Human factors in image guided surgical simulator training : Components, visual-spatial and haptic aspects

Sammanfattning: New image-guided surgical techniques such as arthroscopy and laparoscopy are difficult to learn and master. To handle these techniques, there seems to be a need for higher-level abilities such as visual-spatial abilities. Difficulties in human-system interaction could be studied with a human factors perspective in a similar way as already has been done in high-reliability organizations (HROs) such as aviation. The main goal of human factors is to improve human interaction with systems to: improve performance, reduce errors and increase user satisfaction. Health care in general and surgery in particular share high-risk activities with the HROs but unfortunately lack a sufficiently high level of safety. Training in a virtual reality (VR) simulator could help the surgeon to overcome new demands and difficulties through training. There is a relative lack of haptic feedback in image guided surgery compared to open surgery. More precise, more consistent and much safer procedures could be performed if haptic feedback (force feedback and/or tactile feedback) would be integrated into image guided surgery. A major difficulty is also to extract 3-dimensional (3-D) information from 2dimensional (2-D) images on the monitor. A strive to implement these simulators into a surgical curriculum for a standardized, objective and validated surgical skills training and assessment cannot be done without investigating these human factors aspects further. The general aim of this study was to investigate the influence of three important but scarcely studied aspects of human factors in image guided surgery: components in a simulator training context, visual-spatial ability and haptic feedback. In Paper I, medical students (novices) participated in a study with self-controlled design to analyze the effect of training in a simulator with anatomical graphics and haptic feedback. We also compared the results in this simulator with the established MIST simulator with an artificial environment and no haptic feedback. There was a significant improvement after 1 hour of training regarding time, movement economy and total score. The results in the two simulators were highly correlated. In Paper II, novices participated in a study with experimental design: a randomized, posttest only, control group design. We analyzed the importance of visual-spatial components in a simulator context. Performance in a virtual arthroscopy simulator did not improve after training in other simulators with different visual-spatial components. The performance in the arthroscopy simulator was not significantly correlated with the MIST simulator. In Paper III, novices participated in a study with self-controlled design, which was modified to provide a combination of self-controls and concurrent experimental treatments. We analyzed the correlations between high-level visualspatial ability test scores and performance measures in two simulator tasks differing in spatial complexity. The performance in the most spatial complex simulator task significantly correlated with high-level visual-spatial ability while such ability-performance relationships were not found for the less spatial complex simulator task. In Paper IV, surgical residents participated in a randomized study with cross-over design to analyze the effect of early exposure to haptic feedback during early phases of image guided surgical simulator training. After two hours of training the group who had started with haptic feedback performed significantly better on two diathermy tasks compared to the group who had started without haptic feedback. The results show that the use of surgical simulators as a pedagogical tool in medical student training is promising with steep learning curves. Performance in one visual-spatial context (VR arthroscopy) did not improve after one hour of training on virtual tasks with different context. However, it cannot be excluded, that experienced trainees could improve their performance, as perceived similarity between different situations is influenced by many psychological factors, such as the knowledge or expertise of the person performing the task. A training curriculum in advanced image guided surgical simulators should aim at task and context specificity in order to maximize transfer. However, construction of simulator tasks should be based on task analysis to select essential components and invariant relationships for transfer to the real task. High-level visual-spatial ability is important to possess for surgical novices in the early training phase of a visual-spatial complex simulator task. Our findings further indicate that haptic feedback enhances performance in the early training phase of skill acquisition in image guided surgical simulator training. This is important to consider when designing a multi-modal simulator based curriculum for image guided surgical training.