Quantification of Surgical Technique Using an Inertial Measurement Unit

Abstract

Quantifying the algorithmic information content of hand motion patterns during a surgical task enables the comparison of different groups, novice and expert. Previously, we have shown that the information content/complexity of hand motion patterns during the surgical skill/subtask of knot tying reduces with increased expertise (J Surg Educ 2012;69:306-310). We therefore hypothesized that the information content/complexity of motion patterns would also reduce with expertise during a more complex surgical task of a bench model venous anastomosis. A custom inertial measurement unit was used to record 6-degrees of freedom hand motion from the right hand during a low-fidelity bench model venous anastomosis. Data were obtained from 2 groups as follows: novice (surgical residents, postgraduate year 1 and 2) and expert (attending surgeons). Each data set from the surgical task was processed into a symbolic time series from which the algorithmic entropy measure Lempel-Ziv complexity was calculated. A Student t test was used to test whether the 2 groups were sampled from the same population when using this metric, applying a P value of 0.05 to reject the null hypothesis. The expert surgeons had more complex patterns of motion compared with novice surgeons during this surgical task. This was statistically significant using the Lempel-Ziv complexity metric (P = 0.01). The hypothesis that increased surgical experience would reduce the complexity of hand motion patterns was not found to be true during this surgical task, and the opposite was found to be true. An alternative hypothesis that would combine the previous finding of a reduction in complexity of a subtask with this study's finding of increasing complexity of the whole task with increasing surgical expertise could be due to a hierarchy of pattern structure: experts use simpler subtask motifs in more complex and denser patterns.