Individual learning curves in bronchoscopy : exploring skill acquisition with a low-fidelity prototype.

The thesis is focused on the design of a low-fidelity, low-budget prototype for flexible endoscopy (EndoProto) to explore bronchoscopic skill acquisition. Twenty-four students executed two basic scope insertion tasks with the EndoProto. First, they trained bronchoscopic skills from an allocentric and then form an egocentric perspective, defined by the visualization through a low-budget endoscopic camera. We made use of a within-subjects design. We focused on the exploration of learning through the estimation of learning curves. No learning curves could be estimated. Instead, we explored individual performance and differences between the allocentric and the egocentric tasks. We applied a multi-level mixed-effects model with a binominal distribution, where trials became exchangeable repeated measures.(1) Results showed group effects for wall contacts conditional on Time-on-Task (ToT). (2) Participants performed better in the allocentric compared to the egocentric perspective regarding the performance parameter wall contacts and ToT. (3) We found an interaction effect of routes conditional on tasks. Results were based on 95% CI. Two experts in bronchoscopy were interviewed for additional insights in the effectiveness of the build prototype.Qualitative and quantitative results were discussed according to learning, performance and the placing of the EndoProto on the spectrum of resemblance.