The Effects of Virtual Reality Ultrasound Training in Pairs Compared to Training Alone

Ultrasound is becoming increasingly used in many medical specialties. However, the quality of the ultrasound examination is highly operator dependent and sufficient training of clinicians is challenging. Virtual Reality ultrasound simulation enables training under controlled conditions and it allows trainees to make mistakes before practicing on patients. However, this type of training is expensive due to costs of simulators and clinician supervisors to provide instructions and feedback while the trainees are practicing. Training in pairs, also known as dyad training, may provide a learning advantage to trainees due to shared memory and increased confidence. Moreover, dyad training allows training twice the number of trainees on simulators, which has implications to the cost-effectiveness of VR simulation. In this study, we examine non-inferiority of dyad training compared to the effect of traditional training based on clinician instruction using a VR ultrasound simulator. Methods Study design. A randomized observer-blinded non-inferiority study in which training in pairs is compared to training alone on a VR simulator. Pilot study. Two pilot studies were conducted to determine which modules to include in the training curriculum and how to provide a standardized instruction of trainees. The first pilot study included 10 OB/GYN residents and 10 OB/GYN consultants, who rated and commented each module with regards to how useful they were to pelvic ultrasound training. These ratings were used to develop a training curriculum consisting of a series of basic gynecologic training modules. The second pilot study included 6 medical students, of which half trained in pairs and the other half trained alone with clinician instructions. This enabled the research team to develop a template for the instruction that should be provided by clinician instructors in the group of participants that practiced alone. Participants. Year-six medical students in their final rotations before entering residency enrolled in this study. The only exclusion criterion was previous ultrasound courses or extra-curricular ultrasound training. 30 participants were enrolled to ensure that at least 24 completed the transfer test. Setting. All participants are trained on a VR ultrasound simulator (Scantrainer, Medaphor) in an OB/GYN clerkship site. The Scantrainer is a high-fidelity vaginal ultrasound simulator that provides haptic feedback and realistic imaging. Randomization. All students were randomized to either training in pairs or training alone. Randomization was conducted at a different clinical department using computer-generated list of random numbers. Intervention. All students received 30 minutes of introduction to the simulator including a short introduction to vaginal ultrasound examination. The simulator instruction was provided using a standardized template developed in the pilot study. Constructive feedback was provided when participants encountered problems that they wanted feedback on. The instructor was a clinician ultrasound simulation expert. The participants in the intervention group did not receive any continuous instructor support but were allowed to consult an instructor whenever the participants encountered a module that they could not pass. The total training time was two hours in both groups. Hence, participants in the dyad group received half the 'hands-on' time as the participants in the control group. Outcome measure. The outcome measure was ultrasound examination skills on pelvic ultrasound examination of patients in an ambulatory setting. Participants were asked to perform a systematic ultrasonic transvaginal examination including measurement of endometrial thickness and estimate the volume of the right ovary. All patients were required to provide written consent to participate. A blinded gynecologist rated performance using a previously validated rating scale. Sample size calculation. Based on a previous construct validation study, a difference in performance scores corresponding to two months of clinical experience was regarded as significant, which corresponded a 4.6% difference. Using a SD of 0.2, a power of 0.80 and an alpha of 0.05, the total number of participants required was 24.