Background

During April 2017 to March 2018 in the UK, there were 23.8 million attendances in the emergency department (ED), a 22% increase since 2008–09. Of these, 12% waited over the target four hours to be discharged or admitted, more than double the expected 5% [1]. In addition, there was a 33% increase of patients attending either minor injuries units (MIU) or walk-in centres during the same time period [1]. To address this increase, it is appropriate to consider radiographer-led discharge (RLD) for patients with minor musculoskeletal injuries [2]. RLD utilises reporting radiographers, trained to either discharge patients with normal images or refer for treatment pathways following pre-specified management plans [3] (Table 1). This innovative pathway was recognised for its potential to reduce the pressure on ED and MIU [2] and first piloted by Snaith in 2007 [3]. Despite being successfully trialled 15 years ago [3], it is still not common practice across the NHS [2].

Table 1 Glossary of terms
Full size table

The NHS plan [4] in 2000 offered the opportunity of role extension for allied health professions (AHPs). Radiographers developed image reporting skills, to the accuracy levels of radiologists [5], leading to improved quality through clinical error reduction [6]. This gave potential for improved efficiency and cost effectiveness; values identified in NHS core principles [4]. Since 2013 image interpretation has been included in undergraduate radiography programmes [7]. This training, extended at post graduate level to advanced practice through Master’s degree programmes, allowed radiographers to report clinically [8]. Nationally in 2017, 78% of hospitals utilised reporting radiographers [8]. RLD also required extension of radiographer training specifically for the discharge process [3].

The Snaith RLD pilot study evidenced a 61% reduction in the patients’ length of stay (LOS) in ED. Patients were discharged or referred for treatment at point of image assessment, by the radiographer with the ED consultant reviewing the outcomes for discharged patients the day after [3]. Hot reporting (Table 1) was also introduced into ED, during the study period, reducing patient recalls, where the radiology report differed with the initial image interpretation, by 52% [3]. The 1.75% patient re-attendance rate also compared favourably to re-attendance rates following discharge by junior doctors (13.1%) and nurse practitioners (8.6%) from other similar studies [3]. However, the pilot also noted there were a further 564 (32%) patients hot reported as normal that RLD was not utilised for, as no management plan was provided at initial assessment [3]. Therefore, with a more robust process there is scope for twofold service improvements with RLD, in line with clinical streaming principles [10]. This is via integrated care benefits and adding value to the patient experience by shortening their journey through ED.

There is potential for improved cost effectiveness in emergency care, with image interpretation errors the leading cause of litigation in ED [2]. For example, use of hot reporting reduced missed fracture litigation claim costs by 66% in one NHS Trust [2]. There could be further cost savings through service streamlining with radiographers discharging patients, increasing ED staff capacity for seeing other patients [2].

A literature review of RLD in 2015 focussed on the impact of RLD on quality of ED services and potential barriers to RLD implementation [11], but did not consider cost effectiveness. Also, the search strategy adopted by the review [11] was not comprehensive. Therefore, it was appropriate to complete a scoping review with additional electronic databases, including grey literature [12]. More recent studies were also available which included ED staff perspectives on RLD, an area for future research identified by the 2015 review [11, 13].

RLD has been recognised as an innovative process [2] and its’ potential demonstrated [3]. Combining the aforementioned studies with recent evidence would allow synthesis of what is currently known about RLD [12]. Thus, conducting a scoping review would be a robust approach, summarising the complete evidence base of RLD for patients with minor musculoskeletal injuries, in emergency care. Synthesising the nature and characteristics of current research would allow identification of further research required to assess the potential feasibility of RLD [12].

Methods

The PRISMA extension for scoping reviews [14] and framework described by Arksey and O’Malley [12] were used as appropriate tools for this scoping review and formed the basis of the review protocol [14]. Initially, the research question and relevant studies were identified, with the included study selection made using predefined eligibility criteria. Finally, the data from the studies were extracted, charted and summarised [12].

Identifying the research question

The research question was required to be broader than expected for a systematic review but sufficiently focussed to identify all relevant literature [12, 14]. A primary question was set, based on the Joanna Briggs Institute mnemonic for scoping reviews of population, concept and context [15]. The specific elements included were the population of radiographers and concept of RLD for patients with minor musculoskeletal (MSK) injuries. The context was urgent or emergency care. Key inclusion and exclusion criteria were also developed (Table 2).

Table 2 Inclusion and exclusion criteria
Full size table

The international prospective register of systematic reviews (PROSPERO) was searched for a protocol based on this topic, once the question was framed [16]. In addition, Pubmed, Medline and Google scholar were also searched for existing systematic and scoping reviews on RLD. No protocols or existing reviews were found.

Eligibility criteria

All study designs were considered to add context and depth to the findings, in keeping with a scoping review [12, 14]. As potential barriers for RLD may be linked to implementation and resistance to change [17], rather than feasibility, it was deemed appropriate to include qualitative studies.

The study population was based on the eligibility criteria (Table 2). Therefore, radiographers with advanced training in projection radiography reporting or image interpretation for the appendicular skeleton were included. The concept for inclusion specifically pertained to RLD for minor MSK injuries of the extremities and the context was discharge or treatment referral from emergency or urgent care. As radiographer role extension in the UK was well established following the introduction of the NHS plan in 2000 [4], only UK NHS studies, post 2000, written in English, were included.

Identifying relevant studies

On the 26th April 2019 CINAHL, Embase and MEDLINE databases were searched. The key words identified in Table 3 and relevant Medical Subject Headings (MESH) terms were combined using Boolean operators. The Radiography journal was hand searched by screening article titles in content pages, for articles pre-dating the March 2017 inclusion on Medline, back to January 2000. The journal was not available online before this date.

Table 3 Summary of key words for population, concept and context
Full size table

Grey literature not available through the traditional databases were also searched [12, 14, 18, 19]. This included sources identified by Public Health England [20], Imaging and Therapy in Practice magazine and the University of Exeter repository (ORE). Keywords used for searching were ‘radiographer-led discharge’ and ‘discharge by radiographer’.

A further search of the included articles reference lists and forward citation chasing was conducted [12, 21]. Scopus medical database and Google Scholar were used for the forward citation chasing [22]. With a limited time scale for the review, a time deadline of the Scopus and Google Scholar searching of 9th June 2019 was set, after which no more new studies were included [12].

Study selection

Once the search was completed, the citations were uploaded to Endnote software v.X8 (Thomson Reuters, New York, NY, USA) [12] and duplicates removed. A random sample of titles and abstracts were cross referenced by two reviewers [12], to assess inter-assessor reliability. This was a binary check that eligibility criteria had been correctly interpreted, using a percentage agreement check of 10% of the total studies returned from searches.

Once included articles were agreed, full texts were obtained. Initially, three randomly selected articles were independently reviewed by two reviewers, using the full text screening form, based on the inclusion criteria. The reviewers then met to confirm appropriateness of the form [12, 14]. The remaining full text articles were independently screened against the full text form by both reviewers, who were blinded to each other’s results [14].

Charting and collating the data

The data were extracted using a descriptive-analytical approach and charted under the headings identified by Arksey and O’Malley [12]. The study characteristics were tabulated by aims, design, location, population and intervention, specifically RLD protocol method used. This process allowed emerging themes to be identified [12].  An overview of the characteristics was reported with further synthesis of qualitative and quantitative outcomes included. This thematic charting process identified gaps in the research evidence base, which were reported in the narrative synthesis. Consistent with scoping review methods, study quality was not assessed [12].

Results

4148 studies were identified, reducing to 11 following title and abstract screening. Full text screening left nine studies. RLD was not the main context for two articles (Fig. 1). The two reviewers discussed five articles, for potentially duplicate reporting. It was agreed to include three articles, as the reports differed in context.

Fig. 1
figure 1

Flow of studies in the scoping review with reasons for exclusion

Full size image

Three articles were sourced from electronic databases [3, 23, 24]. Hand searching identified one article [9, 11, 25]. One study was located in the ORE [26]. A storyboard was identified through the PHE search [27]. The author was contacted to request related published articles, but no response was received. One article was provided by the author, an expert in the field, following consultation about RLD at the study outset [13].

Study characteristics

The scoping review process identified a range of study types despite limited study numbers [12, 14]. The studies included an audit [23], two pilot studies [3, 27] a discrete event simulation (DES) modelling study [26], a feasibility study [9], a survey [24] and a mixed methods study [13]. The conference article did not state study type [25] and potentially shared study data with the feasibility study [9]. The modelling and mixed methods articles were possibly based on one study [13, 26]. The mixed methods study was a Health Education England report extending beyond modelling to include interviews, a focus group and training needs analysis (TNA) [13].

There was a geographical spread with two articles based in south west England [13, 26] and two in north east (NE) Scotland [9, 25]. Two studies were based in north east England [23, 24], one in Wales [27] and one in mid-Yorkshire [3]. The review [11] included the NE England [23, 24] and mid-Yorkshire studies [3]. Studies were mostly conducted in general hospital EDs, with the Scottish articles based in community hospital MIUs [9, 25] (Table 4).

Patient numbers in RLD study arms, where stated, ranged from five [27] to 497 [23], with modelling extending to 1303 [26]. The questionnaire uptake was 101 respondents from around 500 contacted [24]. Excluding modelling studies [13, 26], 3 was the maximum number of RLD radiographers per study [3]. Howard identified RLD radiographers driving the process; a relevant concept if radiographer motivation is pertinent to RLD success [9].

Different patient age ranges were included, where stated, with adults included in three studies [3, 9, 23] and paediatrics only in one [27]. Two studies excluded patients under five years old [3, 23] and one study excluded under two year olds [9], whilst Jenkins only included paediatrics [27]. Justifications for paediatric exclusion were difficulty in clinical assessment [3] and potentially more complex symptoms [23].

The primary aim for four studies was reviewing impact of RLD on length of stay (LOS) in the emergency department, either arrival to discharge [3, 25, 26] or x-ray to discharge [23, 27]. The other studies assessed RLD feasibility [9], and radiographer attitudes to RLD [24]. Knapp et al. primarily scoped local reporting radiographer training requirements [13].

Secondary aims varied from impact on ED recall and re-attendance rates [23, 25] to improving patient experience [27] and reviewing use of a DES model as a support tool for using RLD [26] (Table 4).

Table 4 Study characteristics
Full size table

RLD methodology

RLD protocol differed across the studies. The main theme was radiographers discharging patients with normal x-rays, although giving basic treatment advice was also stated in four studies [3, 9, 11, 23, 26]. Three studies required a discharge plan written at initial clinical assessment to be used by the discharging radiographer [3, 9, 23]. Four studies specified extremity only examinations [3, 9, 23, 26], with this detail not stated for the remaining studies.

Two studies modelled RLD pathways with differing variations, based on process mapping within ED [13, 26]. An insightful RLD perspective was also provided by a patient focus group [13]. The survey [24] did not include a standard interpretation for RLD, with the focus being an overview of the RLD concept and radiographer opinion (Table 4).

Quantitative outcomes

The predominant outcome measure was time of patient arrival in ED to discharge for RLD, compared to standard discharge practice (SDC). Six studies reported a significant LOS reduction using RLD [3, 9, 11, 13, 23, 25,26,27] although one study omitted numerical data [9, 25] (Table 4). Henderson et al. [23] reported 17% (21 min) mean RLD LOS reduction, compared with SDC, which also included RLD data (Fig. 2). No standalone SDC data was included; however, with no overlap of confidence intervals (CIs) for RLD results were still statistically significant (Table 5). Snaith [3] reported RLD LOS reduction of 61% (82 min) and after RLD treatment referral of 41% (63 min).

Fig. 2
figure 2

LOS in minutes for RLD compared to standard discharge. Knapp 50% used to align with the other study methods. Knapp 100% modelled continual RLD use [13]

Full size image
Table 5 Study results
Full size table

Jenkins [27] piloted x-ray to discharge LOS, with 72% reduction. Henderson [23] also reviewed x-ray to discharge times for RLD, with 12.9 min (SD 9.81) mean LOS. No SDC data was cited (Table 5).

Varying levels of RLD were modelled, compared to simulated standard discharge (SSD). The results reported here used 50% and 100% RLD, compared to SSD. This was formerly to align with the intervention, comparator results from other studies and latterly a more realistic estimate of RLD uptake (Fig. 2). Both results demonstrated reduced LOS, against SSD [13]. Differing days of the week utilising RLD were also modelled. With 51% of RLD eligible patients presenting at the weekend, impact of RLD on these 48 h gave a reduction in ED overall LOS of 10%. The limiting factor for this was only 55% of all ED patients were imaged [26]. Impact on ED capacity was modelled with around 500 RLD discharges allowing time for roughly 300 additional clinical examinations by ED staff [26].

Secondary outcome measures for recall and re-attendance within 28 days from original attendance showed significant reductions. Henderson et al. [23] identified RLD re-attendance rate was 2.62%, and either not clinically significant or unrelated. SDC re-attendance rates were 7.06% of which 1.75% were clinically significant. The odds ratio of re-attending with the same injury through standard discharge, compared to RLD was significant at 8.36 (95% CI 2.05 to 34.08) [23]. However, given the wide CIs, the study may be underpowered and therefore this may introduce uncertainty into the results [28]. Snaith reported RLD re-attendance rates of 1.75% [3], compared against other study results of 13.1% and 8.6% for junior doctors and nurse practitioners respectively [3]. Howard and Craib saw no patients re-attending and one patient recall, with no change in management with RLD [25] (Table 5). No comparisons were drawn against ED senior clinicians, despite them routinely clinically assessing patients and interpreting images. Their level of expertise was likely to lead to fewer imaging requests so was a relevant comparator for future consideration [13].

Qualitative outcomes

A range of qualitative methods were used [11 13, 24, 27]. The radiographer questionnaire used snowball sampling across hospitals with an estimated 30% response rate [24]. The radiographer respondents were generally positive about RLD, recognising potential benefits. The requirement for salary to reflect the additional training and responsibility was a motivational point [24].

The patient satisfaction survey was 100% positive about RLD [27]. Patients appreciated not having time wasted and A&E staff thought RLD was a good idea and the trial worked well [27]. Knapp et al. used a patient and public involvement (PPI) focus group and ED staff interviews [13]. The PPI group identified positives of reducing waiting times and potential cost effectiveness. The focus group included the benefit of increased patient satisfaction which was also reflected in the patient satisfaction survey responses [27]. The survey was 100% positive about RLD with patients appreciating not having time wasted. ED staff thought RLD was a good idea and the trial worked well [27].

PPI focus group [13] concerns were possible missed diagnosis and increased risk of litigation, the latter also being the main concern in the Lumsden study [24]. Radiographer competency to discharge, patient safeguarding and pathway changes were raised during ED interviews [13]. Appropriate training, competency use and expertise in discharge were discussed in other studies [23, 24].

Discussion

In this scoping review, seven primary studies were identified addressing RLD in either ED or MIU. This limited number of studies may in part be due to a low uptake of research activity in radiography in general [29].

Differing RLD methodologies were described and those investigating LOS evidenced reductions with RLD, compared to SDC [3, 9, 11, 13, 23, 25,26,27]. RLD demonstrated potential to both increase clinical assessment capacity for ED staff [3, 13, 27], and efficiency within the minor MSK injuries patient pathway [11, 23, 26]. This was also observed for remote access general practitioners [25]. RLD was a variable protocol-driven process offering potential of generalisability and widespread implementation [30].

Another theme was reduction in image interpretation errors improving recall and re-attendance rates [3, 23, 25]. This could improve patient outcome [31] and decrease likelihood of litigation [32]; a key concern of radiographers surveyed [24]. Radiographer hot reporting has demonstrated cost effectiveness with significant reductions in interpretive errors, compared to ED clinicians [33]. Therefore, RLD cost effectiveness was also likely, combining hot reporting with improved minor MSK injury pathway efficiency.

The studies identified positive outcomes, albeit mostly with short time frames [3, 27] and small sample sizes which could impact the strength of the results [28]. It was important to understand why RLD was not more widely utilised. The concern of litigation has already been identified [24]. A further consideration was radiography culture, where a less supportive work environment could impede role development [34]. At non-RLD sites, radiographers surveyed preferred commenting on images to RLD; this was the reverse for RLD active sites [24]. This could be further explained through resistance by radiographers to extend their practice [34], or less confidence with an unfamiliar process [24].

The small number of RLD active radiographers, up to three [3] per study was noted. RLD radiographers could be considered champions actively promoting the initiative [30], within a supportive culture [34]. They would have resistance to departmental culture issues through belief in RLD [30]. Generalisability of RLD [35] may therefore be reliant on the presence of champions, rather than a concept accepted by all appropriately qualified radiographers [30].

With low RLD radiographer numbers, inconsistent uptake of RLD could be expected [3, 23]. Integration of RLD would require consistent use of the protocol-driven process [30] requiring more RLD radiographers. This was implemented following one study which extended RLD service to evenings and weekends [23]. Pathways of RLD use on different days of the week were also modelled [26]. With 51% of RLD eligible patients attending ED at the weekend; efficient and potentially cost-effective use of RLD could occur on these days [26].

Further themes emerged around inter-professional working [9, 13, 25] and radiographer training in discharge [3, 13, 23, 24]. Radiographers consistently interpreted images more accurately than they expected to [7]. Therefore, future training emphasis requires focus on discharge [3]. Given the radiographers’ concern over litigation [24], use of protocol-driven pathways and appropriate governance systems [2] could encourage engagement.

ED staff could be motivated to support this competency-based training in discharge, once their increased workload capacity was recognised [3, 13, 27]. This capacity was through a decrease in the number of clinical assessments required with increased use of RLD [26]. In addition, ED clinician engagement in protocol development and implementation should reduce the potential of RLD appropriate patients presenting without a management plan [3, 30].

Strengths and limitations

This is the first scoping review on RLD utilising a comprehensive searching strategy. As such, there is inclusion of both quantitative outcomes and qualitative content allowing contextualisation of the current RLD evidence base.

Ideally there would have been two reviewers at abstract screening stage and reviewing data extraction stages [14]. However, the 10% title and abstract check and full text screening produced full agreement between assessors.

The quality of studies was not assessed [12, 14]. Small sample sizes were identified as limitations [3, 27], with one study having five participants [27]. Larger sample sizes would have increased the power of the study and therefore likelihood of demonstrating true effect of RLD [28]. Henderson [23] included SD and CI in results, which acknowledged variance of LOS, with patients not discharged within the expected four hours [1]. This was omitted by other studies therefore variance of waiting times could impact study results [28].

Potentially there was a further bias with the focus from the radiographer perspective, despite RLD overlapping with ED [25]. Knapp et al. did include interviews with ED staff and PPI focus group, although extending this to ED based studies would address this [13].

Future research

The narrative synthesis evidenced areas where further investigation could be considered. Reduced recall and re-attendance [3, 23] and service streamlining were identified; however financial impact was not explored. Quality-adjusted life-year (QALY) benefits for patients are possible with increased likelihood of receiving the correct treatment at initial presentation [31]. Hot reporting identified £23.40 saving per patient [8], therefore a good rational for extending DES modelling [26] to a cost effectiveness study of RLD across radiology and ED. Savings could offset some of the training costs and salary increase for radiographers, a motivation for engagement with RLD [24]. A previous study identified radiographer reporting as more cost effective than radiologists. However, further work was required for implementation of the pathway [36]. Innovative thinking between radiology and ED would be required to action the cost effectiveness outcome.

The majority of studies focussed on the radiographer role and Henderson et al. [23] recommended a randomised controlled trial (RCT) as further research. Given the variations in RLD, a cluster RCT with process evaluation would be appropriate to aid fidelity of implementation and give context to outcome variations [37, 38]. This process would include other stakeholders’ perspectives, such as ED staff and patients [4]. As small sample sizes have been identified as study limitations, this would ensure use of larger sample sizes and therefore should give more power to the study [37, 38]. Alternatively, action research – problem solving and improving practice whilst actively undertaking the discharge role, would be an alternative research method, encompassing all relevant parties [38, 39].

Emphasis specifically on the discharge element was required, as the innovative element of RLD [2]. This could be achieved through DES modelling [26] of RLD, from the ED perspective, given the capacity for additional clinical assessments already demonstrated [3, 13, 27]. Alternatively, a time and motion study – monitoring and timing the specific RLD activities, would identify inefficient areas or give improvement targets [40] within the discharge process. This research could directly address the requirement to manage the increasing number of ED and MIU patients [3, 10, 26, 25].

Conclusion

A limited evidence base of seven RLD studies of varying sample sizes and heterogeneity identified potential benefits for the patient, radiographer and ED. The minor MSK injuries pathway could be streamlined. Journey time through ED and likelihood of recall or re-attendance could be reduced and ED staff may gain increased clinical assessment capacity. A cost effectiveness study could identify RLD financial savings which could contribute towards radiographer training and salary increase; an incentive for engaging with RLD. Further qualitative work to examine the impact of this interdepartmental initiative may be the final key to implementation. Understanding radiography department culture and considering the perspectives of all involved through process evaluations alongside larger quantitative studies could be sufficient to review RLD feasibility. RLD success is not only reliant on radiographer uptake, but dependant on acceptance by both ED staff and patients themselves.