Relative effectiveness of high- versus low-fidelity simulation in learning heart sounds

Although there are increasing numbers of studies of outcomes of high-fidelity patient simulators, few contrast their instruction with that provided by equivalent low-fidelity, inexpensive simulators. Further, examination of decays in learning and application (transfer) to real patient problems is rare. In this study, we compared the effects of training using a high-fidelity heart sound simulator (Harvey) and a low-fidelity simulator (a CD) on recognition of both simulated heart sounds and those in actual patients.

A pilot study with 10 students was conducted to show the feasibility of the methods and some evidence of modality-specific learning (the Harvey-trained group scored 72% correct on Harvey and 36% correct on CD test examples; the CD-trained group scored 60% correct on both CD and Harvey test examples). A main study was then initiated involving 37 Year 3 medical students from the University of Leeds. They received 1 hour of common instruction, after which one group received 3 hours of specific instruction on Harvey. The second group received 3 hours of instruction using a CD. Six weeks later, both groups were tested blind with real patients with stable heart sounds. Stations were observed by an examiner who scored communication skills and examination skills using 5-point scales.

The Harvey-trained group was slightly but not significantly better than the CD-trained group at identifying heart sounds (3.11 versus 2.47, respectively; P = 0.06). However, there was no difference between the Harvey and CD-trained groups in diagnosis (2.94 versus 2.84, respectively), communication skills (18.9 versus 19.6, respectively) or examination skills (17.4 versus 17.5, respectively).

The study found little evidence that students trained with a high-fidelity simulator were more able to transfer skills to real patients than a control group. Although there was some suggestion that the Harvey-trained group was better at recognising heart sounds, there was no difference between groups in diagnostic accuracy or clinical skills.