Overview
We aimed to conduct a rapid, mixed-methods evaluation of prehospital video triage for suspected stroke patients, as implemented in NC London and East Kent. In order to understand the implementation, experience and impact of these services, our EQs were the following:
EQ1: what evidence exists on prehospital video triage for suspected stroke patients, in terms of implementation, usability, safety and outcomes?
EQ2: are the prehospital video triage services piloted in
NC London and East Kent acceptable to their users (i.e. stroke clinicians and ambulance clinicians)?
EQ3: are the services effective in terms of usability and image/sound quality?
EQ4: do the services support the appropriate, safe conveyance and treatment of suspected stroke patients?
EQ5: which factors influence uptake and impact of these services?
EQ6: which aspects of these services should be retained post COVID-19 and which adaptations (if any) are required to support their implementation?
In this chapter, we first present our principal findings organised by these evaluation questions. Second, we discuss the implications of our evaluation, referring to key themes that emerged from our literature review and empirical work. Third, we present the main strengths and weaknesses of the evaluation, including the evaluation design and data used. Finally, we set out our main conclusions and recommendations for future research, formed in agreement with our clinical and patient collaborators.
Principal findings
In this section, we set out our principal findings, where we draw together the key lessons from our analyses and organise them around our EQs.
What evidence exists on prehospital triage services for suspected stroke patients, in terms of implementation, usability, safety and outcomes?
There is a limited, but growing evidence base on remote prehospital video triage for stroke, which sits alongside more established and substantial literatures covering other aspects of telemedicine for stroke (e.g. systems permitting hospital-to-hospital communication and mobile stroke ambulance units). Much of this evidence was based on pilot or feasibility research, using both simulated and ‘real-world’ settings.
In terms of usability, research suggested that stable network coverage and clear audio-visual signals were important to successful patient assessment. Communication between ambulance and stroke clinicians was also important in ensuring that stroke clinicians could access appropriate patient information.
Training of both ambulance and stroke clinicians was identified as an important facilitator of effective prehospital video triage, for example using simulations to enable a clear understanding of new protocols and effective use of communications technology.
Importantly, both reviews and primary studies on outcomes of prehospital video triage tended to focus on delivery of stroke clinical interventions, for example time from arrival at hospital to brain scan or thrombolysis. They presented little evidence on the impact of prehospital video triage on such key factors as appropriate patient destination, patient safety, and cost-effectiveness.
Were the prehospital video triage services piloted in North Central London and East Kent acceptable to their users (stroke clinicians and ambulance clinicians)?
Our qualitative analysis suggested that in
NC London and East Kent, some ambulance clinicians reported ongoing concerns about whether the potential benefits of accessing specialist secondary care stroke expertise might be outweighed by a delay in patient conveyance. However, most ambulance and stroke clinicians interviewed were supportive and accepting of the prehospital video triage services; for example, both groups cited perceived improvements in appropriate patient conveyance and potential reductions in pressure on services (e.g. unnecessary ambulance journeys and fewer ‘stroke mimics’ to manage in
HASUs).
Ambulance clinicians cited increased confidence and reassurance about their decisions on patient conveyance, and the view that they were learning more about stroke through their communications with stroke clinicians.
Stroke clinicians noted that the service did not involve a significant change in practice, beyond having the opportunity to conduct assessments earlier in the pathway and gain advance knowledge of patients. However, they also described a barrier to acceptability in the common requirement to conduct triage assessments alongside their other professional duties, placing pressure on clinicians and potentially limiting the quality of communication.
Our survey of over 200 ambulance clinicians in
NC London and East Kent confirmed that a substantial majority of respondents (86%) found the prehospital video triage services an improvement on ‘business as usual’ and a similarly high proportion (88%) wanted the new services to continue to operate. However, these positive views were significantly stronger among
NC London ambulance clinicians.
Were the services effective in terms of usability and image/sound quality?
In our interviews with ambulance and stroke clinicians, image and sound quality were seen as sufficient to conduct prehospital video triage assessments. Ambulance clinicians noted that connections could be disrupted, as both built-up areas and highly open spaces had potential for limited Wi-Fi coverage. Under such circumstances, ambulance clinicians reverted to conveyance protocols that operated before the introduction of prehospital video triage. Issues with connectivity and ambient noise in hospital settings were addressed through connection to hospital Wi-Fi and by giving stroke clinicians audio headsets. Our survey confirmed that ambulance clinicians (77%) agreed that the prehospital video triage services were usable in terms of audio-visual quality.
Ambulance and stroke clinicians reported that using the FaceTime communication platform was straightforward to use, although some ambulance clinicians suggested that training and ‘refresher’ courses could be beneficial.
NC London’s approach to training was more active, with both face-to-face training and distribution of video information, whereas in East Kent protocols were distributed via e-mail and an online portal. In line with this, our survey found that a majority of respondents (89%) saw the services as easy to use. However, a higher proportion of
NC London respondents (94%) rated the service as usable than in East Kent (78%): this may reflect the more active approach to training taken in
NC London, where 91% of ambulance clinicians reported having received sufficient training, in contrast to East Kent, where 42% of ambulance clinicians reported having received sufficient training.
Did the services support the appropriate, safe conveyance and treatment of suspected stroke patients?
Our analysis of ambulance journey data showed that only a small percentage of stroke patients did not reach hospital within 60 minutes of leaving the scene of stroke. Publicly reported national stroke audit data on times from symptom onset to stroke patients’ arrival at hospital and
HASUs found that prehospital video triage did not result in increased median stroke patient conveyance times. This suggests that, despite the additional time spent on prehospital video triage consultations, prehospital video triage can be delivered while still supporting timely patient conveyance to stroke services.
Our analysis of national stroke audit data found several significant increases and no significant reductions in delivery of stroke clinical interventions in
NC London and East Kent following introduction of prehospital video triage (i.e. relative to changes observed in
RoE over the same time period). However, it is possible that other factors, such as relocation of the hospitals’ hyper-acute stroke services in both areas, were significant contributors to these improvements.
Our qualitative research suggested that leaders of the triage services were conscious of risks to patient safety and put processes in place to monitor and manage any incidents. Observations of meetings where these data were discussed suggested that there were few reported safety incidents, each of which was explored to identify potential improvements. Our interviews suggested that there was a broader perception among ambulance and stroke clinicians that these services were providing safe care.
Our survey of ambulance clinicians confirmed that prehospital video triage influenced patient destination at least once for most (86%) respondents, while 82% of respondents had no concerns about the safety of the services. However, there was again a significant difference between responses in
NC London, where 91% had no concerns about safety, and East Kent, where 62% had no concerns.
Which factors influenced uptake and impact of these services?
Our qualitative research suggested that several factors helped enable the rapid development, implementation and uptake of prehospital video triage. These factors related to (1) the local and wider contexts into which prehospital video triage was introduced; (2) the interventions themselves; and (3) how they were put into action.
Two national/international contextual factors helped drive these changes. First, there existed a longstanding challenge related to the limited specificity of screening instruments such as
FAST, which led to the conveyance of a substantial proportion of non-stroke patients to stroke units. Second, the COVID-19 pandemic, by adding a significant patient safety risk to the issue of inappropriate patient conveyance, acted as a ‘burning platform’ for change.
These drivers interacted with more local contextual factors. For example, they encouraged adoption of governance processes that were more facilitative of change at system and service levels. The pandemic also raised issues that played on values important to the ambulance and stroke clinicians who were to deliver prehospital video triage, specifically the desire to provide high-quality, safe care to stroke and non-stroke patients.
In terms of the interventions themselves, the prehospital video triage services were attractive to many ambulance and stroke clinicians. They saw the process and interface as straightforward to use. Furthermore, reflecting their professional values (cited above), they indicated that it offered advantages over ‘business as usual’, in terms of getting the patient to the most appropriate service for the best possible care.
Collaborative leadership was key to implementation: ambulance and stroke clinical leads worked in multiple directions to develop the new services. Within their organisations, clinical leads engaged with senior management to gain local endorsement for the services and worked actively with frontline clinicians (e.g. to encourage uptake, provide training and monitor progress of services). They also reached beyond their local organisations; for example, ambulance and stroke clinical leads worked together across organisational boundaries to develop services that would work for ambulance and stroke clinicians alike. They also engaged with wider system governance to obtain support for these changes and aid ongoing preparations for wider roll-out of the services, if successful.
Which aspects of these services should be retained post COVID-19 and which adaptations (if any) are required to support their implementation?
The ambulance and stroke clinicians we interviewed were emphatic that the new prehospital systems represented should be retained. Many suggested that they should be implemented more widely, both in other parts of the country and other health-care specialties. This support for continuation was reflected in our ambulance clinician survey, where 88% of respondents indicated that they would like to see the service continue, and 67% suggested that similar services should be considered for other clinical settings.
Broadly, interviewees suggested that relatively few changes were required to the services. A potential risk to sustainability of these services likely to require adaptation was ensuring sufficient capacity among stroke clinicians to deliver the assessments. The current arrangement, where stroke clinicians conduct assessments alongside their other duties, was felt to be disruptive and placed undue pressure on clinicians (which may have important implications for sustainability of prehospital video triage). Our findings on training suggest that more active approaches are preferred by staff: such approaches, if extended, may have potential to encourage increased collaboration between ambulance and stroke clinicians.
Implications
Our literature review confirmed the findings from other reviews that the evidence base on prehospital video triage for suspected stroke patients is small but growing.15,16 Reflecting a general appreciation of the increasing stability and power of mobile technology, we found recent cases (i.e. published in 2021) of localised activity (in the form of pilots and feasibility studies) in different parts of the world, but little from UK settings. We identified important gaps in relation to the impact of prehospital video triage on appropriate patient destination and patient outcomes, including patient experience, patient safety and cost-effectiveness; we address several of these gaps in our evaluation.
In terms of usability, our qualitative findings confirmed the importance of having a stable Wi–Fi network and clear audio-visual signals, as reported in previous research and reviews.16,66,69 In addition, our qualitative analysis identified the value of back-up processes for situations where clear audio-visual communication is not possible. In line with previous research, we found that training was seen by clinicians as essential to their becoming confident users of the new service.59,61,65 However, in addition, our qualitative analysis suggested the value of ‘refresher’ courses, given ambulance clinicians’ view that they treat suspected stroke patients relatively infrequently. Clinicians also suggested the potential for joint training events that bring together ambulance and stroke clinicians to enable a stronger sense of shared culture across these professions.
Our qualitative and survey findings on acceptability suggested that the stroke and ambulance clinicians using prehospital video triage found it acceptable, with the overwhelming majority wanting the services to continue or even expand. This was in line with previous research, which has suggested high user acceptability of prehospital video triage.16,39,48,65,67 In addition, we were able to identify factors contributing to acceptability. These included increasing ambulance clinicians’ confidence in decisions about patient destination and their understanding of stroke as a condition. An important obstacle to acceptability – and by extension sustainability – of prehospital video triage services was that many stroke clinicians were required to conduct these assessments alongside their standard duties. This was felt to place significant pressure on stroke clinicians.
Our real-world evaluation confirmed findings from recent simulation studies,66,69 which have indicated the importance of clinician communication. However, our qualitative analysis also extended this evidence by providing insights on the distractions that might result from the settings in which ambulance and stroke clinicians conduct the assessment (e.g. ambient noise and competing tasks) and some simple solutions applied (e.g. providing stroke clinicians with communication headsets).
Our systematic review found little published evidence on the impact of prehospital video triage for stroke on patient safety.16,66,69 Our analysis of ambulance journey times suggested that an overwhelming majority of patients who underwent prehospital video triage assessment were conveyed to a HASU within recommended journey time thresholds. Publicly available national stroke audit data on times from symptom onset to stroke patients’ arrival at hospital and a HASU indicated that prehospital video triage did not result in increased median stroke patient conveyance times. This suggests that, despite additional time potentially spent on scene, prehospital video triage can be delivered while still supporting timely patient conveyance to hospital and stroke services. In addition, our qualitative analysis captured examples of how clinical leaders monitored and managed potential patient safety issues (e.g. reviewing potentially inappropriate patient conveyance in terms of both contributing factors and outcomes).
We found no evidence of reductions in delivery of key stroke clinical interventions, and several examples of significant increases in delivery of interventions (although these may be attributed to other service changes that took place concurrently). These patterns were broadly in line with previous research indicating potential to improve delivery of key interventions, such as time to brain scan and time to thrombolysis.78,80,81
Strengths and limitations
This rapid evaluation represented an important opportunity to develop timely lessons on a potentially important change in organising prehospital care for suspected stroke patients, implemented during a highly innovative period for the NHS in England. Our evaluation had several strengths:
We were able to study ‘real-world’ implementation of two prehospital video triage services. These were implemented in parallel, but differed in several important ways, including the ways in which the service was delivered and supported, and the rurality of the area served by the participating ambulance and stroke services; this hopefully allowed our findings to be relevant to a range of contexts.
Our mixed-methods design allowed us to draw on multiple sources of quantitative and qualitative data, letting us address our research questions from several different perspectives.
We were able to access a range of high-quality data in a relatively short time: thanks to our strong collaboration with local clinical leads, we were able to interview many ambulance and stroke clinicians and observe a range of governance meetings, and we also obtained over 200 responses to our ambulance clinician survey. In addition, we received high-quality case-level data on ambulance destinations and were able to analyse publicly available national audit data (to which stroke teams broadly continued to submit data throughout the pandemic).
Using a social science perspective (in particular, Nolte’s conceptual framework for understanding implementation of digital innovations)
17 permitted a rich understanding of the factors contributing to the development, implementation and sustainability of prehospital triage.
There were also several limitations:
Although the services we studied contrasted in some important ways, there were some contextual issues that we could not explore. For example, the services we studied were both based in the south-east of England (because none had been implemented elsewhere at that time), and local stroke services had been reorganised so that the local
HASU was not co-located with an
ED (a separation that may have increased pressure to use the triage service). Similarly, the services studied were introduced during an unprecedented period of change and adaptation in the NHS in England prompted by the COVID-19 pandemic. We were also unable to study alternative prehospital triage systems, for example using telephone discussion of cases. This may limit the extent to which lessons on the implementation and impact of the studied services might be translated to other contexts.
We were unable to interview several stakeholder groups for this evaluation. Key among these groups were patients and carers. This project was highly time-sensitive and obtaining
HRA ethics approval would have caused a substantial delay to commencing work: we therefore designed this project as a service evaluation so that it could commence in a timely fashion. However, this meant that we could not include interviews with patients and carers in our evaluation design and we recognise this as an important limitation to our work. In addition, we interviewed only ambulance clinicians and stroke consultant physicians, and were able to interview only two stroke consultant physicians (including the service lead) in East Kent. This limited the extent to which we could draw conclusions about clinician views of these services. To address some of these gaps, we shared interim findings with patient representatives and members of stroke teams and ambulance services.
Our ambulance conveyance data covered only the areas where prehospital video triage had been introduced and only the time period following the introduction of the triage services; therefore, our analyses had no historical or regional comparators, which limited the nature of analyses we could conduct.
Again as a result of project timelines, we were unable to request national stroke audit data at patient level, meaning we were limited in the types of analysis we could conduct (for instance, we could not apply patient-level risk-adjustment or develop matched controls for our studied areas, and we could not examine the effects of shifting from consultant-led to registrar-led prehospital video triage services).
The national stroke audit data did not cover several potential patient safety issues, such as appropriateness of patient destination (e.g. patients initially identified as non-stroke who then required inter-hospital transfer to a
HASU, and non-stroke patients who were conveyed to a
HASU unnecessarily). However, we observed meetings where these data were discussed, which indicated that such incidents were rare and analysed actively.
Our ambulance clinician survey recruited a convenience sample and although we obtained over 200 responses, these numbers were too small to permit further disaggregation of responses, for example by frequency of use of prehospital video triage.
Conclusions
Our evaluation sought to contribute lessons about prehospital video triage for suspected stroke patients, particularly in relation to implementation, acceptability and usability, and impact on safety.
Implementation
We found that prehospital video triage can be developed and implemented rapidly. By drawing on a relevant theory of implementation and sustainability of innovations, we were able to establish that influential factors included context, implementation approaches and the characteristics of the prehospital video triage services themselves. These factors were interrelated; for example, the COVID-19 pandemic acted as a ‘burning platform’, encouraging more facilitative governance processes and local professional and organisational receptivity to new ways of working. The wider service reorganisations – whereby local stroke units were no longer co-located with ED services – may have encouraged staff to engage more with prehospital video triage. Ambulance clinicians saw training as an important means of becoming confident users of the new service; more active approaches to training are more likely to be viewed positively by clinicians. Collaborative leadership was key to implementation: ambulance and stroke clinical leads engaged with local senior management, frontline clinicians and beyond their local organisations to develop services that would work for ambulance and stroke clinicians alike and to gain ongoing support across the system.
Acceptability and usability
Ambulance and stroke clinicians overall found prehospital video triage acceptable and usable. The technology was seen as straightforward to use and generally reliable. A potentially important factor was the nature of training offered in the two areas, with more active approaches (as employed in NC London) preferred by ambulance clinicians. Stroke clinicians reported concerns about delivering prehospital video triage alongside their other duties, suggesting that addressing this issue would be important to ensuring sustainability of services.
Impact on safety and quality
Almost all stroke patients’ ambulance journeys to HASUs remained within recommended conveyance time thresholds. Publicly reported national stroke audit data on times from symptom onset to stroke patients’ arrival at hospital and a HASU suggested that, despite additional time potentially spent on scene, prehospital video triage can be delivered while still supporting timely patient conveyance to hospital and stroke services. In terms of stroke care delivery, we found several significant increases in delivery of key clinical interventions following the introduction of prehospital video triage (above and beyond what was seen elsewhere in England), although other concurrent changes to service organisation may have played influential roles. Our qualitative data – in terms of both interviews and observations of meetings where safety issues were analysed – suggested that safety was a key priority of the clinicians delivering these services; governance processes gave assurance that the services were indeed delivering safe care as well as wider service and system benefits.
Future research agenda
Although we believe that our evaluation has made several important contributions to the understanding of prehospital video triage for suspected stroke patients, there were several important issues that we could not address. Taken alongside the gaps in evidence identified in our literature reviews, we propose that the following issues should be prioritised in future research:
Qualitative research on patient and carer experience of prehospital video triage is an important gap in current understanding. There would be value in analysing the perspectives of stroke and non-stroke patients and their carers (including people from different ethnic, cultural and socioeconomic backgrounds, and from remote, rural areas), and understanding factors that might influence how people experience prehospital video triage (e.g. immediate symptoms such as dysphagia or loss of ability to communicate verbally). This would provide important learning on how these services work and how they might be improved, for example on how to be accessible to all communities who might require these services.
Qualitative research on the views and
experiences of the wider health-care system would provide a clearer understanding of implementation, delivery and sustainability of these services. Such stakeholders would include clinicians and managers (at varied levels of seniority) in ambulance, stroke and other acute teams, senior organisational managers, commissioners, patient representative groups, and wider system governance (e.g. NHS England/Improvement and
ISDNs).
Quantitative research using national data sets at patient-level would permit more detailed analyses of the impact on prehospital and acute care delivery, conveyance, and patient outcomes for both stroke and non-stroke patients. In particular, such data would permit longitudinal designs using historical (i.e. pre-/post-implementation) and regional comparators, with patient-level risk-adjustment.
Research on cost-effectiveness remains an important gap in knowledge about these services. Quantitative research of the kind described above, integrated with qualitative data on implementation activities, would permit large scale cost-effectiveness evaluation of these services.
Mixed-method research on the issues outlined above may be conducted to address
sustainability of established prehospital video triage services and roll-out
elsewhere. Research on sustainability would permit understanding of how established services develop over time, in accommodating contextual changes and other issues identified here (e.g. stroke clinicians delivering prehospital video triage alongside other duties). Research on roll-out (for instance, to services that have a co-located
ED) would illustrate how prehospital video triage might be adapted to different contexts. Such research would thus increase the likelihood of lessons being engaged with by a wide range of patient, professional and managerial stakeholders based in different settings.
To address many of these issues, the authors recently commenced a new research project, funded by the NIHR
HSDR programme. PHOTONIC (PreHOspital Triage for potential stroke patients: lessONs from systems Implemented in response to COVID19) will run from September 2021 to August 2023 inclusive, with the aim of understanding the implementation, experience and impact of prehospital video triage for suspected stroke. It will employ a mixed-methods approach, using qualitative methods to study implementation and experience from a wide range of stakeholders, including (stroke and non-stroke) patients and carers; it will also analyse patient-level data from national and local data sets to understand the impact of prehospital video triage on patient conveyance, care delivery, patient outcomes and cost-effectiveness, relative to national and regional comparators. For further information, please see https://fundingawards.nihr.ac.uk/award/NIHR133779.
