ContextDespite the growing use of virtual patients (VPs) in medical education, few studies have explored the features and effectiveness of VP‐based medical communication skills training. We undertook a systematic review to summarise the design and evaluation of VP‐based medical communication skills training systems in order to identify features of successful cases.MethodsFollowing PRISMA guidelines, we searched four databases for studies published between 2006 and 2018. Using a refined classification scheme, we extracted data on instructional design (scenario and instructional intervention), technological design (modality and interaction), and evaluation (user experience, learning effectiveness and evaluator). We assessed the quality of studies using the Medical Education Research Study Quality Instrument (MERSQI) and the QualSyst standard assessment criteria.ResultsA total of 14 studies were included for review. Of these, 85.7% (n = 12) were quantitative and 71.4% (n = 10) involved undergraduate students. The most common VP training scenario was history taking followed by the delivery of bad news. Diverse instructional interventions, including tutorials, learning activities, and feedback, were embedded in the VPs. The first‐person perspective animated within‐screen size VP was a popular technological feature. Most evaluations concerned the reality of simulation (for user experience) and skill in expressing empathy (as a learning outcome). Of the eight comparative studies, half reported significant attitude or skill improvements in the VP group. The distinct features of VPs shown to be effective were well‐designed instructional interventions (eg, a pre‐activity with a protocol‐informed tutorial), and post‐activity (eg, debrief or reflection), scaffolding and human feedback, but not system feedback.ConclusionsEvidence‐based VP training can enable students to gain communication skills in a safe and affordable learning environment. Elaborate technology alone cannot guarantee effective learning, but evidence‐based instructional interventions can facilitate its optimal use and bring about better learning outcomes.

No abstract