and all the environmental context that may have a significant impact on usability and clinical integration. This is why an interactive evaluation of configuration models is so important.
Developing the VR configuration model study starts with developing CAD models for the alternative system configurations. All user interfaces should be represented and differ between models, such as handles, power button, emergency stop, break controls, power cord, data cable, etc. In this example, the differentiating features may not be just user interfaces but also general architecture that can impact use-case scenarios like positioning the robot to the OR table/patient, access to the patient and other equipment, and exchanging instruments. The general architecture variations could be the shape of the robot base or the structure of the cantilevered arm or arms. The models should all be rather schematic in their design so as to not bias feature preferences by the overall aesthetic.
Use-Case Scenario Planning
After building the digital OR environment, the next step is planning a use-case scenario—in this case it will be a robotic-assisted surgery (RAS). One approach is to storyboard the scenario to determine what actors need interactive features, or kinematics. This is likely to be an iterative process since evaluating the design in context will reveal unintended interactions. The storyboard (Figure 2) should have an associated script that offers step-by-step details of the scenario so the developer can adjust the environment and actor kinematics accordingly. Participants will then be able to don the goggles and gloves to try out the experience; keep in mind, a room the same size as the area you want the participant to explore will be required.
A study guide is helpful so the moderator can know when to prompt the participant for feedback and have a list of questions ready. Having the participant use a think-aloud approach is an appropriate method as well. As the participant responds to prompts from the moderator and executes tasks, insights regarding why one configuration is preferred over another will become apparent. Questions like: Is there enough space for the surgical team to access the patient and exchange instruments with the robot? If the RAS has to convert to an open procedure, how would the team interact with the robot?
Conclusion
VR has proven itself and the technology has advanced over the years since its introduction. Applying this methodology to configuration alternatives is very effective in generating design inputs for foundational system design decisions. With the latest advancements in VR, it is possible to build much lighter CAD models, without engineering the mechanisms and eliminating the need for physical model fabrication. Using a VR version of a configuration model study can reduce development time and costs by enabling faster iterations.
Since founding BlackHägen Design in 1995, Sean Hägen has been the principal investigator for design research, within both institutional and home environments, across 20 countries. He has published numerous articles on usability research and design, as well as contributed to domestic and international standards for usability development processes, as a member of the AAMI HE committee (ANSI HE75:2009 and IEC 62366-1:2015). In this role, he is a contributing author for AAMI TIR 50 and 59 as well as the revision of ANSI/AAMI HE75. He was the IDSA Medical Section chair for eight years and sat on the Board of Directors for two terms. He currently chairs the IDSA Patient Safety Task Force.
General Study Methodology
The predicate methodology involves designing study models that enable the user to evaluate—at a feature level—disparate user interface configurations prior to system engineering, design for manufacturability, and industrial design. This complements other system-level design inputs that impact requirements like manufacturability, service, sustainability, and power management by providing insights born from the user’s needs and context of use. These study models are not iterations of a design direction under development, rather, they are more of a mock-up or prop.
The methodology is intended to provide a user with the means to experience a set of alternative user interface configurations and demonstrate their preferred features. Each alternative model is intentionally configured to present UI features in a different way. For example, the overall UI is configured in a portrait or landscape orientation, handles are positioned at different attitudes, disposable sets are loaded from different approaches, etc. This approach enables study participants to isolate and determine what their preferred feature configurations are, and indicate why they are preferred in the context of use without choosing an overall design solution.
The downside of this methodology is each study model has to be engineered in CAD to provide the study participant with an appropriate level of interaction in order to experience the UI in a meaningful way. For example, features that open and close will need hinges engineered. Although the models are not going through the extensive process of designing for manufacturability, it still takes significant time to engineer such features. Furthermore, it then has to be fabricated and debugged. All this time and CAD data is effectively “throw away engineering,” because it is not representing the actual design direction yet.
The development time for a study model can be drastically reduced, however, by providing the study participant with a virtual means of evaluating the study models. Evaluating virtual models eliminates the need to engineer all the functional details and, of course, nothing needs to be fabricated. Instead, functions are assigned kinematic characteristics. So, if a display is to tilt and swivel, the extent of those kinematics is defined without designing the mechanisms.
Simulating the Contextual Environment
There is an upfront investment in creating the environment of use, although it is reusable. If the fabricated methodology is utilized, the study would require a high fidelity, simulated operating room (OR) to be utilized for context. In this example, VR study configurations for a surgical robot are to be evaluated and the contextual environment is an OR. All the people and equipment are critical for the realistic evaluation of the system. For instance, the OR environment should have surgical lights, monitors, patient table, booms, IV poles, an anesthesia cart, infusion pumps, etc. This level of contextual realism enables the following parameters to be assessed:
- Footprint integration (relative to people and other equipment)
- Dimension priorities (height, width, and depth)
- Line of sight (relative to people and other equipment)
- Dynamic UI interferences (kinematics)
- Collisions with other equipment
- Access to user interfaces
- Component configuration
- User interface layout
- Maneuverability
Using the surgical robot development example, study participants (potential users and designers) can experience how certain configurations and feature layouts are impacted by the context of use. A simple evaluation in CAD of the robot’s impact on the context of use is inadequate; even though it is 3D, it’s still static and not experiential. In Figure 1, it appears the robot integrates into the sterile field appropriately. However, this does not take into account the device’s kinematics and all the environmental context that may have a significant impact on usability and clinical integration. This is why an interactive evaluation of configuration models is so important.
Developing the VR configuration model study starts with developing CAD models for the alternative system configurations. All user interfaces should be represented and differ between models, such as handles, power button, emergency stop, break controls, power cord, data cable, etc. In this example, the differentiating features may not be just user interfaces but also general architecture that can impact use-case scenarios like positioning the robot to the OR table/patient, access to the patient and other equipment, and exchanging instruments. The general architecture variations could be the shape of the robot base or the structure of the cantilevered arm or arms. The models should all be rather schematic in their design so as to not bias feature preferences by the overall aesthetic.
Use-Case Scenario Planning
After building the digital OR environment, the next step is planning a use-case scenario—in this case it will be a robotic-assisted surgery (RAS). One approach is to storyboard the scenario to determine what actors need interactive features, or kinematics. This is likely to be an iterative process since evaluating the design in context will reveal unintended interactions. The storyboard (Figure 2) should have an associated script that offers step-by-step details of the scenario so the developer can adjust the environment and actor kinematics accordingly. Participants will then be able to don the goggles and gloves to try out the experience; keep in mind, a room the same size as the area you want the participant to explore will be required.
A study guide is helpful so the moderator can know when to prompt the participant for feedback and have a list of questions ready. Having the participant use a think-aloud approach is an appropriate method as well. As the participant responds to prompts from the moderator and executes tasks, insights regarding why one configuration is preferred over another will become apparent. Questions like: Is there enough space for the surgical team to access the patient and exchange instruments with the robot? If the RAS has to convert to an open procedure, how would the team interact with the robot?
Conclusion
VR has proven itself and the technology has advanced over the years since its introduction. Applying this methodology to configuration alternatives is very effective in generating design inputs for foundational system design decisions. With the latest advancements in VR, it is possible to build much lighter CAD models, without engineering the mechanisms and eliminating the need for physical model fabrication. Using a VR version of a configuration model study can reduce development time and costs by enabling faster iterations.
Since founding BlackHägen Design in 1995, Sean Hägen has been the principal investigator for design research, within both institutional and home environments, across 20 countries. He has published numerous articles on usability research and design, as well as contributed to domestic and international standards for usability development processes, as a member of the AAMI HE committee (ANSI HE75:2009 and IEC 62366-1:2015). In this role, he is a contributing author for AAMI TIR 50 and 59 as well as the revision of ANSI/AAMI HE75. He was the IDSA Medical Section chair for eight years and sat on the Board of Directors for two terms. He currently chairs the IDSA Patient Safety Task Force.