gms | German Medical Science

Objective: ECG interpretation is a clinically highly relevant, yet difficult to master skill with frequently poor performance results. Medical students are currently taught without a pre-existing evidence-based gold standard, relying instead on personal experience or schemata. It is largely unknown whether ECG teaching strategies are consistent with actual expert viewing patterns during ECG interpretation. The present study investigated the effects of eye-tracking videos with cued retrospective reporting (CRR) of an expert interpreting ECGs on medical students’ ECG interpretation skills.

Methods: Realisation was in 3 phases.

1.

Eye-tracking video generation: A cardiology expert wearing a head-mounted eye-tracker silently interpreted 15 ECGs of common and consequential diagnoses, resulting in a video visualizing the expert gaze patterns (see figure 1 [Fig. 1], cf. [1]).

2.

CRR audio commentary creation: The expert watched their own gaze behaviour on each ECG and retrospectively verbally explained their visualized gaze, resulting in a 9-minute video containing both gaze visualization and simultaneous verbal explanation.

3.

Student learning intervention: All students (N=91) participated in a previously validated [2], [3] online training session (5h; pre-test on theoretical ECG knowledge, 4 cases on complex clinical ECG scenarios with learning feedback, post-test on quick practical ECG scenarios). The intervention group (INT; n=47) further received the expert eye-tracking CRR as additional learning stimulus and was evaluated against the training as usual group (TAU; n=44) for differences in ECG interpretation skills gain (0-100%).

Results: Analysis showed a 1.25% greater knowledge gain of INT (M=11.05, SD=9.80) vs. TAU (M=9.80, SD=10.76), which did not reach statistical significance, t(89)=0.58, p=.57. When regressing post-test scores on experimental group, sex, age, years in education, prior medical professional qualification, university semester, prior cardiology residency, self-rated ECG experience, and pre-test scores (R²_total=.45), pre-test scores contributed significantly and explained most of the variance (R²_change=.40; B=0.63, p<.05*10^-8). Only self-rated ECG experience further added significantly (R²_change=.02; B=0.04, p<.03), leaving 3% of variance explained by all other nonsignificant predictor variables combined.

Discussion: Results showed a small initial advantage of the CRR teaching method after only 9 minutes of presentation. That, however, was too small to carry through given the current sample size and effect of prior theoretical ECG knowledge. Participant feedback also suggested that pausing videos at crucial points instead of a continuous loop could further improve the usefulness of CRR.

Take home message: Preliminary findings indicate that being let in on an expert’s real-time gaze pattern and associated thought processes during ECG interpretation may be further developed as a useful learning and teaching tool.