Abstract
Aim
To evaluate the impact of higher physical realism of manikins on educational and clinical outcomes during life support education.
Methods
This systematic review was conducted as part of the continuous evidence evaluation process of the International Liaison Committee on Resuscitation (ILCOR). A search of PubMed, Embase, and Cochrane was conducted from January 1, 2005 until April 30, 2024. Studies comparing training with higher physical realism manikins and lower realism manikins were eligible for inclusion. Studies comparing manikins to other forms of training (e.g. screen-based, virtual reality) were excluded. Risk of bias was assessed using Cochrane Risk of Bias 2 (RoB 2) for randomized trials and Risk Of Bias In Non-Randomized Studies of Interventions (ROBINS-I) for observational studies. For outcomes reported by four or more randomized studies, random effects meta-analysis using standardized mean difference was performed.
Results
Of the 1276 articles identified and screened, 21 articles comprised the final review (19 randomized trials, 2 observational studies). Meta-analysis of eight RCTs reporting simulation skill performance in a simulated clinical scenario at course conclusion demonstrated a benefit from the use of higher- realism manikins compared with lower realism manikins (standardized mean difference 0.66, 95% CI 0.08 – 1.25). Meta-analysis of seven RCTs reporting knowledge at course conclusion showed no significant difference between the use of both types of manikins. Significant risk of bias and a high degree of heterogeneity were found among the included studies.
Conclusion
This systematic review found that higher manikin realism during resuscitation training was associated with improved simulated clinical scenario performance at course conclusion; without an effect on knowledge at course conclusion. Future studies should examine the impact of resource requirements for high realism simulation on generalizability and implementation.
Keywords: Life support education, Physical realism, Manikin fidelity, Cardiopulmonary resuscitation
Introduction
Effective education of both laypersons and healthcare professionals in basic and advanced life support is one of the key components in improving survival from cardiac arrest.1 Published studies have examined many novel approaches to instructional design, course delivery, or technology use to improve the effectiveness and durability of life support education.2 The use of manikins with a greater degree of physical realism (so-called ‘high-fidelity’ manikins) has been implemented over the past three decades in the hopes of improving learner engagement by allowing them to ‘suspend disbelief’ to a greater extent and thus participate in life support training in a manner that more closely reflects their actual clinical performance.
An International Liaison Committee on Resuscitation (ILCOR) systematic review by Cheng et al in 2015 examined this research question, and concluded that higher manikin realism was associated with improved clinical performance at course conclusion.3 This review discussed the inherent difficulty in balancing the cost and resource use associated with this technology, as well as the ongoing challenge of ensuring that the features demonstrated by a manikin with higher physical realism were incorporated into instructional design that appropriately addressed the needs of learners in a given setting. With recent advancements in simulation, new features of manikins, and updated educational approaches in teaching life support, the International Liaison Committee on Resuscitation (ILCOR) Education, Implementation and Teams Task Force (EIT) decided to perform this updated systematic review of contemporaneous evidence. We evaluated outcomes of life-support education studies comparing training with manikins with higher physical realism to training with manikins with lower physical realism.
Methods
This systematic review was part of the continuous evidence evaluation process by the International Liaison Committee on Resuscitation (ILCOR) Education, Implementation and Teams Task Force (EIT) and it was registered with PROSPERO (registration number CRD42024535044). The PROSPERO registration was subsequently updated with new wording for the intervention of interest (‘physical realism’ instead of ‘fidelity’) and the subset analysis of laypeople as learners was eliminated.
The systematic review was driven by a research question using the PICOST (Population, Intervention, Control, Outcomes, Study Design, and Timeframe) format: In learners training in any form of life support (P), does instruction using manikins with a greater degree of physical realism (I), compared to instruction with a lesser degree of physical realism (C), affect educational outcomes (skill, knowledge, attitudes) or clinical processes (change in healthcare provider behavior, patient outcomes, cost or resource utilization) (O)? Randomized controlled trials and non-randomized studies (non-randomized controlled trials, interrupted time series, controlled before-and-after studies, cohort studies) were included in any language as long as there was an English abstract available. Unpublished studies (e.g., conference abstracts, trial protocols) were excluded (S). Databases were searched over a timeframe from January 1, 2005 to April 30, 2024 (T); the 2005 start date was chosen based on the dates of published studies included in the previous systematic review being from 2005 onward.
Physical realism was defined as the presence of physical features (visible, palpable, audible) designed to increase the resemblance of the manikin to an actual patient. Eligible studies compared outcomes between groups using manikins of two different levels of realism; this was most commonly a static manikin compared with a higher degree of physical realism simulator, but no specific levels of realism were determined a priori to be a prerequisite for eligibility. Studies that compared manikins of any level of realism to other content delivery modalities (e.g. screen-based, virtual reality) were excluded.
Search strategy
We searched EMBASE.com, Medline, and Cochrane and citations of included studies were subsequently reviewed for additional eligible studies (Appendix).
Study selection and data collection
The titles and abstracts of all potentially eligible studies were screened for inclusion by pairs of independent reviewers. The full text of included studies were checked against the inclusion and exclusion criteria independently in these pairs of reviewers. Any disagreements between the reviewers at either stage were resolved by discussion finding consensus. Data from each study were independently extracted by a pair of reviewers and grouped separately according to the predefined outcomes.
Risk of bias assessment
Risk of bias was assessed by the same pairs of reviewers. The Cochrane Risk of Bias Tool for randomized trials (RoB 2) was used for randomized studies4; the Risk of Bias In Non-randomised Studies of Interventions (ROBINS-I) for observational studies.5
Data analysis and synthesis
Two reviewers independently assessed the certainty of evidence of each outcome using the GRADE approach (Grades of Recommendation, Assessment, Development and Evaluation). In case of disagreement, consensus was reached by discussion. We conducted both quantitative and qualitative syntheses of the evidence. For specific outcomes reported by more than four studies with sufficient data (point estimate and estimate of variance), we performed random-effects meta-analysis to quantitatively pool results using a standardized mean difference (SMD) to allow direct comparison. In interpreting the clinical significance of our results, we emphasized confidence intervals (CI) in relation to Cohen’s effect size (or SMD) classifications, where SMD > 0.8 large; SMD 0.5–0.8 = moderate; SMD 0.2–0.5 = small; and SMD < 0.2 = negligible.6 Data was entered into Review Manager (RevMan5, The Cochrane Collaboration, Oxford, UK) to calculate SMD, 95% confidence intervals and statistical heterogeneity. We quantified between-study inconsistency using the I2 statistic, which estimates the percentage of variability not due to chance. I2 values > 50% indicate large inconsistency or heterogeneity. Studies that could not be combined in the above quantitative synthesis of results were analysed and summarized in a narrative fashion.
Results
The search identified 1250 articles after duplicates were removed. Of these, 1217 articles were excluded during title and abstract screening, leaving 33 full-text articles to be screened for eligibility. During full-text review, 12 articles were subsequently excluded, leaving 21 articles comprising the final review.7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27 The PRISMA diagram for the review is shown in Fig. 1. Fourteen of the included studies were among the ones included in the previously published systematic review3; seven additional studies are included in the present review. All studies involved healthcare professionals and/or healthcare trainees. 15 studies were performed in North America8, 9, 10, 11, 12, 13, 14, 15, 16, 18, 22, 24, 25, 26, 27; 4 in Asia7, 19, 21, 23; one in Europe17; and one in Australia.20 12 studies used adult scenarios7, 10, 14, 15, 16, 17, 18, 20, 21, 22, 23, 24; 4 used pediatric scenarios9, 12, 25, 27; 4 used neonatal scenarios.8, 11, 13, 19 Table 1 contains the characteristics of these included studies.
Fig. 1.
PRISMA Diagram.
Table 1.
Data extraction table.
| Study, year | Study Design | Subjects and Setting | Number of participants | Content | Outcome measures | Results |
|---|---|---|---|---|---|---|
| Aquel, 2014 | RCT | Nursing students, Jordan | 90 (45 per group) | BLS | K (pre-post test); S (CPR post test); both outcomes at completion and 3 months |
Course conclusion CPR skills: CG 11.58 (1.63) vs IG 13.13 (1.01) (p < 0.001) Knowledge: CG 11.22 (0.90) vs IG 12.67 (1.07) p < 0.001). 3 months: CPR Skills: CG 10.31(1.88) vs IG 12.80 (1.44) (p < 0.001); Knowledge: CG 10.07(1.44) vs IG 12.27 (1.14)p < 0.001) |
| Campbell, 2009 | RCT | 1st year residents, Canada | 15 (IG 8, CG 7) | NRP, 2 scenarios | S (number of redirections from facilitator; time to task completion; time to intubation) | Fewer redirections from facilitator in IG; no difference in time to intubate or other tasks |
| Cheng, 2013 | RCT (factorial design) | Resident teams, USA | IG 45 teams; CG 45 teams | Single PALS scenario | S (clinical performance (CPT); teamwork (BAT)); K (MCQ test) | No significant differences |
| Conlon, 2014 | RCT | EM residents, USA | 54 (IG 18; CG 36 (18 in lo-fi, 18 in mid-fi)) | ACLS | S (clinical performance); K (written exam) | No significant differences |
| Curran, 2015 | RCT | Medical students, Canada | IG 31; CG 35 | NRP | S (clinical, teamwork) | No significant differences |
| Donoghue, 2009 | RCT | Residents, USA | IG 25; CG 26 | PALS scenarios | S (clinical performance) | Greater skill improvement in IG (p = 0.007) |
| Finan, 2012 | RCT | Neonatal fellows, Canada | 16 (crossover) | NRP scenarios | S (megacode performance; ANTS score) | No significant differences |
| Hoadley, 2009 | RCT | Allied heath professionals, USA | IG 29; CG 24 | ACLS | S (clinical scenarios); K (Written test) | No significant differences |
| King, 2011 | RCT | Nursing students, USA | 49 (IG 24, CG 25) | ACLS | K (written test); S scenario checklists at 2 wk and 2 mons | No significant differences |
| Lo, 2011 | RCT | Medical students, USA | 86 (IG 45, CG 41) | ACLS | S (megacode performance at completion and at 1 year) | At completion: IG > CG (83% vs 70%, p < 0.001); at one year NSD |
| Massoth 2019 | RCT | Medical students, Germany | 135 (IG 67, CG 68) | ALS | K: written test; S: practical performance assessed by video analysis; A: self-assessment, confidence | Knowledge: No significant difference (p > 0.05). Skill: CG superior in breathing control (p = 0.012), continous chest compression while charging the defibrillator (p = 0.017), electrocardiogram analysis (p = 0.021), time interval between electric shocks (p = 0.016). |
| McCoy 2018 | RCT | Medical students, USA | 70 (IG 35, CG 35) | BLS | S: compression rate, depth, recoil, compression fraction; time to EMS activation | Skill: mean compression depth IG 4.57 cm vs CG 3.89 cm (p = 0.02) Compression fraction IG 72% vs CG 70% (p = 0.01) Mean compression rate IG 123.3 cpm vs CG 116 cpm (p = 0.06) Full recoil: NSD Time to activation of EMS: IG 24.7 sec vas CG 79.5 s (p = 0.007) |
| Nimbalkar, 2015 | RCT | Medical students, India | 101 (IG 51, CG 53) | NRP | S: NRP megacode scenario; K: NRP written test (both at course completion and 3 mos) | No significant differences |
| Owen, 2006 | RCT | Residents, Australia | 41 (IG 20, CG 21) | Adult cardiac arrest scenario | S: clinical performance score; K: written test | K: IG > CG (p = 0.026); S: NSD |
| Rishipathak 2020 | RCT | EMS students, India | 100 (unsure of group size) | ACLS | A (learner satisfaction and confidence) | Higher confidence in IG; higher preference (2 out of 5 items) in IG |
| Roh 2014 | Observational study | nursing students, South Korea | 163 (IG 28, CG 135) | Adult cardiac arrest scenario | A: self-efficacy scale | Greater improvement in self-efficacy scores in IG (p < 0.001) |
| Rogers, 2009 | Observational study | Nursing students, USA | 34 (IG 16; CG 18) | ACLS | S: scenario performance; K: written test | S: higher scenario score in IG (p = 0.01); K: greater pre- to post- test improvement (p = 0.002) |
| Settles, 2011 | RCT | Allied health professionals, USA | 146 (IG 73, CG 73) | ACLS | S: scenario performance (3–6 mos, 6–9 mos); K: written test | No significant differences |
| Stellflug, 2018 | RCT | Healthcare providers, USA | 94 (IG 28; CG 66) | PALS | S: time to task performance; Knowledge: PALS written exam | Skill: NSD; Knowledge: NSD btw groups at course completion; improved knowledge retention at 6 months post course in IG group (exam score 27.7 vs. 26.2, p = 0.042) |
| Thomas, 2010 | RCT | Residents, USA | 62 (IG 31, CG 31) | NRP | S: NRP megacode score | No significant differences |
| Tufts, 2021 | RCT | Medical students, USA | 50 (IG 27, CG 23) | PALS scenario | S: resuscitation task completion; A: self-reported confidence | Skill: IG > CG in 4 of 11 specific tasks (p < 0.001 for all); self confidence IG > CG in 3 of 14 domains |
Abbreviations: RCT: randomized controlled trial; S: skill; K: knowledge; A: attitude; IG: intervention group (higher fidelity); CG: control group (lower fidelity); BLS: Basic life support; CPR: cardiopulmonary resuscitation; NRP: Neonatal Resuscitation Program; PALS: Pediatric advanced life support; ACLS: Advanced cardiac life support; EM: emergency medicine; NSD: no significant difference; PPV: positive pressure ventilation; MCQ: multiple choice questions.
Risk of bias assessments are shown in Table 2; all studies were found to have at least some concern for bias. Thirteen studies (twelve RCTs and one observational study) had previous assessment of bias when they were included in the 2015 systematic review3; the risk of bias assessment of the eight additional included studies are reported in Table 2.
Table 2.
Assessment of bias tables for studies from 2015 to present.
| RCTs (RoB 2) | |||||||
|---|---|---|---|---|---|---|---|
| 1st Author | Year | Randomization | Deviations from interventions | Missing data | Outcome measurement | Outcome reporting | Overall |
| Aqel | 2014 | Some concerns | Some concerns | Some concerns | Some concerns | Some concerns | Some concerns |
| Massoth | 2019 | Low | Low | Low | Some concerns | Some concerns | Some concerns |
| McCoy | 2018 | Low | Low | Low | Some concerns | Some concerns | Some concerns |
| Nimbalkar | 2015 | Low | Low | Low | Some concerns | Low | Some concerns |
| Rishipithak | 2020 | Some concerns | Some concerns | High | High | Some concerns | High |
| Stellflug | 2018 | Some concerns | Low | Low | Some concerns | Low | Some concerns |
| Tufts | 2021 | Some concerns | Low | Low | Some concerns | Some concerns | Some concerns |
| Observational studies (ROBINS-I) | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| 1st Author | Year | Confounding | Participant selection | Intervention classification | Deviations from interventions | Missing data | Outcome measurement | Outcome reporting | Overall |
| Roh | 2022 | Serious | Low | Low | Low | Serious | Serious | Low | Serious |
The GRADE summary of evidence is shown in Table 3. Using GRADE criteria, evidence for each of the outcomes was found to be of low or very low certainty.
Table 3.
GRADE Summary of Evidence.
|
Question: Higher fidelity manikins compared to lower fidelity manikins for life support education Setting: Life support education settingsBibliography: | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
|
Certainty assessment |
№ of patients |
Effect |
Certainty | Importance | ||||||||
| № of studies | Study design | Risk of bias | Inconsistency | Indirectness | Imprecision | Other considerations | higher fidelity manikins | lower fidelity manikins |
Relative (95% CI) |
Absolute (95% CI) |
||
| Skill −- megacode/scenario performance at course conclusion | ||||||||||||
| 89, 10, 12, 13, 14, 16, 19, 24 | randomised trials | seriousa | seriousb | not serious | seriousb | none | 279 | 271 | − | SMD 0.66 SD higher (0.08 higher to 1.25 higher) |
⊕○○○ Very lowa, b |
CRITICAL |
| Knowledge −- written / MCQ test at course conclusion | ||||||||||||
| 77, 9, 10, 14, 17, 19, 24 | randomised trials | seriousa | seriousb | seriousc | seriousb | none | 511 | 505 | − | SMD 0.24 SD higher (0.23 lower to 0.71 higher) |
⊕○○○ Very lowa, b, c |
CRITICAL |
| Time to clinical task completion (course conclusion) | ||||||||||||
| 38, 18, 25 | randomised trials | seriousa | seriousd | not serious | not serious | none | 71 intervention and 108 control subjects in 3 RCTs examined time to task completion. Two RCTs found faster time-to-task completion (EMS activation (McCoy) and intervention/reassessment cycle (Stellflug)). One RCT (Campbell) found no difference in time to intubation during NRP training. | ⊕⊕○○ Lowa, d |
IMPORTANT | |||
| Skill −- teamwork (course conclusion) | ||||||||||||
| 39, 20, 26 | randomised trials | seriousa | seriousb | not serious | seriousb | none | 96 intervention / 97 control subjects in 3 RCTs examined teamwork with different levels of manikin fidelity. One RCT (Thomas) found improved teamwork behaviors following training with higher fidelity manikins compared with lower fidelity; one RCT (Owen) found improved teamwork behavior during one of three assessment scenarios following training with higher fidelity; one RCT (Cheng) found no difference in teamwork scores between groups. | ⊕○○○ Very lowa, b |
IMPORTANT | |||
| Skill −- clinical performance at 3 months or greater | ||||||||||||
| 37, 16, 24 | randomised trials | seriousa | seriousb | not serious | seriousb | none | 163 intervention / 159 control subjects in 3 RCTs examined clinical skill (scenario-based assessment) at 3 months or greater after training. One RCT (Aqel) in nursing students found better clinical performance in a CPR scenario 3 months after training with higher fidelity manikins; two studies of ACLS skills found no difference at 3 months post training (Lo) or at one year post training (Settles). | ⊕○○○ Very lowa, b |
IMPORTANT | |||
| Knowledge −- written/MCQ test at 3 to 6 months | ||||||||||||
| 37, 24, 25 | randomised trials | seriousa | seriousb | seriousc | seriousb | none | 146 intervention / 184 control subjects in 3 RCTs examined knowledge (written test assessment) at 3 months or greater following training. Two RCTs found improved knowledge following higher fidelity manikin training (3 months after BLS training (Aqel); 6 months after PALS training (Stellflug)); one RCT found no difference in ACLS knowledge at 6–9 months post training (Settles). | ⊕○○○ Very lowa, b, c |
IMPORTANT | |||
| Skill −- CPR parameters at course conclusion | ||||||||||||
| 27, 18 | randomised trials | seriousa | seriousb | not serious | seriousb | none | 2 RCTs with 80 intervention subjects and 80 controls. One RCT found a greater degree of improvement as measured at course completion by the American Heart Association CPR skills checklist among subjects trained on higher fidelity manikins (Aqel); one RCT found better compression depth and compression fraction immediately post training among subjects trained on higher fidelity manikins. (McCoy) | ⊕○○○ Very lowa, b |
IMPORTANT | |||
| Affective responses | ||||||||||||
| 108, 10, 11, 13, 14, 17, 21, 23, 24, 27 | randomised trials | very seriousa, d | seriousb | seriousc | very seriouse | none | 10 RCTs with 347 intervention subjects and 471 controls. Seven RCTs found subjects reporting higher effectiveness of training with higher fidelity manikins (Campbell, Conlon, Curran, Massoth, Rishipathak, Roh, Tufts); three RCTs found no difference. (Finan, Hoadley, Settles) | ⊕○○○ Very lowa, b, c, d, e |
IMPORTANT | |||
CI: confidence interval; SMD: standardised mean difference.
Explanations.
Blinding of participants not possible in RCTs.
Different assessment tools across multiple studies.
No proven correlation between written exam scores and clinical performance.
Different tasks being assessed across studies.
Data from survey instruments measuring different constructs.
Skill performance
Eight RCTs with 279 intervention subjects and 271 controls reported skill at course conclusion assessed by overall clinical performance in a simulated scenario.9, 10, 12, 13, 14, 16, 19, 24 Four RCTs assessed performance in an adult scenario10, 14, 16, 24; two in a pediatric scenario9, 12; and two in a neonatal scenario. 13, 19 Two studies used the NRP Megacode checklist as the outcome of interest13, 19; two used the ACLS Megacode checklist14, 16; the remaining four studies used instruments created or modified for their specific studies. 9, 10, 12, 24 Four of the eight studies included an explicit description of rater training for use of the outcome instruments in the study. 9, 13, 16, 19 On meta-analysis, a significant effect favoring manikins with higher realism was found (standard mean difference 0.66, 95% CI 0.08 – 1.25). A high degree of heterogeneity among studies was found (I2 = 90%). (Fig. 2).
Fig. 2.
Meta-analysis of RCTs reporting scenario-based clinical skill performance at course conclusion.
In addition to the eight RCTs included in the above meta-analysis, two additional RCTs with 51 intervention study participants and 56 controls compared skill at course conclusion but did not report sufficient measures of variance for inclusion in the meta-analysis. Both RCTs found no difference in skill performance at course completion. 11, 20 Four RCTs measured and reported skill performance before and after training in addition to comparing low- and higher-realism manikin groups. 9, 12, 19, 20 All four studies demonstrated improvement in performance post-training when compared with pre-training irrespective of the level of manikin realism.
Three RCTs with 71 intervention study participants and 108 controls examined time to critical task completion at course conclusion.8, 18, 25 Two RCTs found faster time-to-task completion (EMS activation) 18 and intervention/reassessment cycle in simulations with higher realism manikins. 25 One RCT found no difference in time to intubation during NRP training.8
Three RCTs with 96 intervention study participants and 97 controls examined teamwork performance at course conclusion; all three used a separate outcome instrument on which raters were specifically trained.9, 20, 26 One RCT found improved teamwork behaviors following training with higher realism manikins compared with lower realism; one RCT found improved teamwork behavior during one of three assessment scenarios following training with higher realism20; one RCT found no difference in teamwork scores between groups.9
Three RCTs with 163 intervention subjects and 159 controls examined clinical performance at 3 months or longer following course conclusion. 7, 16, 24 One RCT in nursing students found better clinical performance in a CPR scenario 3 months after training with higher realism manikins7; two studies of advanced cardiac life support (ACLS) skills found no difference at 3 months post training or at one year post training.16, 24
Two RCTs with 80 intervention study participants and 80 controls examined specific CPR parameters at course conclusion. One RCT found a greater degree of improvement as measured at course completion by the American Heart Association CPR skills checklist among study participants trained on higher realism manikins7; one RCT found better compression depth and compression fraction immediately post training among study participants trained on higher realism manikins.18
Knowledge acquisition and retention
Seven RCTs with 511 intervention study participants and 505 controls reported on knowledge at course conclusion as measured by a written test. 7, 9, 10, 14, 17, 19, 24 Five RCTs assessed performance in an adult scenario7, 10, 14, 17, 24; one in a pediatric scenario,9 and one in a neonatal scenario.19 On meta-analysis, no effect favoring higher realism manikins was found (standard mean difference 0.24, 95% CI −0.23 – 0.71). Significant heterogeneity was again found amongst these studies (I2 = 92%).(Fig. 3).
Fig. 3.
Meta-analysis of RCTs reporting knowledge assessment at course conclusion.
In addition to the seven RCTs included in the above meta-analysis, three additional RCTs with 72 intervention study participants and 112 controls and one observational study of 34 study participants compared knowledge at course conclusion but did not report sufficient measures of variance for inclusion in metaanalysis.15, 20, 22, 25 One RCT found improved knowledge at course completion20; two RCTs and one observational study found no difference.15, 22, 25
Three RCTs with 146 intervention study participants and 184 controls reported on knowledge assessed at 3 months or longer after course conclusion. 7, 24, 25 Two RCTs found improved knowledge following higher realism manikin training (3 months after BLS training7; 6 months after PALS training25); one RCT found no difference in ACLS knowledge at 6–9 months post training.24
Attitudes
Ten RCTs with 347 intervention study participants and 471 controls reported on learner preference and/or confidence at course conclusion.8, 10, 11, 13, 14, 17, 21, 23, 24, 27 Seven RCTs found study participants reporting higher effectiveness of training with higher realism manikins8, 10, 11, 17, 21, 23, 27; three RCTs found no difference.13, 14, 24
Clinical outcomes, costs and resources
No studies included in the review examined care delivery to patients, patient clinical outcomes, cost efficiency, or resource utilization.
Discussion
This systematic review provides low to very low certainty evidence that using manikins with higher degrees of physical realism during life support training may lead to improved educational outcomes. In particular, clinical skill performance in a simulated scenario at course conclusion was found to be significantly improved in learners trained on higher realism manikins. We found no effect of manikin realism on knowledge at course completion. Our findings support and strengthen the recommendations yielded by the 2015 systematic review by Cheng et al. examining this same research question.3
All of the included studies had significant risk of bias, and large heterogeneity was found among pooled studies in meta-analyses. Additionally, significant variability existed between studies with respect to instructional content, learner group backgrounds and experience levels, and the types of assessment method(s) used. Aside from the outcomes subjected to meta-analysis, the other outcomes (e.g. CPR parameters, task performance) were reported only by three or fewer studies; even in these small subsets, there were differences in the study population and life support course being studied. All of these issues make drawing strong conclusions about the effect of manikin realism difficult. A total of six studies reported either skill or knowledge at remote time points (as opposed to at course conclusion); this subset of studies showed a similar degree of variability, making conclusions about the durability of learning and skill and/or knowledge decay impossible.
The recommendations made by ILCOR in 2020 for use of higher realism manikins stipulate that they should be used by centers that have sufficient resources to implement them in an optimal way.28, 29 This includes financial resources for the purchase and maintenance of such simulators; infrastructure such as space and information technology; and personnel with sufficient training and expertise. Importantly, none of the included studies in this review directly addressed the need for financial, infrastructural, and personnel resources associated with higher realism manikin use. It is not possible to extrapolate these recommendations to settings lacking these resources. A recent systematic review by Issa et al. examined the impact of manikin realism across all medical education areas in lower and middle income countries and found no significant benefit in the used of higher realism training models, leading to a recommendation that lower realism models be used based on the associated resource requirements.30 While Issa and colleagues did not focus exclusively on life support education, it is possible that the cost necessary to use such technology outweighs the educational benefit in such settings. Further studies should more closely examine resource requirements for resuscitation training with more realistic manikins and determine which manikin features are the most contributory for learning for the most learners.
Limitations, knowledge gaps, and future research
Several limitations to the findings of this review should be acknowledged. As described above, the studies included in this review demonstrated marked heterogeneity, making summative conclusions about the impact of manikin realism on educational outcomes difficult. Data with sufficient detail for meta-analysis was reported in eight RCTs for skill and seven RCTs for knowledge at course conclusion; however, given that these assorted studies were conducted using differing content (e.g. PALS cs. NRP vs. ALS) a significant risk of bias remains based on inconsistency. The application of unified assessment of the learning outcome in future studies is highly recommended to enable comparative studies and meta-analysis.
The results of this review highlight several important knowledge gaps with regard to manikin realism. The majority of included studies were performed at single institutions; determining the generalizability of these results would require studying the uniformity across institutions of the types of manikins used and the associated instructor training. Higher realism manikins are associated with significant cost; no included studies attempted to examine cost effectiveness or return on investment. Finally, as discussed above, the substantial resource requirements for implementing training with higher realism manikins, in terms of cost, personnel, and infrastructure, make it impossible to translate these findings to resource-limited settings.
Conclusions
This systematic review provides very low certainty evidence supporting the use of manikins with higher degrees of physical realism during life support education. In one meta-analysis, an association between higher manikin realism use and improved clinical performance at course conclusion was found. A high degree of risk of bias and heterogeneity among studies makes inference of a stronger association with learning outcomes difficult. Future studies should examine the impact of the resource requirements for higher realism manikin use as well as the features in such manikins that maximize functional task alignment with specific learner needs and objectives.
Conflicts of interest
Each of the authors are members of of ILCOR’s task force on Education, Implementation, and Teams. RG is the ERC Director of Guidelines and ILCOR, and Chair of ILCOR’s task force on Education, Implementation, and Teams. RG, AL and AC are Editorial Board members for Resuscitation Plus. AD and AC are first authors of two of the included studies in the review and did not participate in risk of bias assessment and GRADE rating, other members of the task force writing group reviewed and extracted these studies. AC and AL are co-authors of the previous ILCOR systematic review from which data on 12 of the included studies was extracted; two separate authors (AD and KA) were responsible for the assessment of quality of that review using A Measurement Tool to Assess Systematic Reviews (AMSTAR-2) prior to its inclusion.31
CRediT authorship contribution statement
Aaron Donoghue: Writing – review & editing, Writing – original draft, Visualization, Validation, Methodology, Investigation, Formal analysis, Data curation, Conceptualization. Katherine Allan: Writing – review & editing, Writing – original draft, Validation, Methodology, Data curation, Conceptualization. Sebastian Schnaubelt: Writing – review & editing, Writing – original draft, Validation, Methodology, Investigation, Data curation, Conceptualization. Andrea Cortegiani: Writing – review & editing, Writing – original draft, Validation, Methodology, Investigation, Data curation, Conceptualization. Robert Greif: Writing – review & editing, Writing – original draft, Validation, Supervision, Project administration. Adam Cheng: Writing – review & editing, Writing – original draft, Validation, Methodology, Investigation, Data curation, Conceptualization. Andrew Lockey: Writing – review & editing, Writing – original draft, Validation, Methodology, Formal analysis, Data curation, Conceptualization.
Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgements
We would like to thank Jennifer Lege-Matsuura, MSLIS, AHIP, and Elizabeth S. Blake, MLS, AHIP, both from the Holman Biotech Commons at the University of Pennsylvania, Philadelphia, PA, USA; and Julie Reed, MA, MLIS, from the Department of Continuing Education at the Children’s Hospital of Philadelphia, for their support in developing, reviewing, and conducting the search strategies.
The following ILCOR EIT Taskforce members are acknowledged as collaborators on this systematic review: Cristian Abelairas-Gómez, Jan Breckwoldt, Kathryn Eastwood, Barbara Farquharson, Ming-Ju Hsieh, Tracy Kidd, Ying-Chih Ko, Kasper G. Lauridsen, Yiqun Lin, Tasuku Matsuyama, Sabine Nabecker, Kevin Nation, Alexander Olaussen, Chih-Wei Yang, and Joyce Yeung.
Footnotes
Supplementary data to this article can be found online at https://doi.org/10.1016/j.resplu.2025.100940.
Appendix A. Supplementary material
The following are the Supplementary data to this article:
References
- 1.Soreide E., Morrison L., Hillman K., et al. The formula for survival in resuscitation. Resuscitation. 2013;84:1487–1493. doi: 10.1016/j.resuscitation.2013.07.020. 20130803. [DOI] [PubMed] [Google Scholar]
- 2.Cheng A., Nadkarni V.M., Mancini M.B., et al. Resuscitation education science: educational strategies to improve outcomes from cardiac arrest: a scientific statement from the american heart association. Circulation. 2018;138:e82–e122. doi: 10.1161/CIR.0000000000000583. [DOI] [PubMed] [Google Scholar]
- 3.Cheng A., Lockey A., Bhanji F., et al. The use of high-fidelity manikins for advanced life support training–A systematic review and meta-analysis. Resuscitation. 2015;93:142–149. doi: 10.1016/j.resuscitation.2015.04.004. 20150414. [DOI] [PubMed] [Google Scholar]
- 4.Sterne J.A.C., Savovic J., Page M.J., et al. RoB 2: a revised tool for assessing risk of bias in randomised trials. BMJ. 2019;366:l4898. doi: 10.1136/bmj.l4898. 20190828. [DOI] [PubMed] [Google Scholar]
- 5.Sterne J.A., Hernan M.A., Reeves B.C., et al. ROBINS-I: a tool for assessing risk of bias in non-randomised studies of interventions. BMJ. 2016;355(i4919) doi: 10.1136/bmj.i4919. 20161012. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Cohen J. Routledge; 2013. Statistical power analysis for the behavioral sciences. [Google Scholar]
- 7.Aqel A.A., Ahmad M.M. High-fidelity simulation effects on CPR knowledge, skills, acquisition, and retention in nursing students. Worldviews Evid Based Nurs. 2014;11:394–400. doi: 10.1111/wvn.12063. [DOI] [PubMed] [Google Scholar]
- 8.Campbell D.M., Barozzino T., Farrugia M., et al. High-fidelity simulation in neonatal resuscitation. Paediatr Child Health. 2009;14:19–23. doi: 10.1093/pch/14.1.19. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Cheng A., Hunt E.A., Donoghue A., et al. Examining pediatric resuscitation education using simulation and scripted debriefing: a multicenter randomized trial. JAMA Pediatr. 2013;167:528–536. doi: 10.1001/jamapediatrics.2013.1389. [DOI] [PubMed] [Google Scholar]
- 10.Conlon L.W., Rodgers D.L., Shofer F.S., et al. Impact of levels of simulation fidelity on training of interns in ACLS. Hosp Pract (1995) 2014;42:135–141. doi: 10.3810/hp.2014.10.1150. [DOI] [PubMed] [Google Scholar]
- 11.Curran V., Fleet L., White S., et al. A randomized controlled study of manikin simulator fidelity on neonatal resuscitation program learning outcomes. Adv Health Sci Educ Theory Pract. 2015;20:205–218. doi: 10.1007/s10459-014-9522-8. [DOI] [PubMed] [Google Scholar]
- 12.Donoghue A.J., Durbin D.R., Nadel F.M., et al. Effect of high-fidelity simulation on Pediatric Advanced Life Support training in pediatric house staff: a randomized trial. Pediatr Emerg Care. 2009;25:139–144. doi: 10.1097/PEC.0b013e31819a7f90. [DOI] [PubMed] [Google Scholar]
- 13.Finan E., Bismilla Z., Whyte H.E., et al. High-fidelity simulator technology may not be superior to traditional low-fidelity equipment for neonatal resuscitation training. J Perinatol. 2012;32:287–292. doi: 10.1038/jp.2011.96. [DOI] [PubMed] [Google Scholar]
- 14.Hoadley T.A. Learning advanced cardiac life support: a comparison study of the effects of low- and high-fidelity simulation. Nurs Educ Perspect. 2009;30:91–95. [PubMed] [Google Scholar]
- 15.King J.M., Reising D.L. Teaching advanced cardiac life support protocols: the effectiveness of static versus high-fidelity simulation. Nurse Educ. 2011;36:62–65. doi: 10.1097/NNE.0b013e31820b5012. [DOI] [PubMed] [Google Scholar]
- 16.Lo B.M., Devine A.S., Evans D.P., et al. Comparison of traditional versus high-fidelity simulation in the retention of ACLS knowledge. Resuscitation. 2011;82:1440–1443. doi: 10.1016/j.resuscitation.2011.06.017. [DOI] [PubMed] [Google Scholar]
- 17.Massoth C., Röder H., Ohlenburg H., et al. High-fidelity is not superior to low-fidelity simulation but leads to overconfidence in medical students. BMC Med Educ. 2019;19:29. doi: 10.1186/s12909-019-1464-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.McCoy C.E., Rahman A., Rendon J.C., et al. Randomized controlled trial of simulation vs. standard training for teaching medical students high-quality cardiopulmonary resuscitation. West J Emerg Med. 2019;20:15–22. doi: 10.5811/westjem.2018.11.39040. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Nimbalkar A., Patel D., Kungwani A., et al. Randomized control trial of high fidelity vs low fidelity simulation for training undergraduate students in neonatal resuscitation. BMC Res Notes. 2015;8:636. doi: 10.1186/s13104-015-1623-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Owen D.D., McGovern S.K., Murray A., et al. Association of race and socioeconomic status with automatic external defibrillator training prevalence in the United States. Resuscitation. 2018;127:100–104. doi: 10.1016/j.resuscitation.2018.03.037. [DOI] [PubMed] [Google Scholar]
- 21.Rishipathak P., Hinduja A., Sengupta N. A comparative analysis of self-efficacy in low fidelity vs high fidelity simulation post advanced cardiac life support (ACLS) sessions on cardiac arrest algorithm amongst EMS students of Pune, India. Indian Journal of Public Health Research and Development. 2020;11:415–419. doi: 10.37506/v11/i1/2020/ijphrd/193853. [DOI] [Google Scholar]
- 22.Rodgers D.L., Securro S., Jr., Pauley R.D. The effect of high-fidelity simulation on educational outcomes in an advanced cardiovascular life support course. Simul Healthc. 2009;4:200–206. doi: 10.1097/SIH.0b013e3181b1b877. [DOI] [PubMed] [Google Scholar]
- 23.Roh Y.S. Effects of high-fidelity patient simulation on nursing students' resuscitation-specific self-efficacy. Comput Inform Nurs. 2014;32:84–89. doi: 10.1097/cin.0000000000000034. [DOI] [PubMed] [Google Scholar]
- 24.Settles J., Jeffries P.R., Smith T.M., et al. Advanced cardiac life support instruction: do we know tomorrow what we know today? J Contin Educ Nurs. 2011;42(271–279) doi: 10.3928/00220124-20110315-01. [DOI] [PubMed] [Google Scholar]
- 25.Stellflug S.M., Lowe N.K. The effect of high fidelity simulators on knowledge retention and skill self efficacy in pediatric advanced life support courses in a rural state. J Pediatr Nurs. 2018;39:21–26. doi: 10.1016/j.pedn.2017.12.006. [DOI] [PubMed] [Google Scholar]
- 26.Thomas E.J., Williams A.L., Reichman E.F., et al. Team training in the neonatal resuscitation program for interns: teamwork and quality of resuscitations. Pediatrics. 2010;125:539–546. doi: 10.1542/peds.2009-1635. [DOI] [PubMed] [Google Scholar]
- 27.Tufts L.M., Hensley C.A., Frazier M.D., et al. Utilizing high-fidelity simulators in improving trainee confidence and competency in code management. Pediatr Qual Saf. 2021;6:e496. doi: 10.1097/pq9.0000000000000496. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Greif R., Bhanji F., Bigham B.L., et al. Education, implementation, and teams: 2020 international consensus on cardiopulmonary resuscitation and emergency cardiovascular care science with treatment recommendations. Resuscitation. 2020;156:A188–A239. doi: 10.1016/j.resuscitation.2020.09.014. 20201021. [DOI] [PubMed] [Google Scholar]
- 29.Greif R., Bhanji F., Bigham B.L., et al. Education, implementation, and teams: 2020 international consensus on cardiopulmonary resuscitation and emergency cardiovascular care science with treatment recommendations. Circulation. 2020;142:S222–S283. doi: 10.1161/CIR.0000000000000896. 20201021. [DOI] [PubMed] [Google Scholar]
- 30.Issa M., Furia F., Whaiba A., et al. Physical realism of simulation training for health care in low- and middle-income countries-a systematic review. Simul Healthc. 2024;19:S41–S49. doi: 10.1097/SIH.0000000000000761. [DOI] [PubMed] [Google Scholar]
- 31.Shea B.J., Reeves B.C., Wells G., et al. AMSTAR 2: a critical appraisal tool for systematic reviews that include randomised or non-randomised studies of healthcare interventions, or both. BMJ. 2017;358:j4008. doi: 10.1136/bmj.j4008. 20170921. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.



