Recruitment

The study sample comprised registered nurses who undertake face-to-face triage of adults as part of their role in UK EDs. Recruitment involved a convenience sampling approach via advertising on social media platforms (X, formerly Twitter, FaceBook, ResearchGate), online nursing forums (Royal College of Nursing emergency care forums) and the first author’s professional networks. As this was an exploratory study with few available statistics on numbers of UK triage nurses, a target sample size of 20-150 was set based on Daniel’s (2011) guidance on sampling choices. The only benchmark identified by the authors was in a survey of Australian triage nurses by Varndell et al (2019) with a sample size of 33.

Data collection

The authors developed an online survey comprising four sections:

The adapted version of the TDMI consisted of 20 statements organised under four subscales – cognitive characteristics, critical thinking, experience and intuition – with each subscale containing five statements. Respondents scored each statement on a five-point Likert-type scale, ranging from ‘strongly agree’ to ‘strongly disagree’.

The quantitative questions on training, the vignette and its associated qualitative questions were developed by the authors based on findings from Gorick et al’s (2023) systematic review. Respondents were asked to rate their satisfaction with the amount, quality and content of triage training provided in their department on a 5-point scale ranging from ‘very satisfied’ to ‘very dissatisfied’. Face and content validity of the quantitative questions on training and the qualitative vignette questions was established via review by two senior researchers from the University of East Anglia and four senior medical staff from a large teaching hospital in the East of England; these questions were also pilot tested with ten nurses working in emergency areas at the teaching hospital, who were representative of the target population.

Data collection ran from 1 June 2023 to 31 July 2023. Respondents accessed the survey via a link embedded in the social media posts or in the online forums or in the email sent to the first author’s professional networks.

Data analysis

Data from the demographics, TDMI and quantitative questions on training were analysed using the Statistical Package for the Social Sciences v28.0.1.1, using both univariate and bivariate descriptive statistics to establish a quantitative perspective of respondents’ decision-making practices. Distribution was assessed using the Shapiro-Wilks test.

Responses to the qualitative vignette questions were analysed using latent content analysis with inductive coding based on Bengtsson’s (2016) framework. Codes were established by all three authors inductively coding independently, being blinded to each other’s decisions, then comparing results, discussing potential themes and agreeing the final themes.

Ethics

Ethical approval was granted by the University of East Anglia Faculty of Medicine and Health Sciences ethics committee (ETH2223-1877, granted 24 May 2023). Informed consent was gained by participants digitally signing a form positioned at the start of the survey. To ensure participants’ anonymity, no identifying information was captured. All data were securely stored on a university computer, with access only available to the authors. Respondent information sheets, consent forms, survey questions and the vignette are available from the first author.

Demographics

The survey was completed by 51 respondents, with a median age of 32 years (interquartile range (IQR) 13). Respondents had a median of seven years since qualifying (IQR 14), a median of six years working in their current ED (IQR 6) and a median of five years working in triage (IQR 7). The respondents used various job titles to describe their role: 19 (37%) used ‘registered nurse’ or variations; 23 (45%) used variations of sister or charge nurse at deputy and senior level; two (4%) were in educator roles; two (4%) were in managerial roles (matron and head of nursing emergency care); and five (10%) were advanced care practitioners. Respondents’ demographics are presented in Table 1.

Triage decision-making inventory

Most respondents strongly agreed or agreed with the statements under each subscale (Figure 1), which suggests they had a high self-reported level of triage proficiency.

Figure 1.

Respondents’ scoring for triage decision-making inventory subscales (n =51)

All statements within each subscale were positively moderately to strongly correlated with each other, with r (51) values of .47-.82 (P=<0.001). However, there was an overall low to no correlation (r=-.2-.26) and no P=<0.05 between demographics related to experience and the TDMI subscales, and only a positive weak significant correlation (r(50) =.29, P=0.041) between Agenda for Change banding and the experience TDMI subscale.

For the statement ‘I feel comfortable making acuity decisions,’ there was a strong positive correlation with number of years since qualifying (r(51) =.57, P=<0.01), number of years working in the ED (r(51) =.48, P=<0.01) and number of years working in triage (r(51) =.52, P=<0.01), suggesting that respondents’ confidence in their triage is linked to experience.

Training

Respondents’ answers to the questions on triage training are shown in Table 2. Of the 39 (76%) respondents who had received training in the past, only ten (26%) had received this training within the last year; 13 (33%) did not feel that this training had prepared them for triage. A total of 31 (61%) respondents said their department delivered training, however the frequency varied and none of the departments provided weekly training.

Of the 14 East of England respondents, eight (57%) had never received triage training compared with four (11%) of the 37 respondents from other areas. Eight (57%) of the 14 respondents from the East of England reported that their department did not provide triage training, compared with ten (27%) of the 37 respondents from other areas.

Respondents’ satisfaction with the training provided is shown in Figure 2. Thirty-six (71%) respondents commented in the desired training section; comments included ‘specific presentations’, ‘assessment techniques’, ‘information about the systems supporting triage’, ‘just more training’. Seven (19%) of these 36 respondents stated they did not want more training.

Figure 2.

Respondents’ satisfaction with their department’s training (n =51)

Vignette

Four themes were generated through analysis of the respondents’ answers to the questions that accompanied the vignette:

This section provides a commentary on respondents’ answers and includes quotes to illustrate the themes.

Experience, stress and training

Some respondents were highly experienced, while others were relatively newly registered and may not have gained the in-depth knowledge and experience required to support safe and effective triage, as identified in Gorick et al’s (2023) systematic review. However, the respondents scored highly across all four subscales of the TDMI, which is consistent with findings from previous surveys of triage nurses (Aktaş and Alemdar 2017, Ghazali et al 2020), suggesting a high self-reported level of triage proficiency.

In the TDMI, respondents had slightly lower levels of intuition compared with the other three TDMI subscales. This contrasts with findings in Aktaş and Alemdar (2017) and Soola et al (2022) where intuition was the highest scoring subscale; however, higher levels of intuition in the triage nurse participants in these studies were associated with completion of enhanced training, which most respondents in the present study had not completed.

The vignette element of the study revealed how respondents experience and react to stress. Previous international research has identified high levels of secondary traumatic stress in ED nurses, which affected their ability to work effectively (Yuwanich et al 2015, Wolf et al 2020). This is echoed in the present study, where some respondents felt anxious due to the large numbers of patients who required triage and their concerns about patient safety, which led them to question their triage abilities. Respondents’ levels of experience influenced their stress levels, coping mechanisms, risk management and assessment practices. More experienced respondents tended to ‘normalise’ stressful situations, while less experienced respondents did not appear to use this as a coping strategy.

Experience, effective triage practice and coping with stressful situations have been found to be linked to training. For example, Gorick et al (2023) found that triage nurses’ knowledge, experience and coping skills strengthen their abilities to make accurate triage decisions and that these elements are gained through formal education and training as well as clinical practice. In addition, Tam et al (2018), in a literature review of triage accuracy, stated that training was necessary to improve nurses’ decision-making and maintain triage accuracy. In the present study, training delivered in respondents’ departments was variable, which reflects the international literature. For example, in Australia, Varndell et al (2019) reported a geographical variation in relation to education requirements of triage nurses, while Wireklint et al (2021), in Sweden, reported variation in type and frequency of triage training for nurses.

In the present study, most respondents had received triage training in the past, however for a large proportion this was at beginner level and only ten (20%) had received this training within the last year. Most respondents were not satisfied with the amount, quality and content of training provided in their department (Figure 2), which supports US (Wolf et al 2018) and Korean (Moon et al 2021) qualitative research that found triage nurses were dissatisfied with training they had received.

Overall, these results, supported by the first author’s previous work (Gorick and Rai 2023, Gorick et al 2023), suggest there is a need for greater investment in triage training for nurses, including further research into what constitutes effective training, to enhance nurses’ confidence, coping strategies and triage practice.

Decision-making practices

The responses to the vignette show that respondents favoured the use of visual assessment supplemented by objective measures, particularly for patients with red-flag symptoms, over the use of triage algorithms. This supports previous research that has reported that nurses prefer using clinical judgement than triage assessment tools (Roscoe et al 2016, Gorick et al 2023).

Respondents also used visual observation to assess the triage environment, then focused on creating space in which to safely assess patients. These elements of practice align with Gorick et al’s (2023) assertion that triage nurses use situational awareness to inform their decisions and take actions based on this awareness to balance patient acuity against situational pressure. Respondents in the present study were resourceful when creating space in which to safely assess patients, for example by ‘curtaining off’ a section of a room or using a spare trolley. This is reflected in a Canadian study of triage nurses’ decision-making (Reay et al 2016), which reported that participants’ need for space resulted in them ‘pushing boundaries’ by using non-clinical areas to ensure patient safety.

In the present study, respondents also pushed boundaries by relying on non-departmental clinical colleagues and non-clinical departmental colleagues. One respondent discussed how they would retain ambulance crews to observe a patient and/or to identify emerging issues in the waiting room, enabling the respondent to focus on other patients. However, delaying ambulance crews when space is available can result in a poor experience and outcomes for the patient transported by the crew to the ED and for those waiting for ambulances in the community (Association of Ambulance Chief Executives 2021). Triage nurses must therefore release ambulance crews quickly, when appropriate, to reduce the risk of patient harm.

NHS England (2022) states that only clinically trained staff should assess patient acuity in the ED, however some respondents described relying on reception staff to alert them if a patient was acutely unwell. The risks of such an approach can be seen in the case of Darnley v Croydon Health Services NHS Trust (2018), where the seriousness of the patient’s condition was missed by an ED receptionist resulting in significant harm to the patient. The supreme court found that the receptionist gave ‘incomplete and misleading information’ which was ruled as negligent (Okninski 2019). It is vital, therefore, that nurses do not rely on non-clinical colleagues to observe patients.

Strengths and limitations

To the best of the authors’ knowledge, this is the first national survey of UK triage nurses and the results are supported by a rigorous methodology.

The recruitment process limited the study sample to those who use specific social media platforms and online forums, which risked self-selection bias. Data were self-reported and may therefore have been subject to observer bias.

Although the numbers of UK triage nurses are unknown, the number of respondents was within the target range for the study sample, with good geographic spread. A significant number of respondents were from the East of England, which may have influenced results, however all data were compared between samples including and excluding the East of England. Only two areas of significant difference were identified: whether respondents had received triage training and whether respondents’ departments provided triage training, but these did not inform further analyses and were not therefore discussed.

Finally, the study did not explore the duration of training or preference of clinical judgement over the use of triage tools in depth, which may have provided greater context for analysis and discussion.