Augmentation of Nonvisual Feedback in Surgical Simulation and its Effect on Training Outcomes

Medicine is arguably one of the most difficult career paths one can choose. In no small part, this is due to the intensive education and training required, which accompanies one throughout one’s career. All specialties have their own challenges and ways of addressing them. One of the newer methods is using virtual reality simulation. In the field of minimally invasive surgery, with improvements in graphics, cheaper computer components, and influence from fields such as aerospace to standardize training, the use of surgical simulation training has led to several successful commercial products aimed at helping to acquire and maintain skills efficiently and effectively. However, despite an increase in their variety, simulators may not be capable of providing training for crucial elements of surgery. Moreover, they may not be using pedagogically effective methods to train users. One of the challenges in the implementation of a simulator is the selection of the medical procedure components that are replicated in the virtual environment. The implementation of a whole patient and surgery may not be technically feasible, as the human body is a complex assembly, and the operating room is a complex setup. Therefore, partial task training on critical components has become a practical approach, with training focusing on smaller anatomical models and single components of the medical procedure. The decision on which components to select and why is not trivial. To achieve this goal, this work began by examining the operating room and the procedures therein using human factors frameworks and methodologies. The analysis found areas of surgery, specifically in the example field of laparoscopy, that combined cognitive, social, and physical challenges. These identified areas are prone to mistakes and complications that could cause severe damage to the patient. In medical education, these areas should be the focus of additional training efforts. Parallel to these efforts to analyze operations and operating rooms, the state of current commercial surgical simulators was also examined from a human factors’ perspective. The commercial simulators available on the market were developed at different times and for different clients. The strengths and weaknesses of each simulator could significantly alter the training outcomes of surgeons. In this work, simulators were examined within a novel human factors’ framework to determine areas for improvement. No single simulator was found to be with no such areas. Two areas were found to be common across all simulators: cognitive overload during training and low fidelity of haptic feedback. This thesis focused on addressing these areas by developing, optimizing, and validating methods to reduce trainees’ cognitive load and improve haptic feedback. The focus was on commercially viable implementations that could be both efficient and effective. To address the issue of low-fidelity haptic feedback, audio-vibratory feedback was implemented with surgical instrumentation to assess the ability to amplify the range of perceived roughness while a tool interacts with a physical surface. Inspired by orthopedic surgical simulators that face the challenge of having only one anatomical model available for training, an expanded range of roughness allows surgeons to learn to pay attention to surgical cues that can help improve surgical outcomes. Three studies examined how auditory, vibratory, and combined audio-vibratory cues expanded the range of perceived roughness. These cues were played over low-cost actuators and headphones. All methods were found to be effective. Among them, the vibratory cues were the most effective, with an increase of up to 171%. Additionally, these studies were performed with a scissor grip—an underexamined grip style in haptic research, but a very common one in the operating room. The results indicated that a low-cost method was effective in expanding the range of roughness when interacting with a surface using a tool in a scissor grip. To explore ways to reduce trainees’ cognitive load, various audio and visual cues were examined during an arthroscopic diagnostic tour of the knee. These cues were implemented on a surgical simulator and examined in a study involving novices. The cues provided guidance to the participants as they trained and practiced and were implemented in a cost-effective manner using visual cues on computer monitors or commercially available headphones. The cues were selected to utilize sensory modalities that may not be used in training simulators. Once again, it was found that a low-cost method was effective. Specifically, verbal cues were found to lead to a significantly lower overall cognitive load compared to other forms of feedback, including visual feedback and a control group receiving no feedback. The efforts to address the issues of low-fidelity haptics and cognitive load were examined in a multiday study in which novices practiced a diagnostic knee tour on an arthroscopic knee simulator. Vibratory feedback for texture augmentation and user guidance cues were implemented and assessed. This novel approach included applying technologies inspired by human factors–based analyses to surgical simulation. The surgical simulator quantitatively and automatically assessed the novices’ performance, while additional feedback values related to overall workload and system usability were investigated. In contrast to the previous findings suggesting that the verbal cues reduced the workload compared to other feedback cues and the control group without feedback augmentation, this study indicated that the simple act of practicing was the most effective method for improving skills and reducing the workload. The results further suggested that additional feedback cues, including vibratory feedback, could hinder training and affect performance, indicating that the cognitive overload issue was not successfully addressed. In conclusion, the human factors–based approach to examining both surgery and surgical simulation identified areas that could be improved. Potential low-cost solutions to these areas were implemented, assessed, and partially validated. The implementation of low-cost haptic feedback amplified the range of perceived roughness when using a tool interacting with a surface. Furthermore, using alternative sensory modalities to guide users in training helped reduce the cognitive workload. However, simply adding these forms of feedback was not more effective than the act of practice itself, suggesting that their implementation in a longer training scenario poses a considerable challenge. Future studies could examine how more experienced users react to these forms of feedback and expand the forms of feedback to address these sensory modalities and areas of perception (e.g., softness or elasticity)