Access provided by
London Metropolitan University
• To recognise the need for emergency nurses to receive high-quality triage training
• To be aware that lectures, simulations and workshops can be effective training interventions for improving triage practice
• To appreciate there is a need for further research to ensure that the evidence base for triage training is effective for application to practice
Triage accuracy is important to ensure effective treatment and management of patients in the emergency department, however this requires nurses to receive high-quality triage training. This article reports the results of a scoping review that aimed to establish what research on triage training exists and what research is required to improve such training. Sixty-eight studies which used a range of training interventions and outcome measurements were reviewed. The authors conclude that the heterogeneity of these studies makes comparison challenging and that this, combined with low methodological quality, requires caution when applying the results in practice. The authors recommend establishing a gold standard for measuring triage training outcomes.
Emergency Nurse. doi: 10.7748/en.2023.e2163Peer review
This article has been subject to external double-blind peer review and checked for plagiarism using automated software
Gorick H, Rai AS (2023) Training nurses to triage: a scoping review. Emergency Nurse. doi: 10.7748/en.2023.e2163
Published online: 04 May 2023
Triage is an important interaction for patients in emergency departments (EDs), representing an assessment of their acuity and setting the course for their journey through the hospital (Gorick 2022). Accurate triage ensures that patients receive the treatment they need in the time frame in which they need it, thus preventing patient harm and ensuring effective flow through the ED (Farrohknia et al 2011, Yurkova and Wolf 2011).
A systematic review of triage performance in emergency medicine found that errors occur at unacceptably high levels and identified the need to improve interrater reliability and triage performance in terms of identifying patients at risk of adverse outcomes (Hinson et al 2019). Research that explored whether postgraduate qualifications make a difference to triage accuracy found no significant effect (Ekins and Morphet 2015, Jordi et al 2015), while Tam et al (2018), in a review of triage accuracy, recommended that staff should receive monthly refresher training to increase accuracy and improve patient outcomes.
Triage training is vital to ensure triage accuracy, but it is important to understand what research is available regarding training interventions and what further research is required. Undertaking a scoping review of a topic enables assessment of current knowledge and can help to direct future developments (Pham et al 2014). However, a search undertaken by the authors before the one described in this article identified only one scoping review of nurse triage training (Hardy and Calleja 2019) and no systematic reviews. The authors therefore conducted a scoping review to assess current knowledge of triage training.
The aim of this scoping review was to establish what research exists and what future research is required to improve nurse triage training.
• Triage training is vital to ensure triage accuracy
• To improve triage requires an understanding of what research of training interventions is available
• Lectures, simulations and workshops appear to be effective in improving triage practice and may provide direction for development of future training programmes
• Future training programmes should undergo rigorous validation with comparisons against controls to demonstrate their effectiveness
• Establishing a gold standard for measuring triage training outcomes would focus future research
The scoping review followed the six-stage format outlined by Arksey and O’Malley (2005): specify the research question; identify relevant literature; select studies; map out the data; summarise, synthesise and report the results; and include expert consultation (optional). In this review the optional sixth stage was not included.
The research question was: ‘What is the available research about training interventions for triage in the ED?’. Relevant studies were identified using the search terms shown in Table 1.
|Nurs*||Emergenc* ADJ2 Department*||Triag*||Teac*|
|Nurse (MeSH)||Accident* ADJ2 Emergenc*||Acuit*||Trai*|
|A&E||Sever* adj3 ill*||Educat*|
|‘A & E’||Triage (MeSH)||Lesso*|
|Emergency department (MeSH)||Patient acuity (MeSH)||Improv*|
Searches were not restricted by date or geography to capture all available research. Although this meant studies that might not be considered current were included, it reduced the risk of excluding relevant research. Initially language was not restricted, but studies in languages other than English where a high-quality translation could not be obtained were excluded for practical reasons.
Searches for relevant studies were made on Ovid MEDLINE, Cumulative Index to Nursing and Allied Health Literature (CINAHL), Cochrane and the British Nursing Index using targeted search strategies up to 27 September 2022. Relevant conference proceedings and organisations were also searched, including: OpenGrey; EThOS; the King’s Fund; the World Health Organization; the National Institutes of Health; the National Institute for Health and Care Excellence; Manchester Triage Group; the Australasian College for Emergency Medicine; and the Emergency Nurses Association. Citations in studies were checked for relevance.
Citations for all studies were exported to EndNote 20 reference management tool for duplicate removal and organisation before transfer to Rayyan software for screening. Full texts were obtained where possible through the authors’ university library; where these were not available, intra-library loans were obtained and the lead authors were contacted for manuscripts.
Inclusion and exclusion criteria (Box 1) were developed to screen included citations. Primary and secondary research on triage training was included except for studies that surveyed levels of training, as they did not describe the training. Where data were reported in multiple sources, for example in a thesis and a publication, the most complete set of data was used.
Screening took place using Rayyan, with initial title and abstract screening followed by full-text screening using the criteria in Box 1.
The authors screened each study blinded to each other’s decisions, compared results and discussed any disagreements until both agreed. Kappa for title and abstract screening was 0.82 (95% confidence interval (CI) 0.79-0.85) and for full-text screening was 0.84 (95% CI 0.75-0.94).
Data were extracted from included studies for analysis using a standardised form in a spreadsheet programme. Both authors initially extracted from the same ten studies then checked agreement (Kappa 0.85 95% CI 0.58-1.12). As there was almost perfect agreement, the rest of the studies were charted individually.
Methodological analysis was undertaken using critical appraisal tools from the Joanna Briggs Institute (https://jbi.global/critical-appraisal-tools) relevant to the study type. Appraisal was undertaken by both authors while blinded to each other’s decisions, then compared, with differences discussed and resolved (Kappa 0.8 95% CI 0.76-0.85). Overall methodological quality of studies was rated using a five-point scale – very high, high, moderate, low, very low.
Results were analysed following the methods proposed by Arksey and O’Malley (2005), first by quantitative analysis of the extent, nature and distribution of the studies using descriptive statistics and second by narrative analysis of the content to gauge the extent of the research.
Database searches yielded 6,929 records, which was reduced to 3,667 following duplicate removal. Title and abstract screening resulted in 115 reports being moved to full-text review. Full texts could not be retrieved for nine reports, six were not available in English and 34 were not about triage training and were therefore excluded, leaving 66 studies. Three records from websites, 20 from organisations and five from citation searches were added, resulting in 28 reports for full-text screening of which 26 were excluded as they were not about triage training, leaving two studies. A total of 68 studies was included in the review.
General characteristics of the studies reviewed are presented in Table 2.
|Type of triage studied||Overall process of triage||51||75|
|Triage of specific illness||17||25|
|Place of publication||UK||3||4|
A summary of the methodologies used in the studies reviewed, and methodological quality, is shown in Table 3. Overall, methodological quality was low, with only 25 (37%) studies rated as high or very high quality.
|Randomised controlled trials||6||9|
|No specific design||8||12|
|Training intervention type (24 studies used a combination or comparison of interventions)||Lectures||29||43|
|Combination of interventions||13||19|
|Comparison of interventions||11||16|
|Outcome measurement (multiple outcome measurement tools were used in some studies)||Tested levels of knowledge||17||25|
|Self-reported levels of knowledge||8||12|
|Accuracy in simulated triage||17||25|
|Accuracy in actual triage||15||22|
|Emergency department key performance indexes||10||15|
|Evaluation of training||18||26|
|No outcome measurement||4||6|
|Methodological quality||Very high||5||7|
Figure 1 shows a graphical representation of the disposition of the studies stratified by training intervention, outcome measurement and methodological quality.
The design in most of the studies reviewed was interventional (n=54, 79%), with 42 (62%) studies using a quasi-experimental design. Of these 42 studies, 38 (90%) used pre-/post-testing, but only eight (19%) used comparison groups, controls or both (supporting data are available from the corresponding author). In 14 (33%) of the 42 quasi-experimental studies, study follow up was not completed or was not adequately analysed, in 16 (38%) outcomes were not measured in reliable ways and in 13 (31%) there was a lack of appropriate statistical analysis.
Only six (9%) of the 68 studies reviewed were randomised controlled trials (RCTs), none of which used blinding. Of these six RCTs, two (33%) did not measure outcomes in reliable ways and two (33%) did not describe the characteristics of the treatment groups sufficiently to be able to assess similarity at baseline. Of the remaining 62 studies, four (6%) were cross-sectional, one (2%) was a case series and one (2%) was a cost analysis. Of the 14 non-interventional studies, four (29%) used a qualitative design, two (14%) were literature reviews and eight (57%) used no specific design.
A total of 13 (19%) studies used a combination of training interventions, but only 11 (16%) compared interventions, of which six (55%) were RCTs.
A total of 29 (43%) of the studies reviewed used multiple outcome measurements, 35 (51%) used one and four (6%) did not use any. Of the 58 studies that used quantifiable outcome measurements, eight (14%) showed no statistically significant outcomes. However, when only those with high or very high methodological quality were examined (n=25) this figure reduced to four (16%). When this group was limited further to those that used a comparison between interventions (n=9), only one (11%) showed no significant differences.
However, the outcome measurements used to represent triage accuracy were often abstracted versions of participants’ ability to triage using tested and self-reported knowledge of triage, ED key performance indexes (KPIs) and self-reported confidence in triage ability. Although most of the outcome measurements were validated for use in research, using only these tools to represent triage ability risks providing incorrect measurements (Coster 2013). This is because rather than direct measurements of triage ability, they measure factors that influence nurses’ ability to accurately undertake triage. Furthermore, some studies only measured participants’ perceptions of interventions or mandatory training compliance rather than triage ability which, while useful, provided no indication of the effectiveness of the interventions.
Only 32 (55%) of the 58 studies that used quantifiable outcome measurements directly assessed triage accuracy, but not all of them assessed this in clinical practice, with 17 (53%) having measured triage accuracy in simulation contexts. Of the 25 studies rated as high or very high quality, 13 (52%) directly measured triage accuracy, six (46%) of which were measured in simulation scenarios and seven (54%) measured in clinical practice. Of these 13 studies, only six (46%) used a comparison and three (23%) (one of which used a comparison) had non-significant outcomes.
Only five studies (Kriengsoontornkij et al 2010, Delnavaz et al 2018, Hoseini et al 2018, Recznik 2018, Ghazali et al 2019) were rated as high or very high quality, used a comparison between interventions and/or a control, measured outcomes through assessment of triage accuracy in simulation or clinical practice and had statistically significant outcomes.
Some studies included in this scoping review used qualitative methodologies (n=4, 6%) or literature reviews (n=2, 3%). The studies that applied qualitative methodologies used these to assess the training interventions which, although useful, provides no information on the efficacy of the intervention. One literature review (Doherty 2016) considered peer shadowing but lacked measurable outcomes, while the other literature review (Recznik and Simko 2018) described different methods of interventions for paediatric triage education.
Each of the interventions used in the studies reviewed are discussed below to establish a comprehensive overview of the findings.
Lectures were the most used interventions (n=29, 43%) and ranged from one-off sessions to regular weekly or monthly sessions, lasting from 20 minutes to several hours. Lectures were found to be effective, with most studies showing statistically significant increases in triage ability following the intervention, although Arroabarren et al (2018) was the only high-quality study that featured solely lectures that were not aimed at a specific illness. Grossmann et al (2014) found no significant differences in triage ability following lectures; Olsson et al (2022) also found no significant differences, even after combining lectures with simulations.
Simulation was used in 19 (28%) studies and ranged from low fidelity, for example role-playing and paper cases, to very high fidelity, for example using training manikins, multiple disaster simulation or virtual reality. Efficacy of the simulations was shown to be very high, with most studies reporting significant improvements following the intervention. Studies that used higher fidelity simulations demonstrated greater improvement in participants’ triage ability compared with those that used lower fidelity simulations. Only two studies that explored simulations with no comparisons were rated as high quality – Jang et al (2020), who reported a statistically significant increase in triage ability, and Campbell et al (2022), who noted no significant improvements.
Delnavaz et al (2018) and Hu et al (2021) compared simulations with lectures and reported improved outcomes in both types of intervention, although there was greater improvement in outcomes with the simulation interventions. Recznik (2018) assessed two methods of simulation in cross-over trials and found that participants’ triage accuracy significantly increased in both methods.
Of the studies reviewed, 18 (26%) used workshops as the intervention, although it should be noted that the authors used a broad definition to label the interventions. Workshops consisted of mixed education methods, including group discussion, role-play, skills sessions and educational aids; however, several studies that used the term ‘workshops’ gave no description of what the intervention involved. Three studies that used controls reported significant levels of improvement following a workshop intervention (Hoseini et al 2018, Ghazali et al 2019, Kaiafas and Bennett 2021).
Smith et al (2013) compared lectures with simulations and a combination of simulations and workshops, and reported that all groups showed increased tested triage knowledge. Only one study that explored the use of workshops reported a lack of significant positive outcomes (French et al 2021).
Six (9%) of the studies used online courses, which were generally in-person, face-to-face courses that were delivered online rather than designed for online delivery. One study (Greci et al 2013) used an online interaction to simulate a patient surge, which presented a novel method of teaching triage; however, the quality of the study was low and it measured self-reported knowledge.
Outcome measurements for online interventions mostly assessed user acceptability – with good levels of acceptance – but lacked evaluation of the effectiveness of the intervention. Only four studies examined the effectiveness of online interventions. Two of these, Atack et al (2005) and Rankin et al (2013), assessed triage accuracy; while Atack et al (2005) (moderate quality) reported that the intervention improved accuracy, Rankin et al (2013) (high quality) reported no significant difference between the experimental and control groups. Greci et al (2013) measured self-reported knowledge and reported that this increased following delivery of an online intervention, while Yazdannik et al (2018) compared online courses with workshops by measuring tested knowledge and showed improvements in both interventions, with greater improvements in the online intervention.
Peer shadowing was used in four (6%) studies, three of which combined this with lectures. The interventions consisted of working with more experienced colleagues and learning from their decision-making processes. Kriengsoontornkij et al (2010), who explored peer shadowing combined with lectures, measured triage accuracy in practice and reported significant increases. However, the other studies that considered peer shadowing either lacked measurable outcomes (Jakobsen and Villumsen 2011, Doherty 2016) or used only self-reported confidence as the outcome and was low quality (Baston and Simms 2002).
Specific interventions were discussed in six (9%) studies. One study (Jarvis and de Freitas 2009) compared a ‘serious game’ – that is, a game where the purpose is to educate rather than to entertain – with tabletop exercises, in which the serious game group showed better simulated triage accuracy; however, the study was low quality. Jang et al (2021) investigated the effects of peer-based learning on triage accuracy and found that accuracy improved, although ED waiting times did not change significantly; furthermore, the quality of the study was moderate.
Other specific interventions included reflective practice, which showed significant positive effects on triage accuracy and ED KPIs (Saban et al 2021), and ‘the clock model’ tool for ‘clinical reasoning in the ED’ (Schumaker and Bergeron 2016). The methodological quality of Schumaker and Bergeron’s (2016) study was low however, as they evaluated the intervention rather than its effect on triage accuracy.
Nine reports were excluded from the final review as the full text could not be retrieved and six were excluded because they were not available in English. These 15 studies may have contained relevant information about training in triage.
While the methodological quality of the studies was measured using validated critical appraisal tools, the outcomes are subjective assessments by the authors based on the results of these validated tools. However, these assessments were made by both authors blinded to each other’s decisions and agreement was at a very good level, increasing the rigour.
The results of this scoping review suggest that lectures, simulation and workshops appear to be effective training interventions for improving triage accuracy and may provide direction for the development of future training programmes. However, due to the heterogeneity of the studies, the mix of outcome measurements used – many of which did not directly measure triage accuracy – and methodological quality, the authors advise significant caution before applying these results to practice. The authors also suggest that future training programmes should undergo rigorous validation with comparisons against controls to demonstrate their effectiveness.
The authors recommend establishing a gold standard for measuring triage training outcomes and that triage accuracy should be assessed in clinical practice. Further high-quality research into the effectiveness of training interventions using these outcomes will be required. Finally, a meta-analysis of training interventions may provide good evidence for practice but given the heterogeneity of the research this may prove challenging to undertake.
The current literature on training in triage is heterogeneous, covering a wide variety of interventions and using multiple outcome measurements. Additionally, the methodological quality of studies is low, with only 25 (37%) of those reviewed rated high or very high quality, and most interventional studies lack comparison. There is a need for further research to ensure that the evidence base for training in triage is effective for application to practice. The authors recommend establishing a gold standard for measuring triage training outcomes and that triage accuracy should be assessed in practice. Where this is not possible, assessing accuracy by simulation should be used as an alternative.
Developing an e-learning package to provide chemotherapy updates
Cytotoxic chemotherapy is potentially carcinogenic,...
Improving nurses’ skills through e-learning
This article examines the development of an interactive...
Giving staff confidence to discuss sexual concerns with patients
This article describes a countywide event to raise awareness...
Emergency assessment bay for cancer patients
The annual Quality in Care (QiC) Excellence in Oncology...
Services for women with metastatic breast cancer in the US
This article describes the experience of a nurse on an...