Abstract
Formal surgical skill assessment and critical path analysis are not widely used in orthopaedic surgical training due to the lack of technology for objective quantification, reliability, and the discrimination insensitivity of existing methods. Current surgical skill assessment methods also require additional instrumentation, cost and time. Such problems can be overcome by a novel method that records the motion of surgical instrumentation for the purposes of documentation, surgical-skill assessment, and safety analysis. This method uses an existing computer-aided-orthopedic-surgery (CAOS) navigation system and does not compromise its functions of real-time tracking, rendering, or simulation. The stored data allows realistic playback in 3D of the complete bone cutting/refining process. This concept and its sensitivity were previously tested and validated using a robotic arm as a reliable actuator for a surgical instrument moving in controlled paths. In this study, the system was used to evaluate the surgical skills of actual orthopaedic residents in a hospital/lab setting.
Two chief orthopaedic surgery residents participated in the experiment. Each one cut all five distal cuts on four synthetic (right) femurs to accommodate the same femoral implant using NoMiss, an in-house built system for Navigated Freehand bone cutting. The motion of the surgical saw was recorded in real time by NoMiss during the whole procedure, but the real purpose of the experiment (and the recording) was not revealed to the residents until the end of all tests. Based on the data recorded by the navigation system, the following parameters were analyzed: cutting time, area-of-the-cut/time ratio, trajectory of the saw, errors in distance off the plane as well as errors in roll and pitch angles.
While no significant difference among the two subjects was found in bone cutting time (mean 531s vs. 642s, p=0.099), subject 1 (S1) was faster than subject 2 (S2) in total time, which included cutting, reshaping of the bone, and implantation (mean 719s vs. 958 s, p=0.035). Area-of-the-cut/time ratio revealed higher (not significant) proficiency for S1 compared to S2 (mean 16 mm2/s vs. 13 mm2/s, p=0.084). Nevertheless considering individual cuts, there was significant difference in the posterior chamfer cut (mean 9 vs. 5 mm2/s, p=0.015). The analysis of the trajectory of the saw showed less conservative motion (and less consistency) for S1 than for S2 (average total length of trajectory 8.6m (sd=2.1m) vs. 8.1m (sd=0.4m), as well as larger paths in between cuts (average 39% vs. 33% of the total trajectory).
The system/method was able to characterize different subjects without additional instrumentation, cost, time, awareness of or distraction to the user. Slightly better performance was detected for S1 compared to S2 presumably signifying superior skills. The main differences in this case appeared in the cutting of the chamfers, which might be considered the trickiest of the distal cuts in a navigated freehand cutting environment. A larger number of subjects with a wide level of expertise should be analyzed under similar conditions to establish quantitative acceptance limits (e.g. numerical determination for pass/fail criteria).
Correspondence should be addressed to Diane Przepiorski at ISTA, PO Box 6564, Auburn, CA 95604, USA. Phone: +1 916-454-9884; Fax: +1 916-454-9882; E-mail: ista@pacbell.net
