Effects of High‐Fidelity Simulation Training on Learning Outcomes and Satisfaction for Practising Registered Nurses: A Systematic Review and Meta‐Analysis

ABSTRACT

Background

Ample evidence has shown the benefit of high‐fidelity simulation (HFS) in promoting learning in pre‐licensure nursing students, but the evidence for practising registered nurses has not been synthesised.

Objective

To evaluate the effects of HFS training on learning outcomes and satisfaction in practising registered nurses.

Methods

In this systematic review and meta‐analysis, we searched PubMed, CINAHL, ERIC, Cochrane Library, Web of Science and China National Knowledge Infrastructure (CNKI) for studies published in English or Chinese from database inception to 31 May 2023 (updated on 20 April 2025). All randomised controlled trials (RCT) or quasi‐experiments that compared HFS training with traditional methods (e.g., lecture) for practising registered nurses and reported learning outcomes and satisfaction were included. Risk of bias was assessed by the Cochrane risk‐of‐bias tool for randomised trials (RoB 2) and non‐randomised trials (ROBINS‐I). Inverse‐variance random‐effect models were used to calculate standardised mean differences (SMDs) with 95% confidence interval (CI). We followed the PRISMA 2020 guideline.

Results

Of 1404 records, eight eligible studies (five RCTs and three quasi‐experiments) involving 275 practising nurses were identified. Two RCTs had high risk of bias, while others showed some concerns or moderate risk of bias. Meta‐analyses showed that HFS could promote knowledge acquisition (SMD = 0.65, 95% CI, [0.35, 0.95], p < 0.01, I² = 0%), professional skills (SMD = 0.72, 95% CI, [0.41, 1.04], p < 0.01, I² = 0%) and learning satisfaction (SMD = 1.24, 95% CI, [0.35, 2.13], p < 0.01; I² = 67%), compared with traditional methods. The pooled effect on self‐confidence was marginally insignificant (SMD = 0.59, 95% CI, [−0.04, 1.22], p = 0.07; I² = 67%).

Conclusion

Compared with traditional training methods, HFS is effective in promoting knowledge acquisition, professional skills and learning satisfaction and may enhance self‐confidence among practising nurses. To strengthen the evidence base, more rigorous RCTs with larger sample sizes, adequate reporting of HFS design, and standardised outcome measures are warranted.

Protocol Registration

PROSPERO (CRD42022358717). No Patient or Public Contribution.

Summary

• High‐fidelity simulation (HFS) is effective in enhancing knowledge acquisition, professional skills and learning satisfaction among practising registered nurses.

• The findings support the use of HFS for continuous nursing education and professional development in advanced nursing practise.

• More rigorous trials with larger samples and standardised measures are warranted to strengthen the evidence for HFS in nursing.

1. Introduction

High‐fidelity simulation (HFS) often refers to the use of computer‐controlled, full‐size manikins that closely mimic realistic clinical scenarios, providing highly immersive and interactive simulated experiences for learners (Lioce et al. 2022). This approach allows learners to develop both technical and non‐technical skills within a safe and controlled setting, enhancing their readiness to real‐world clinical practice. Despite the advancements in simulation technologies, such as virtual reality, HFS remains a critical tool in developing clinical competencies (Jiang et al. 2024).

Recent meta‐analyses have shown HFS's effectiveness in enhancing knowledge and professional skills in undergraduate nursing students (Lei et al. 2022; Li et al. 2022; Tonapa et al. 2023). Additional benefits identified through meta‐analyses included improvements in critical thinking, clinical judgement and communication skills (Lei et al. 2022), as well as collaboration, caring and learning interest (Li et al. 2022) and self‐confidence (Tonapa et al. 2023). A systematic review further showed the benefit of HFS in improving clinical reasoning skills (Alshehri et al. 2023). HFS has become the most popular simulation‐based learning method in baccalaureate and prelicensure master's programs in the US (Smiley 2019).

Emerging studies have suggested that HFS may also be effective for practising registered nurses. Integrated reviews indicated benefits in managing clinical deterioration in acute care nurses (O’Rourke et al. 2023) and in enhancing knowledge acquisition and confidence among critical care providers (Boling and Hardin‐Pierce 2016). A systematic review that mostly included nurse practitioners provided preliminary evidence that HFS can promote learning and satisfaction compared with traditional methods (Warren et al. 2016). However, we were not aware of any meta‐analyses that specifically examined the effectiveness of HFS training for practising registered nurses.

Despite its potential benefits, HFS is often underutilised in continuous nursing education. This may be due to factors such as resource constraints, lack of institutional support and the challenges of integrating simulation into busy work schedules (Al‐Ghareeb and Cooper 2016). Given the differences in prior knowledge and clinical experiences between undergraduates and practising professionals, findings from studies focusing on nursing students may not be applicable to registered nurses. Synthesised evidence is warranted to substantiate the use of HFS for advanced nursing training. In this systematic review and meta‐analysis, we evaluated the effectiveness of HFS training in promoting learning outcomes and satisfaction among practising registered nurses.

2. Methods

This systematic review and meta‐analysis was guided by the Cochrane Handbook for Systematic Review of Interventions (Higgins et al. 2023). The process and findings were reported in accordance with the Preferred Reporting Items for Systematic Review and Meta‐analysis (PRISMA) 2020 Statement (Page et al. 2021). This review was prospectively registered with PROSPERO (CRD42022358717). AI technologies were not used to develop any portion of this manuscript. Supporting Information is available in the supplementary section.

2.1. Inclusion and Exclusion Criteria

Eligible studies included in this review needed to meet the following criteria. First, only RCTs and quasi‐experimental studies comparing HFS training with traditional educational methods (e.g., lecture, discussion, static manikin) were considered. Participants must include only practising registered nurses. HFS‐based teaching methods should be the primary interventions, with or without any complementary interventions that were employed in both the intervention group and control group. For instance, (Starodub et al. 2020), which included a video lecture in both groups, was eligible for inclusion. In contrast, in (Jung et al. 2023) the intervention group received HFS training and video lectures, while the control group received usual training. Consequently, this study was excluded from the review. Studies also needed to report on at least one of the following outcomes: knowledge acquisition, professional skills, self‐confidence or learning satisfaction.

2.2. Search Strategy

A systematic electronic search was conducted in PubMed, Cumulative Index to Nursing and Allied Health Literature (CINAHL), Education Resources Information Center (ERIC), Cochrane Central Register of Controlled Trials, Web of Science and China National Knowledge Infrastructure (CNKI) to identify original peer‐reviewed papers. We included studies published in English or Chinese published from inception to 31 May 2023. An updated search was conducted on 20 April 2025. The search terms included keywords synonymous with nurse and HFS and subject indexing terms (e.g., medical subject headings, MeSHs) combined using Boolean operators. Supplementary Appendix A shows the search strings used in each database. Backward citation search was conducted by manually checking the reference lists of articles eligible for full‐text screening.

2.3. Data Extraction

Endnote 20 was used to manage the records identified from the search and to remove duplicated records. Two independent (JZ and GWZ) reviewers screened all titles, abstracts and full‐text articles for inclusion. Using a standard data extraction form, two reviewers (JZ and GWZ) independently extracted data from eligible studies, including authors’ names, year of publication, country, population, settings, study design, HFS designs (clinical scenario, duration of prebriefing, simulation and debriefing, manikin brand), teaching method of comparison, sample size, outcome measures and summary statistics (e.g., mean, SD, median, IQR) of the outcome. Following the NLN Jeffries Simulation Theory framework, the outcomes were categorised as knowledge acquisition, professional skills, self‐confidence and learning satisfaction (Jeffries et al. 2015). For studies that reported the outcomes at multiple time points, we prioritised outcomes measured at the final time point. For studies using multiple instruments of similar outcomes, we selected the most commonly used instrument. Any disagreements in the study selection and data extraction processes were resolved through discussion between the two reviewers, with a third reviewer (TTL) as arbitrator.

2.4. Quality Appraisal

Methodological quality of RCTs was assessed using the Cochrane revised tool for risk of bias in randomised trials (RoB 2) (Sterne et al. 2019), while the quality of quasi‐experimental studies was assessed by the risk of bias in non‐randomised studies of interventions (ROBINS‐I) (Sterne et al. 2016). Two independent reviewers (JZ and GWZ) completed the quality assessment, with an arbitrator (TTL) resolving potential disputes. Additionally, the Cochrane online tool “robvis” was employed to visualise the result of the risk of bias assessment (McGuinness and Higgins 2021).

2.5. Data Synthesis

Statistical analysis was conducted in R software (version 4.4.1) using the ‘meta’ package. For studies that reported summary statistics as sample median, interquartile range or minimum and maximum, an online calculator was used to estimate the sample mean and standard deviation (SD) (Cai et al. 2021; McGrath et al. 2020). Studies with significant skewness away from normality were excluded from the meta‐analyses (Shi et al. 2023). Effect sizes (Hedge's g) were presented as standardised mean difference (SMD). An SMD of 0.2, 0.5 and 0.8 indicated small, moderate and large effect sizes, respectively (Andrade 2020). A priori inverse‐variance random‐effects model was used to pool the SMDs, using restricted maximum likelihood as the variance estimator. Fixed‐effect models were also conducted as sensitivity analyses to examine the robustness of the results (Higgins et al. 2023). The Higgins' I² statistics were computed to quantify the percentage of the variability in effect estimates that was due to heterogeneity rather than sampling error. Cochran's Q test was used to test for significant heterogeneity. Sources of heterogeneity were investigated through the leave‐one‐out method and subgroup analysis by study design (RCT vs. quasi‐experiments) (Higgins et al. 2023). An assessment of publication bias using funnel plots was planned if the meta‐analysis included ten or more studies (Sterne et al. 2011).

3. Results

3.1. Study Selection

We identified 1346 records from the database search and an additional 58 from backward citation searching. After removing 197 duplicates, 1149 records were screened by titles and abstracts. Twenty‐four full‐text studies were evaluated for eligibility based on the predefined eligibility criteria. Ultimately, five RCTs (Corbridge et al. 2010; Jansson et al. 2014; Lemarie et al. 2019; Starodub et al. 2020; Zhang et al. 2019) and three quasi‐experimental studies (Andrighetti et al. 2012; Luo et al. 2021; O’Leary et al. 2016) were included in this systematic review. All studies were included in the systematic review, but one study (O’Leary et al. 2016) was excluded from the meta‐analysis due to severe skewness of the outcome data. Figure 1 shows the PRISMA flow diagram.

FIGURE 1.

The PRISMA flowchart. [Colour figure can be viewed at wileyonlinelibrary.com]

3.2. Study Characteristics

The eight studies included in this review were published between 2010 and 2021, involving a total of 275 clinical practising registered nurses. Table 1 summarises the study characteristics. The studies were conducted across five countries, mostly in the United States (3/8) and China (2/8). Sample sizes ranged from 20 to 60, with most studies targeting nurses practising in emergency departments or intensive care units (5/8). The HFS training designs varied widely among the studies, including differences in scenarios, location and duration. However, the scenarios were all related to high‐stake procedures or critical and emergency care. Control conditions also varied, including different traditional training formats such as presentations, videos, lectures and discussions. All outcome measures were self‐reported, except in (Jansson et al. 2014), where professional skills were observer‐rated.

TABLE 1.

Characteristics of included studies.

Author, year, country	Population, study design, trial registration	Intervention	Comparison	Outcome measures a
Corbridge et al. (2010) US	20 adult, geriatric and acute care advanced practice nurse students (I/C = 10/10), RCT, not registered	Scenario: Principles of mechanical ventilation Location: Simulation center Duration: Prebriefing NR; Simulation NR; Debriefing NR Manikin brand (model): Laerdal (SimMan)	Online, narrated PowerPoint presentation of the principles of mechanical ventilation. Duration: NR	Knowledge: A self‐designed 12‐item questionnaire Satisfaction: A self‐designed 5‐item survey. Last measurement: End of training.
Jansson et al. (2014) Finland	25 critical care nurses (I/C = 13/12), RCT, not registered	Scenario: Nursing management of patients requiring mechanical ventilation. Location: Adult mixed medical‐surgical ICU. Duration: Prebriefing: 20 min; Simulation: 10 min; Debriefing: 60 min. Manikin brand (model): Gaumard (HAL).	No intervention. Duration: NR	Knowledge: Ventilator Bundle Questionnaire Skills: Ventilator Bundle Observation Schedule (observer‐rated). Last measurement: 6 months after randomization.
Lemarie et al. (2019) France	30 critical care nurses (I/C = 15/15), RCT Clinicaltrial.gov: NCT02379234	Control condition + HFS Scenario: Continuous renal replacement therapy (CRRT). Location: an environment reproducing an ICU room Duration: 6 h, 3 session of 2 h each including prebriefing and debriefing. Manikin brand (model): Laerdal (Nursing Kelly).	Theoretical (2 h) and practical training (4 h) on CRRT in an ICU.	Knowledge: A self‐developed 40‐item questionnaire covering CRRT principles, alarms and generator settings. Self‐Confidence: A numerical scale (0–10). Last measurement: End of training.
Starodub et al. (2020) US	42 critical care nurses and 10 emergency care nurses (I/C = 24/28), RCT, not registered	Control condition + HFS Scenario: Delivery of targeted temperature management (TTM) for cardiac arrest patients. Location: Academic setting. Duration: Prebriefing 5 min; Simulation: 60 min; Debriefing: NR. Manikin brand (model): Laerdal (SimMan 3G).	30‐min TTM lecture +1‐h video lecture on TTM case studies.	Knowledge: A 20‐item multiple‐choice test adapted from the Hypothermia and Resuscitation Training Institute at Penn (HART) biannual conference. Skills: A self‐developed TTM skills checklist. Self‐Confidence and Satisfaction: Each measured on a numerical scale (1–10). Last measurement: 6 ± 2 weeks after training.
Zhang et al. (2019) China	60 emergency care nurses (I/C = 30/30), RCT, not registered	Scenario: Diagnosis, assessment, healthcare education and emergency care. Location: Simulation ward. Duration: Prebriefing NR; Simulation 20 min, Debriefing: NR. Manikin brand (model): NR.	Case study, group discussion and role play on emergency care. Duration: NR.	Knowledge: A self‐designed questionnaire. Skills: A self‐designed questionnaire. Last measurement: End of training.
Andrighetti et al. (2012) US	28 nurses enrolled in a graduate midwifery education program (I/C = 18/10), Quasi‐experiment.	Scenario: Managing postpartum haemorrhage complications (nine nurses) or shoulder dystocia (nine nurses) Location: Simulated labor and delivery unit Duration: Prebriefing NR; Simulation NR; Debriefing NR. Manikin brand (model): NR.	Discussion of postpartum haemorrhage management using a question‐and‐answer approach (nine nurses) or discussion and video on managing shoulder dystocia and then practice of hand manoeuvres using a static manikin Duration: NR.	Self‐confidence: The student satisfaction and self‐confidence in learning scale (SSSCLS). Last measurement: End of training.
Luo et al. (2021) China	30 newly graduated registered nurses (I/C = 16/14), Quasi‐experiment.	Scenario: Acute myocardial infarction, fracture of the lower leg, chronic obstructive pulmonary diseases and intestinal obstruction. Location: Simulation laboratory. Duration: Prebriefing: 10 min; Simulation: 15; Debriefing: 35 min. Total 1 h for each scenario Manikin brand (model): Laerdal (SimMan 3G).	Case study using lecture and discussion about cases including acute myocardial infarction, fracture of the lower leg, chronic obstructive pulmonary disease and intestinal obstruction. Duration: 4 h, 1 h for each case.	Skills: Lasater Clinical Judgement Rubric Self‐confidence: The student self‐confidence subscale in SSSCLS. Satisfaction: Simulation Design Scale Last measurement: End of training.
O’Leary et al. (2016) Australia	30 nurses from paediatric critical care unit (I/C = 15/15), Quasi‐experiment.	HFS + a print copy of the slides used in control group. Scenario: Recognising and managing paediatric deterioration (infant sepsis). Location: Paediatric critical care unit. Duration: Prebriefing: NR; simulation: 10 min; debriefing NR. Manikin brand (model): Laerdal (SimBaby).	Standard instruction sessions comprised of a 15‐min didactic presentation with 14 slides.	Knowledge: A self‐developed 5‐item multiple‐choice questionnaire. Last measurement: End of training.

Abbreviations: C = control group; I = interventional group; NR = not reported.

^a All measures were self‐reported unless otherwise specified.

3.3. Quality Appraisal of Included Studies

Based on the RoB 2 tool, three RCTs showed some concerns of bias, while two were rated as high risk (Figure B1. in Supplementary Appendix B). None of the RCTs adequately described the randomization process, particularly allocation concealment. (Corbridge et al. 2010) were rated as high risk because of differing methods of outcome measurement, with the intervention group completing the survey in‐person and the control group online. Jansson et al. (2014) were also rated as high risk due to differing retention rates between the intervention group and control group (87% [13/15] vs. 67% [10/15]), raising concerns of bias due to missing outcome data. While blinding of participants was not feasible due to the nature of the intervention, there was no evidence or strong reason to believe that the trial context led to deviation in implementing the protocol intervention. Therefore, all studies were rated as low risk in domain 2.

Based on ROBINS‐I, the risk of bias of all three quasi‐experimental studies was rated as moderate because of potential confounding inherent in the non‐randomised designs (Figure B2). (Luo et al. 2021) additionally had a moderate risk of bias in missing data due to differential retention rates between groups (84% [16/19] vs. 70% [14/20]).

3.4. Effects of HFS on Learning Outcomes

3.4.1. Knowledge Acquisition

Five RCTs (Corbridge et al. 2010; Jansson et al. 2014; Lemarie et al. 2019; Starodub et al. 2020; Zhang et al. 2019) and one quasi‐experiment (O’Leary et al. 2016) reported on knowledge acquisition. Data from O’Leary et al. (2016), which showed a higher post‐test score in the HFS group than control group (median [min‐max]: 4 [2–5] vs. 2 [2–4]), were excluded from the meta‐analysis due to substantial skewness in post‐test score. The meta‐analysis included 185 practising registered nurses (92 in the interventional group and 93 in the control group). Most studies (3/5) used self‐reported instruments developed by the authors. The random‐effects model revealed a moderate effect of HFS on knowledge acquisition (SMD = 0.65, 95% CI, [0.35, 0.95], p < 0.01), with no notable heterogeneity (I² = 0%, p = 0.62) (Figure 2). Results remained consistent in the fixed‐effects model.

FIGURE 2.

Effects of HFS on knowledge acquisition. [Colour figure can be viewed at wileyonlinelibrary.com]

3.4.2. Professional Skills

Three RCTs (Jansson et al. 2014; Starodub et al. 2020; Zhang et al. 2019) and one quasi‐experiment (Luo et al. 2021) assessed professional skills, involving 165 practising registered nurses (83 in the interventional group and 82 in the control group). Half of the studies used self‐developed instruments. The meta‐analysis with a random‐effects model showed a moderate pooled effect on professional skills (SMD = 0.72, 95% CI, [0.41, 1.04], p < 0.01), with no notable heterogeneity (I² = 0%, p = 0.57) (Figure 3). The findings were consistent in the fixed‐effects model.

FIGURE 3.

Effects of HFS on professional skills. [Colour figure can be viewed at wileyonlinelibrary.com]

3.4.3. Self‐Confidence

Self‐confidence was reported in two RCTs (Lemarie et al. 2019; Starodub et al. 2020) using self‐developed instruments and two quasi‐experiments (Andrighetti et al. 2012; Luo et al. 2021) using the Student Satisfaction and Self‐Confidence in Learning (SSSCLS). A total of 140 practising registered nurses (73 in the interventional group and 67 in the control group) were involved. The random‐effects meta‐analysis showed no statistically significant effect on self‐confidence (SMD = 0.59, 95% CI, [−0.04, 1.22], p = 0.07) (Figure 4). However, the fixed‐effect model yielded significant results (SMD = 0.48, 95% CI, [0.14, 0.83], p < 0.01). No obvious source of substantial heterogeneity (I² = 67%, p = 0.05) was found using the leave‐one‐out method (Figure C1 in Supplementary Appendix C). Subgroups analysis revealed the study design contributed to heterogeneity. Specifically, RCTs showed no effect (SMD = 0.08, 95% CI, [−0.35, 0.52]), while quasi‐experiments showed a large significant effect (SMD = 1.19, 95% CI, [0.61, 1.76]) (Figure C2).

FIGURE 4.

Effects of HFS on self‐confidence. [Colour figure can be viewed at wileyonlinelibrary.com]

3.4.4. Learning Satisfaction

Two RCTs (Corbridge et al. 2010; Starodub et al. 2020) and one quasi‐experiment (Luo et al. 2021) reported on learning satisfaction, involving 102 practising registered nurses (50 in the interventional group and 52 in the control group). Only (Luo et al. 2021) used an existing instrument for measuring satisfaction. The random‐effects meta‐analysis showed a large pooled effect on learning satisfaction [SMD = 1.24, 95% CI, (0.35, 2.13), p < 0.01], with high heterogeneity (I² = 67%, p = 0.05) (Figure 5). The fixed‐effect model yielded similar results [SMD = 1.09, 95% CI, (0.66, 1.52), p < 0.01]. The leave‐one‐out method showed a notable decrease of I² to 11% after omitting (Corbridge et al. 2010), which reported a very large effect on learning satisfaction (SMD = 2.35, 95% CI, [1.16, 3.54]) (Figure. C3). Nevertheless, the pooled effect remained large and significant after excluding this study (SMD = 0.90, 95% CI, [0.41, 1.39], p < 0.01).

FIGURE 5.

Effects of HFS on learning satisfaction. [Colour figure can be viewed at wileyonlinelibrary.com]

3.4.5. Publication Bias

Publication bias was not examined, as all meta‐analyses involved fewer than 10 studies.

4. Discussion

To our knowledge, this is the first meta‐analysis synthesising evidence on the effects of HFS training on learning outcomes and satisfaction among practising registered nurses. Meta‐analyses showed significant moderate effects of HFS training on knowledge acquisition, professional skills and a large effect on learning satisfaction. The marginally insignificant results were indicative of the effectiveness of HFS training in promoting self‐confidence. The high heterogeneity, coupled with discrepant results between the random‐effect and fixed‐effect models, suggested the need for more definitive studies. Overall, this study adds to the literature by showing that HFS training can be utilised in continuing education and professional development for practising registered nurses, potentially leading to improved quality of care and better patient outcomes.

The HFS training in the included studies predominantly focuses on the management of critical or emergency conditions. The moderate effect sizes observed for knowledge acquisition (SMD = 0.65) and professional skills (SMD = 0.72) corroborated previous findings from integrated reviews, which indicated HFS's potential in improving knowledge acquisition and management of clinical deterioration in critical care nurses (Boling and Hardin‐Pierce 2016; O’Rourke et al. 2023). Our results also align with a previous meta‐analysis that included both undergraduate and postgraduate nursing students, which showed the effectiveness of HFS in improving knowledge and performance in managing life‐threatening scenarios (La Cerra et al. 2019). Importantly, these findings suggest that HFS training is not only beneficial for nursing students but also provides valuable opportunities for those already in the field to refresh and refine their clinical skills, ensuring they remain adept in managing high‐stakes and often infrequent situations.

Learning satisfaction also emerged as a significant outcome, with a large effect size (SMD = 1.24), highlighting HFS's potential to enhance learner engagement and satisfaction compared to traditional methods. This finding is consistent with previous studies that identified increased satisfaction and interest in learning among nursing students (Warren et al. 2016). However, the findings for self‐confidence were inconclusive. The random‐effects model did not yield a statistically significant effect, although the fixed‐effect model indicated a significant moderate effect (SMD = 0.48). The discrepancy likely reflects the high heterogeneity (I² = 67%) among the studies. While the fixed‐effect model assumes a common effect and yields significant results, the random‐effects model accounts for between‐study variability. This results in wider confidence intervals and a non‐significant finding. Subgroup analysis revealed that quasi‐experimental studies showed a large and significant effect on self‐confidence (SMD = 1.19), indicating that the study design and context of the training might influence outcomes. Notably, findings from meta‐analyses in undergraduate nursing students were also mixed: (Li et al. 2022) reported a small and insignificant effect of HFS on self‐confidence (SMD = 0.22), while (Tonapa et al. 2023) found a moderate and significant effect (SMD = 0.56). More studies are needed to provide more conclusive evidence.

The quality of evidence presented in the review varied across the included studies. Utilising the RoB 2 and ROBINS‐I tools, we identified several concerns regarding bias, particularly the inadequate descriptions of the randomization process in RCTs. While some studies reported moderate to high risks of bias, the overall consistency in findings for knowledge acquisition and professional skills enhances our confidence in the effectiveness of HFS training. An additional issue is the lack of pre‐registration of the trials in the Primary Registry in the WHO Registry Network and inadequate reporting of the designs of the HFS (e.g., learning duration). Future research should follow EQUATOR guidelines, particularly the CONSORT extensions for health care simulation research (Cheng et al. 2016), in the trial design, execution and reporting to enhance the quality of evidence. Another common limitation is the small sample sizes, which is typical in studies involving practising registered nurses. This challenge is likely due to the difficulty in engaging working nurses in research, especially as many HFS studies for clinical nurses are conducted within postgraduate programs that have smaller class sizes or focus on specific hospital departments. Conducting multicentre studies may provide a viable solution to enhance participant recruitment in future research. Nonetheless, this limitation underscores the value of conducting meta‐analyses, which pool data from smaller studies to strengthen the overall evidence base for HFS training in nursing education.

Several limitations should be acknowledged. First, there is notable variability in the study designs, including HFS training protocols, control conditions and assessment tools, in the included studies. The paucity of studies precluded subgroup analyses or meta‐regression to identify study characteristics or effective HFS features that can optimise learning outcomes. Second, we could not examine learning outcomes that were not reported in the reviewed studies, such as clinical judgement and communication. Third, an assessment of publication bias was not feasible due to insufficient studies. Finally, while our studies included papers published in English and Chinese, studies published in other languages were not included.

5. Conclusion

This meta‐analysis provides evidence that HFS training is beneficial for enhancing knowledge acquisition, professional skills and learning satisfaction among practising registered nurses. Given the critical role of continuous education in nursing, integrating HFS into ongoing training programs could significantly benefit clinical practice and patient outcomes. Further studies with a randomised design, larger sample sizes and uniform outcome measures are warranted. Additionally, considering the resource‐intensive nature of HFS, exploring its integration with other simulation modalities such as computer‐based simulation could offer more flexible and cost‐effective training solutions.

Author Contributions

J.Z. and T.T.L.: conceptualization, validation and writing – original draft. J.Z., G.Z. and T.T.L.: data curation. J.Z.: formal analysis and visualisation. J.Z., P.S. and T.T.L.: methodology. All authors: writing – review and editing. T.T.L.: supervision.

Funding

The authors have nothing to report.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Data S1: Supplementary Appendix.

Data Availability Statement

Data sharing not applicable to this article as no datasets were generated or analysed during the current study.