Objective <p>Virtual patients (VP) such as high-fidelity patient simulators/manikins, VP avatars, or human simulated patients have been increasingly used in health professions and psychiatry residency education. The authors aimed to review extant literature and summarize the different types of VP adopted, training contexts, and its impact according to the Kirkpatrick model for evaluation of training outcomes.</p> Methods <p>The authors adopted the Joanna Briggs Institute methodology for evidence synthesis and used a five-step framework, searching several databases (MEDLINE, Web of Science, PsycInfo, PubMed, Scopus, and Cochrane) for pertinent articles published up to August 2024. Studies were included if they (1) were empirical studies involving use of VP in psychiatry residency training, (2) with evaluation of training outcomes, and (3) were published in English.</p> Results <p>Overall, 31 articles met the inclusion criteria. Among studies that were mostly conducted in the West, human simulated patients were most used, followed by VP avatar and high-fidelity simulators/manikins. Most studies across specific topics (ranging from anxiety to psychotic and alcohol use disorders) reported outcomes up to level 3 of the Kirkpatrick model. Apart from user acceptability, other training outcomes included improvements in knowledge about the topics covered, clinical care competency related to history taking, risk assessment, use of psychotropic medications, and neurostimulation treatment.</p> Conclusions <p>Future studies may consider multi-site collaborations to examine the impact of such VP interventions in different psychiatry residency training environments, compare across different VP modalities, include assessments at higher Kirkpatrick levels of training outcomes, and monitor longitudinal changes in learner outcomes.</p>