Video analysis of endoscopic cutting task performed by one versus two operators

Abstract

Background

In the performance of complex laparoscopic tasks, one question is whether the task should be distributed between two operators or accomplished bimanually by one operator. The authors hypothesized that superior task performance results when two operators work collaboratively in a dyad team as opposed to one operator performing the task bimanually. Furthermore, in a visually misaligned condition, the performance of a team will be more robust than that of a single operator working alone.

Methods

The suture-cutting task was performed by 24 right-handed subjects in a mock surgical setup using a laparoscopic grasper and a pair of laparoscopic scissors. The cutting task was performed by 8 subjects bimanually (using both limbs) and 16 paired subjects unimanually (using their preferred limbs). The image of the work plane was displayed either vertically or superimposed over the work plane. In half of the conditions, the camera was rotated 45°, causing misalignment between the actual and displayed work planes. Movements were videotaped. Important movement events were identified and used to subdivide the task into subtasks. Durations of the subtasks and attempts for grasping and cutting were analyzed using a mixed-design multivariate analysis of variance (MANOVA).

Results

For a number of subtasks, the dyad group showed shorter durations than the bimanual group. The 45° rotation of the camera degraded both bimanual and dyad performance, resulting in prolonged movement times for all subtasks. The learning process was facilitated by the superimposed display in that grasper and scissor reaching times improved over trials, as compared with the vertical display.

Conclusion

The results indicate the superior role of team collaboration, as compared with the single operator, in a complex remote manipulation such as a laparoscopic cutting task. This enhanced task performance is achieved because of the larger capacity for information processing. These results may have some relevance for optimizing performance of endoscopic surgery.

Appendix

Temporal data were recorded from the VCR counterclock, which was synchronized with the video recordings. The counter was based on 60 min, 60 s, and the number of frames (30 frame/s time intervals), with the counter increasing as the time duration increased. Thus, once the counter read 00:00:29, the next frame increased the counter to read 1 s, indicated by 00:01:00. When counter read 00:59:29, the next frame increased the counter to 01:00:00. In contrast, when temporal data are inputted to Microsoft EXCEL, in which data are sorted and analyzed, EXCEL requires data to be in the format of 60:60:60 (conventional time format) for the software to “recognize” it. Therefore, when inputting VCR counterclock data into EXCEL, we multiplied the last two digits by 2 to make it a ratio of 60 rather than 30. Time durations between defined events were then calculated by subtracting the respective time codes and converted them to seconds.

The conversion steps are as follows:

VCR counterclock = 00:31:27

Total time duration = 00:31:54

= 31 seconds + 54/60 seconds

= 31.9 seconds