Fig 1.
Participants trained with three separate training structures prior to performing real-world assessment of acquired knowledge. A) Text-based study tools consisted of a flash card for each item that included two images (full instrument and an enlargement of the distinguishing feature), and a text-based description including item uses, physical characteristics, and alternate names. B) The tablet-simulated environment consisted of items of a given set of instruments presented on the lower half of the tablet screen (4 viewable at once, with the remaining visible by scrolling left/right), and an empty ‘surgeon’s’ tray presented above. Following an auditory request, participants were required to swipe the instrument into the empty tray. Text, then image hints were provided if the correct idem was not identified within a given period. C) For real-world procedures (real-world training + ALL test procedures), pseudo-surgical instruments were arranged in a predefined order on a ‘surgical prep table’ located directly in front of the participant. Following correct identification, instruments were placed by the experimenter on an empty workspace. Incorrect instruments were returned to the prep table.
Table 1.
Performance across training conditions and assessment timing.
Fig 2.
Performance of real-world test procedures immediately and 1-weeek delayed from training.
For all plots, each dot represents an individual participant. A black diamond represents the population mean, with error bars indicating standard deviation. A) Accuracy by training protocol. Real-world-practice resulted in better instrument identification accuracy during retention testing than either tablet-based simulation training or text-based study. B) Response time by training protocol. Less time was required to perform the instrument hand-off following real-world-practice and tablet-based simulation compared to text-based study. C) Errors by training protocol. Differences in the number or errors per incorrect instrument were observed only during the immediate retention test. During delayed retention testing, all protocols resulted in equivalent errors. For all plots, individual subject data are shown by coloured circles, with the group mean and st. dev indicated in black.* p < 0.05, ** p < 0.01, *** p < 0.001, m p < 0.10.
Fig 3.
Instrument identification accuracy as a function of procedural repetition (A/B) and training time (C/D). Solid blue lines indicate the model LOESS fit (span = 0.5). The solid line represents the point estimate at which the LOESS model indicates that performance to be plateaued. The shaded area around this line indicates the 95% confidence interval (CI) around this estimate. Note that while highly overlapping point estimate CIs are observed between real-world and tablet-based training as a function of repetition, these CIs do not overlap as a function of training time, indicating that a performance plateau is reached significantly faster during tablet-based training.