The learning curve for a colonoscopy simulator in the absence of any feedback: no feedback, no learning

Abstract

Background: The hypothesis of this study is that working on the simulator without a structured feedback does not change performance; hence, any effects shown after structured feedback would amount to useful learning of the procedure. The aim was to investigate the learning curve for the HT Immersion Medical Colonoscopy Simulator without any structured feedback. This could then be potentially applied to validate the learning curve on the simulator when structured feedback is provided. There are no previous studies on this matter.

Methods: Candidates were asked to perform colonoscopy on the HT Immersion Medical Colonoscopy Simulator. Modules 3 and 4 were used at random. In total, each candidate was asked to perform five consecutive virtual colonoscopies on the same module. These five episodes were collectively referred to as one trial. A time result of 3,600 sec (1 h) was used to denote perforation. No guidance or feedback was given to candidates before, during, or after each procedure. A total of 26 postgraduate doctors were recruited, including nine research fellows, five preregistration house officers, six specialist registrars, and six consultants. Fourteen candidates recorded five attempts each (i.e., one trial each) on the same module of the colonoscopy simulator (14 trials over 70 episodes). Another 12 candidates recorded five attempts (i.e., one trial each) on two modules of the colonoscopy simulator (24 trials over 120 episodes). Hence, 190 episodes were recorded in total, representing 38 trials.

Results: There was no improvement in performance on the simulator from first attempt to the fifth in the absence of feedback. If there was any initial gain in any measurable outcome, this was lost in subsequent attempts indicating lack of learning. The outcomes measured included time taken to complete the test, percentage of the mucosa visualized, depth of the instrument inserted, and the path length used. The results were statistically significant for all outcomes.

Conclusions: This study demonstrates that in the absence of feedback, it is not possible to improve performance on the HT Immersion Medical Colonoscopy Simulator. Thus, there is no learning curve for the machine. The information from this study is vital for using the simulators in training and assessment because any improvement in learning curves shown after training on simulators can be presumed to be due to learning the procedure and not the simulator.