CoChat

AI-Powered Simulation for Nursing Education: Mixed Methods Systematic Review

Open access
Jiang, Hongzhan, Wang, Ziyan, Shen, Wanting, Meng, Meiqi, Yang, Dan, Li, Xuejing, Hao, Yufang · 2026 · Journal of Medical Internet Research

The DOI matches its Crossref record; 66 of the 66 references it cites were found in a registry or an index; 16 of the 52 sources its claims cite were checked for support.

Abstract

Background: Traditional simulation-based nursing education is often constrained by high costs, resource intensity, and limited scalability. AI-powered simulations offer dynamic, scalable, and personalized alternatives. However, the empirical evidence regarding their pedagogical effectiveness and learner acceptance remains fragmented. Objective: This study aimed to systematically evaluate and synthesize evidence on the effectiveness and learner perceptions of AI-powered simulations in nursing education. Methods: Following PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines, we systematically searched 11 electronic databases (PubMed, CINAHL, Embase, Web of Science, Cochrane Library, Scopus, SinoMed, CNKI, Wanfang, VIP, and Google Scholar) for studies published between January 2014 and September 2025. Two independent reviewers performed study selection, data extraction, and quality appraisal using design-specific tools (risk of bias 2 tool [RoB 2; Cochrane Bias Methods Group] for randomized controlled trials [RCTs], Risk Of Bias in Nonrandomized Studies of Interventions [ROBINS-I; Cochrane Bias Methods Group] for nonrandomized studies, Mixed Methods Appraisal Tool [MMAT] for mixed methods, Joanna Briggs Institute [JBI] for qualitative, and Agency for Healthcare Research and Quality [AHRQ] for cross-sectional studies). Quantitative data were synthesized narratively, and qualitative findings were integrated using JBI meta-aggregation. A convergent segregated approach with joint display was used to generate meta-inferences. Results: Nineteen studies involving 1253 participants (primarily prelicensure nursing students, with some interdisciplinary cohorts) were included. AI modalities comprised generative AI/large language models (n=7), AI-driven virtual patients/mannequins (n=5), AI-enhanced virtual/mixed reality (n=5), and chatbots (n=2). Three studies were RCTs, 4 were quasiexperimental with control groups, 3 were uncontrolled pre-post studies, 4 were mixed methods, 4 were qualitative, and one was a cross-sectional survey. Quantitative synthesis showed that evidence from RCTs and controlled quasiexperimental studies indicates significant improvements in cognitive knowledge and affective outcomes, including self-efficacy and communication confidence; however, effects on complex psychomotor skills were inconsistent, with one RCT finding AI-assisted simulation inferior to standardized patient simulation. Findings from uncontrolled designs are preliminary. Qualitative meta-aggregation revealed that learners valued safe, repeatable, nonjudgmental practice environments that reduced anxiety and bridged the theory-practice gap. Persistent challenges included technical frustrations, “robotic” interactions, lack of nonverbal cues, and system instability, collectively constituting an “authenticity gap.” Conclusions: AI-powered simulations show promise for developing foundational clinical reasoning and communication skills in nursing education, though the evidence base is limited by the predominance of uncontrolled designs, reliance on self-reported measures, and absence of longitudinal data on skill retention or clinical transfer. Due to current technological limitations in replicating physical and emotional authenticity, AI should be implemented as a complementary tool alongside traditional simulation methods and clinical placements, rather than as a replacement. Future research should prioritize longitudinal outcomes, standardized competency measures, RCTs with active comparators, and implementation strategies addressing technical barriers.

Overview

CoChat Summary

TL;DR

This mixed methods systematic review synthesized 19 studies (1253 participants) on AI-powered simulations in nursing education, finding short-term improvements in knowledge, clinical reasoning, communication confidence, and learner acceptance. However, evidence is limited by few RCTs, uncontrolled designs, self-reported outcomes, and a persistent learner-perceived “authenticity gap.”

Key findings

  • Nineteen studies involving 1253 participants were included; AI modalities were generative AI/LLMs (n=7), AI-driven virtual patients/mannequins (n=5), AI-enhanced VR/MR (n=5), and chatbots (n=2).
  • Among 11 studies assessing knowledge or skills, 8 (73%) reported improvements; among 9 assessing attitudes or self-efficacy, 8 (89%) reported improvements, but one RCT found AI-assisted simulation inferior to standardized patient simulation for breast self-examination skills.
  • Usability and acceptance were generally moderate to high: SUS scores ranged from 61.6 to 78.56, TAM perceived usefulness was 5.78/7, and 72%–93% of students expressed willingness to recommend AI simulations.
  • Qualitative synthesis identified valued safe, repeatable, nonjudgmental practice but persistent technical frustrations, “robotic” interactions, lack of nonverbal cues, and system instability, collectively termed the “authenticity gap.”

Methods

  • PRISMA-guided mixed methods systematic review with a convergent segregated design; protocol registered in PROSPERO (CRD420251208020).
  • Searched 11 databases (PubMed, CINAHL, Embase, Web of Science, Cochrane Library, Scopus, SinoMed, CNKI, Wanfang, VIP, Google Scholar) from January 2014 to September 2025; 3181 citations screened, 100 full texts assessed, and 19 studies included.
  • Two independent reviewers selected studies, extracted data, and appraised quality using RoB 2, ROBINS-I, MMAT, JBI, and AHRQ tools; quantitative data were synthesized narratively and qualitative findings integrated via JBI meta-aggregation and joint display.

Limitations

  • The paper states that heterogeneity in AI modalities, intervention durations, and measurement instruments precluded meta-analysis and standardized effect sizes; most studies used uncontrolled pre-post designs, and no study examined skill retention or clinical transfer beyond short-term postintervention windows.
  • The paper also notes small samples, inclusion of some nonnursing learners, near-uniform positive findings suggesting publication bias, and geographic concentration in high-income regions with no studies from Africa, South America, or low-income countries.
  • Evident from the text: only 3 RCTs were included, many outcomes relied on self-reported measures, and the review itself could not formally test publication bias due to heterogeneity.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.2196/95167

Cites

66 of 66 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Figures

Figure 1. PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) flow diagram displaying the process of included records.
Figure 1. PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) flow diagram displaying the process of included records.CC BY · PubMed Central

Verification

Claims the paper makes

12 Valid · 2 Not valid · 25 Not checked · 93 Own results and interpretations

56 more claims · 2 with their cited sources checked

Create a free account to see all 132 claims and every source verdict.

See all claims — free account
  • It may state two separate claims as one. 58
  • Probably not a substantial claim. 17

Create a free account to see these claims.

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

3 more claims about this paper

Create a free account to see every claim other papers make about this one.

See all claims — free account

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

5 of 5

Citations

Cites

Metadata

Authors
Jiang, H., Wang, Z., Shen, W., Meng, M., Yang, D., Li, X., & Hao, Y.
Affiliations
Beijing University of Chinese Medicine (CN)
Corresponding author
Hao, Yufang
Venue
Journal of Medical Internet Research
Published
2026-07-21
Volume
28
Pages
e95167-e95167
ISSN
1438-8871 (ISSN-L)
OpenAlex
W7169864056
Cited by
4
References
65
Citations by year
2026: 4
FWCI
4.16
Type
journal article
Language
en
Open access
Licence
cc-by
Locations
  • Journal of Medical Internet Research (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • PubMed Central (publishedVersion, cc-by)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion)
Topics
  • Social Sciences › Education › Simulation-Based Education in Healthcare
  • Health Sciences › Medicine › Health Informatics › Artificial Intelligence in Healthcare and Education
  • Social Sciences › Education › Innovations in Medical Education
Keywords
nursing education, simulation-based training, large language models, generative AI, virtual patients, chatbots, virtual reality, mixed reality, learner perceptions, self-efficacy, psychomotor skills, standardized patients, prelicensure nursing students, student anxiety, personalized learning, theory-practice gap, standardized test scores
MeSH terms
Artificial Intelligence, Education, Nursing, Humans

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

The claim verdicts this report counts are experimental: their accuracy is not yet measured.

A language model drew the claims from the paper on 2026-10-10, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).

A second model labelled each claim on 2026-10-10 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).

Cited sources are checked with a language model’s help (gpt-6-sol).

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.