Virtual suturing for training in vascular surgery

Today's health professionals are facing a crisis in training needs: on one hand, the working hours of junior doctors and experience at the operating table are being reduced; on the other, patients are growing ever more critical and litigious. VR simulators may be able to provide a solution, but whilst hardware costs have fallen in recent years, they are still expensive when compared to conventional methods and few have been adopted. The challenge for researchers has been to create realistic, but affordable, surgical interfaces and to provide convincing assessments of the resulting systems.

There are several common forms of assessment in the surgical simulation literature. The most popular would seem to be the construct validity test, in which experts' performance is contrasted with that of novices. Although this method provides a useful check, it is argued that the results are often unreliable, given the short-term nature of the test and the difficulty of separating practice and learning effects. Moreover, a wider literature search shows that consistency and persistency of performance are much more highly respected in eg. military and aviation contexts.

A design for a virtual suturing simulator, dubbed FESTIVALS, is proposed which is based upon principles established in motor psychology over the last few decades. In particular, practice variability is promoted by requiring the user to employ both hands in facilitating access. Also, a delayed feedback schedule is introduced to provide feedback on errors.

The Finite Element Method is adapted to build an accurate deformation model to support bimanual working and real-time haptic display. Evaluation of this system showed that the FESTIVALS system possessed good training and retention characteristics. In addition, a usability study collected feedback from clinicians which showed a generally favourable response and allowed several recommendations for future development.

By considering discrete phases of the suturing data collected in these evaluations, it is also possible to show that construct validity held for several metrics. This is of particular interest because it appears to show that experts were much more capable of planning specific movements in advance, suggesting a more highly developed technique for error-control. This finding led to a novel proposal for an error-correction model of expertise.