U.S. flag

An official website of the United States government

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Ramsay AIG, Ledger J, Tomini SM, et al. Prehospital video triage of potential stroke patients in North Central London and East Kent: rapid mixed-methods service evaluation. Southampton (UK): National Institute for Health and Care Research; 2022 Sep. (Health and Social Care Delivery Research, No. 10.26.)

Cover of Prehospital video triage of potential stroke patients in North Central London and East Kent: rapid mixed-methods service evaluation

Prehospital video triage of potential stroke patients in North Central London and East Kent: rapid mixed-methods service evaluation.

Show details

Chapter 3Literature review

Overview

What was already known?

  • There is a well-established evidence base on telestroke networks (which link stroke centres to rural or community hospitals using telemedicine) and mobile stroke units (MSUs) (highly equipped vehicles with CT scanners, laboratory equipment and telemedicine equipment, as well as stroke staff).
  • Less is known about prehospital ‘mobile’ stroke telemedicine, which connects emergency clinicians/paramedics to hospital-based stroke clinicians.

What this chapter adds

  • We conducted a systematic review into ‘mobile telestroke’ or ‘in-ambulance telemedicine’, a process that enables communication between ambulance crews and hospital-based doctors in cases of suspected stroke.
  • Mobile telemedicine systems in stroke care have been piloted and found to be technically feasible. These systems enable two-way audio and video communication between emergency personnel and stroke clinicians, provided that there is good network coverage [e.g. fourth-generation broadband cellular network technology (4G)]. Indeed, reliable two-way audio and video communication is deemed especially important for remote clinical assessment of stroke.
  • Several factors support successful implementation. These include staff training, compatibility with existing systems, clear communication, reliable network coverage, co-design with staff, high-performing microphones and a stable audio connection (especially to avoid clinicians having to repeat information or misunderstandings).
  • Barriers to implementation include background noise, poor usability and weak or inconsistent network coverage.
  • Few process evaluations or RCTs have been conducted on in-ambulance telemedicine for stroke compared with studies of telestroke networks and MSUs, especially with regard to the impact on destinations (including patients remaining at home in the case of mimics) and treatment times.
  • Little is known about cost-effectiveness, patient privacy and patient experience in mobile telestroke in prehospital settings.

Background

This chapter presents a two-phase review of published literature about digital and communication systems that enable stroke specialists to assess potential stroke patients remotely while patients are attended to by emergency services personnel. The focus was on the prehospital pathway, prior to patient admission to a general hospital or specialist stroke centre.

‘Telestroke’ has been defined as ‘the process by which electronic, visual and audio communications (including the telephone) are used to provide diagnostic and consultation support to practitioners at distant sites, assist in or directly deliver medical care to patients at distant sites, and enhance the skills and knowledge of distant medical care providers’.38 By comparison, MSUs are highly equipped ambulances, which usually incorporate a ‘CT scanner, point-of-care laboratory and a device for teleconsultation with the hospital neuroradiologist’,39 along with highly trained staff. MSUs therefore represent a high-tech (and often high-cost) solution, one deemed particularly useful where access to hyper-acute services is limited. However, although much research has focused on MSUs or hospital-based telestroke networks that leverage telemedicine to provide remote neurological cover in rural areas, there is growing recognition of innovations within the prehospital care pathway and ambulances that provide an alternative to MSUs and may improve time to treatment over standard care. The literature, therefore, suggests that prehospital telemedicine may be one way to reduce travel times as it can (1) connect emergency medical services (EMS) clinicians to hospital-based stroke teams and (2) alert hospitals to an incoming patient and direct patients to the most appropriate point of care (e.g. a specialist stroke team). ‘Prehospital telestroke’/‘mobile telemedicine’ is therefore identified as a promising, novel development to support clinical examination in the field by EMS clinicians, with remote support provided from hospital-based staff and stroke experts.16,39–41 This topic was the focus of a two-phase literature review, which we outline below. Phase 1 of the review was used to inform the design of research materials in our rapid evaluation.

Methods

We sought to understand the nature of the published evidence base on digital interventions used in the triage of potential stroke patients. Owing to a rapid evaluation timeframe, the literature search was split into two distinct phases: one conducted early in the project and one that could report towards the end once we had empirical findings.

Phase 1: umbrella ‘review of reviews’ – a rapid, exploratory search

To guide this rapid evaluation and quickly identify key terms and relevant topics, a rapid search for existing reviews was undertaken from June–July 2020. According to the Joanna Briggs Institute (JBI) Manual for Evidence Synthesis, umbrella review permits consideration of multiple wide-ranging aspects of a given issue.42 We deliberately broadened the search to quickly identify existing systematic, scoping, narrative or rapid reviews on the use of innovative technologies to support remote triage of stroke and suspected stroke patients by ambulance and paramedic crews on site.

The aims of the rapid umbrella review were to:

  • identify relevant reviews on the topic (systematic and other types, e.g. scoping reviews)
  • identify conceptual frameworks or theories used to understand the implementation of digital interventions in this context
  • identify gaps in research or evaluation knowledge
  • determine how our rapid evaluation and future research might address these gaps (and avoid duplication).

The following headings were used to guide the search:

  • population – patients suspected of stroke, ‘stroke mimic’ or TIA
  • phenomenon of interest – paramedic and ambulance crews (i.e. EMS clinicians) using technologies to triage and treat stroke (or suspected stroke) patients at the prehospital stage
  • context – ambulance clinicians triaging patients in ambulances or in patients’ homes or care settings (i.e. non-hospital or hyper-acute clinical settings)
  • outcomes – for example efficiency, safety and outcomes compared with usual practice and stroke management pathways.

Inclusion criteria

Using a select number of key word strings, the following databases were accessed and searched from 1 to 4 July 2020: Cochrane Library, ACM Digital Library, Web of Science Core Collection, Epistemonikos and PROSPERO. Searches were by title/abstract and topic. The search was limited to outputs published in English between 2010 and 2020 (Table 2). One member of the team (JL) executed the search. The following parameters were used to identify existing reviews:

TABLE 2

TABLE 2

Rapid umbrella review (phase 1): search terms

  • eligible for inclusion – systematic/rapid/scoping/literature reviews (including published protocols) and evidence syntheses on the topic
  • not eligible for inclusion – pilot and feasibility studies, primary research studies, reviews on stroke rehabilitation, reviews of patient behaviour or person-generated health data, diagnostic test reviews (e.g. FAST) and reviews of thrombolysis (without reference to telestroke or ambulance/paramedics).

As this was a rapid ‘review of reviews’, we excluded empirical studies that could later be identified in phase 2, as well as reviews on less relevant topics such as stroke rehabilitation and patient behaviour change to prevent stroke. Key terms to drive the search were ‘stroke’, ‘ambulance’, ‘paramedic’, ‘triage’, ‘digital’, ‘telestroke’ and ‘review’.

Sharing lessons from umbrella review

The findings from this phase were summarised in a ‘review of reviews’ paper that was circulated to the wider team and clinical collaborators for comment. Feedback was provided through team communications and meetings with the evaluation team and clinical collaborators on the study. The findings were used to inform the design of research materials (e.g. topic guides) and the design of the literature search in phase 2.

Phase 2: systematic review

In addition to placing this evaluation and its conclusions in the context of a wider literature, it was decided to conduct a systematic review for four reasons. First, the most recent and relevant review identified in phase 1, Lumley et al.,16 focused on a variety of prehospital interventions in the stroke pathway (e.g. biomarkers), whereas we were interested primarily in a very specific intervention covered in that scoping review: ‘mobile telemedicine’. This meant that we needed to focus specifically on technologies that enabled communication between ambulance clinicians and stroke clinicians and centres. Second, following discussion of the phase 1 search results and new guiding questions with an information specialist at UCL, the published search strategy used in Lumley et al.16 was discussed and it was noted that there were some medical subject index headings (MeSH) and key words that had not been included but which would be important in the context of this particular evaluation (e.g. we added terms such as ‘bidirectional communication’, ‘iPad’, ‘clinical informatics’ and ‘HASU’). In addition, Lumley et al.16 had framed part of their review on process outcomes (e.g. ‘detect’ and ‘diagnosis’) and the information specialist advised that this may have missed some literature of relevance to this review. We therefore wished to employ a fine-grained search strategy focused on mobile telemedicine, the prehospital emergency context and communication between health professionals, requiring a modified design. We also wanted to identify any social science or human–computer interaction literature associated with this topic to provide an interdisciplinary focus, requiring new terms to be added (see Report Supplementary Material 3 for the detailed MEDLINE search we developed). Third, the phase 1 search revealed a relatively large literature on MSUs. Because we were interested in evidence about potentially lower-cost digital technologies that support two-way communication in the field, we required a different strategy from that for other reviews, and to go beyond the literature on MSUs. Finally, as the two pilots evaluated here were implemented in the context of COVID-19, it was important to capture any more recent studies that had been implemented and written up during the last year (2020–21), although it was noted that, owing to publication time delays, more recent empirical studies and reports might be small in number.

The final systematic review search strategy involved close working between a researcher (JL) and a UCL information specialist (Debora Marletta, UCL librarian) and extensive trialling of key terms and Boolean operators across multiple databases. The review was guided by PRISMA 2020 recommendations24 to ensure that the approach taken was both transparent and explicit, and a final protocol was published on PROSPERO following team discussion and review.43 Collaboration between the researcher and the information specialist was essential to maximise the sensitivity of the search and to sense-check returns, helping ensure that the final records included insights about the implementation of digital and communication technologies within prehospital acute stroke pathways. The information specialist checked all final search strategies that had been refined until consensus was reached. In addition, a ‘test list’ of potentially eligible papers was used drawing on phase 1 in order that JL could confirm that the final search strategy was picking up relevant papers.

We developed a specific title and research questions for the systematic review in autumn 2020, which were discussed and agreed on as a team. The phase 2 review aimed to identify implementation factors associated with digital and communication technologies that could impact on health-care service quality and health outcomes, and any conceptual frameworks and social science perspectives that could help explain these relationships. The search strategy was framed around the participants, intervention, and health context and setting, but not limited by outcomes in order to capture the wider literature. Instead, we were interested in a variety of outcomes of interest, such as implementation factors, safety and security of digital and communication systems, usability, user experience, acceptability (staff and patients), cost-effectiveness, overall clinical effectiveness, and ambulance destination outcomes (e.g. faster treatment times, but also issues of usability, staff experience, enhanced communication and safety). We also wanted to confirm any evaluation gaps, such as whether or not economic analyses had been conducted.

The final review title was ‘The implementation of digital interventions and communication technologies in emergency care pathways to support the remote assessment and triage of patients suspected of stroke: a systematic review’.

The review questions were as follows:

  • Which human, technological, and usability factors are associated with the implementation of digital and communication technologies in ambulance settings (e.g. mobile telemedicine) that enable two-way communication between paramedics and hospital physicians in the care of stroke and potential stroke patients?
  • Which system and local contextual factors are important for the implementation of these kind of technologies in the care of stroke and potential stroke patients?
  • What service-level outcomes (e.g. clinical, financial and resource impacts) are associated with the use of these kind of technologies in the care of stroke and potential stroke patients?
  • What evidence exists about the safety and security of these kind of technologies and systems when used at the prehospital admission stage? Relatedly, are any adverse effects reported in the literature?
  • What (if any) conceptual frameworks have been used to understand the implementation of these kind of technologies and systems?

Database search strategies and specific operators were developed and piloted by a researcher and information specialist during the first part of 2021. Six databases were used for the systematic review:

  1. the Cochrane Systematic Reviews Database
  2. The University of York Centre for Reviews and Dissemination [covering the Database of Abstracts of Reviews of Effects (DARE), the NHS Economic Evaluation Database (NHS EED) and Health Technology Assessment (HTA) database]
  3. MEDLINE (via Ovid)
  5. EMBASE (1980–present)
  6. Web of Science Core Collection.

The final searches were run across the databases from July to August 2021. Searches were saved so that the main reviewer received automatic alerts about new records. Report Supplementary Material 3 provides details of the search strategies executed across a selection of these databases.

Eligibility criteria

The review was limited to including articles published in English from 2010 onwards. The population eligible for inclusion comprised patients of any age suspected of having acute stroke, ‘stroke mimic’ or TIA and assessed remotely by paramedics/ambulance clinicians and stroke doctors using digital and communication technologies. The exclusion criteria for the review were as follows:

  • diagnostic instruments used by ambulance clinicians/paramedics only (e.g. FAST+ test) without the input of a hospital-based stroke doctor
  • processes and interventions not facilitated by digital and communication technologies
  • patients treated in MSUs equipped with specialist diagnostic equipment (e.g. scanners)
  • patients treated for stroke, ‘stroke mimic’ or TIA in hospital accident and emergency (A&E), EDs or clinical settings only without prior triage by ambulance or EMS
  • commentaries/editorials, grey literature, conference proceedings or opinion pieces
  • non-peer-reviewed studies.

There was no restriction based on study type. RCTs, feasibility studies, pilots, service evaluations, implementation studies (qualitative), health economic studies, and so on, were all deemed relevant.

Data extraction and screening

A total of 4577 records were downloaded from the databases into EndNote [version 20, Clarivate Analytics (formerly Thomson Reuters), Philadelphia, PA, USA], which was used to identify duplicate entries (Figure 2). The process used for deduplication was recommended by the information specialist.45 Duplicates were automatically identified by EndNote and manually checked by one researcher (JL) using this method. Records were next uploaded to the software Rayyan.ai for screening of titles and abstracts, a package that also facilitates team collaboration. To ensure reliability in the screening of titles and abstracts, a screening tool was developed (see Report Supplementary Material 4) and piloted on the first 200 records, and the results were discussed with a second reviewer (AIGR). This led to a validated final screening protocol for searching all titles and abstracts, which was led by one reviewer (JL). If any papers were identified in which eligibility was unclear, the title and abstract were discussed by two reviewers and an agreement was reached. This process was important for screening a large number of papers that addressed telestroke ‘hub-and-spoke’ configurations, but excluded discussion about emergency care pathways and ambulance/paramedic teams.

FIGURE 2. Systematic review: PRISMA flow diagram.

FIGURE 2

Systematic review: PRISMA flow diagram. Based on PRISMA 2020 – flow diagram for systematic reviews which include searches of databases and registers only). CRD, Cochrane Systematic Reviews Database; WOS, Web of Science Core Collection.

One reviewer (JL) identified the 288 records for review against the inclusion criteria and validated screening tool. This resulted in the identification of around 60 relevant reports for inclusion. A second reviewer (AIGR) was involved in screening this final group, using Rayyan.ai.46 Any reports where it was uncertain whether or not they should be included were discussed until consensus was reached. For example, a number of studies incorporated a prenotification telephone call from emergency clinicians to stroke clinicians or units; these were excluded following discussion because they did not involve the explicitly use of video or two-way communication to support remote triage and diagnosis. Two reviewers (JL and AIGR) approved the final group of 47 included reports.

Data extraction

The full texts were reviewed, and the thematic findings are summarised below (for detailed study characteristics, see Appendix 1, Table 11). We extracted and recorded information in Microsoft Excel about (1) study details (e.g. title, year, publication, authors, study type, the intervention, setting and population); (2) details of the study design and methods (e.g. trial, economic evaluation, feasibility, pilot or observational study); (3) results, including any implementation factors associated with implementation of the intervention; and (4) researcher observations regarding relevance.

Findings

Phase 1 identified 15 relevant reviews within the search parameters (Table 3). The findings are summarised as follows:

TABLE 3

TABLE 3

Review of reviews summary table

  • Types of research –
    • The most recent review on the topic,16 having identified published protocols and studies up until June 2019, was highly relevant to our review.
  • Defining terms –
    • The terms ‘telestroke’ and ‘telemedicine’, when used in reference to prehospital stroke care/emergency diagnosis, identify pilots, primary research, RCTs and protocols on the topic.
    • There are varied service innovations in prehospital stroke care, including diagnostic algorithms for dispatch teams, MSUs and remote deployment of treatment on scene, and new handover protocols and templates.
    • MSUs are highly equipped specialised ambulances, originally piloted in Germany, and form part of a prehospital stroke management response. They are a high-tech solution and potentially useful where access to hyper-acute services is limited. They require specialist teams and training. Stroke specialists may be onboard or accessed via telemedicine facilities within MSUs.
  • Aims/focus of reviews –
    • Reviews focused strongly on the impact of prehospital triage using communication technologies on time taken to reach hospital, time taken to deliver clinical treatment, mortality and clinical outcomes.
    • In contrast, there was little focus on the impact of prehospital triage using communication technologies on the following key outcomes: appropriate patient destination, ambulance and stroke clinicians’ shared decision-making, patient safety, patient satisfaction, cost-effectiveness, and communication between stroke and ambulance clinicians.
    • However, we recognised that such detail may be present in the underpinning primary studies.
  • Quality of reviews –
    • Mixed quality: systematic and systematic scoping reviews identified alongside descriptive reviews based on a limited number of databases.
    • A lack of social science concepts and theory is apparent in the literature. There is more focus on treatment and journey times and outcomes (e.g. there is a lack of insight reported from science and technology studies, although concepts may have been applied in the original studies). One exception by French et al.15 applied normalisation process theory to telestroke.
  • 4G may be more effective for mobile telemedicine in stroke care (e.g. in supporting video communication).
  • The telemedicine and telestroke literatures discuss general barriers to implementation, for example technical issues such as poor transmission or network speeds (especially for earlier systems), low usability, level of compatibility with systems already in use and the costs of equipping ‘high-tech’ ambulances; however, there was a lack of cost information about mobile telemedicine compared with MSUs overall.

Key words that were frequently identified across the papers and therefore informed the design of the phase 2 systematic review, were ‘telemedicine’ OR ‘telestroke’, ‘Stroke’ OR ‘Acute Ischaemic Stroke’, ‘emergency medical services’ (EMS), ‘Mobile Stroke Unit’ (MSU), ‘Prehospital’, ‘Paramedics’, ‘Emergency medical technicians’, ‘ambulances’.

Phase 2: systematic review

The systematic review identified 47 papers that met the inclusion criteria (32 primary studies39,41,59–88 and 15 reviews16,39,40,47,48,53,89–97; for details, see Appendix 1, Table 11). One record was a trial registration;98 this led to the identification of a research paper,80 which was included in place of the trial registry record. The publication date ranges suggested a fairly slow research trajectory, with signs of growth from 2019 onwards. The majority of studies originated in the USA and Germany, with a small number of studies found from across Scandinavia, Asia and mainland Europe. The UK studies were based in Scotland and no studies were based in the NHS in England. Below we provide a summary of the findings. We included other reviews to ensure that no underlying studies were missed, to provide continuity with phase 1 and to confirm any evidence and knowledge gaps.

Insights from available reviews

In a systematic review into prehospital EMS telehealth, Winburn et al.40 identified 68 studies, the majority of which focused on stroke and acute cardiovascular care, suggesting a broader trend in this area. Amadi-Obi et al.89 also found that the literature on prehospital telemedicine largely focused on stroke. Generally, existing reviews described how technological innovations in prehospital stroke management may improve acute care management and workflows (e.g. timely assessment) and potentially lead to faster access to intravenous (i.v.) thrombolysis treatment. This followed the ‘time is brain’ rationale and importance of improvement interventions in the hyper-acute time window (e.g. Aude Bert et al. 2013).47

Interventions to reduce delays to treatment included ‘advanced notification’ by EMS, alerting hospital stroke teams of an incoming patient, as well as the administration of therapies within ambulances. However, within-ambulance treatment was not the primary focus of our review. We found limited information in the literature about the impact of ambulance telemedicine on destinations or pathway determination in stroke care. This included leveraging digital technology to allow patients to remain at home and be referred to an outpatient TIA clinic rather than be taken to hospital following a remote diagnosis.

Previous reviews identified the concept of ‘video examination of stroke patients in ambulances for earlier stroke recognition’,47 and there was a wider interest in ‘adjunctive technology’ at the prehospital stage to support the stratification of patients by ambulance crews.16 For example, a systematic review about ‘advances in TeleStroke’53 identified a number of issues in the prehospital ‘stroke rescue chain’, viewing telemedicine as a potential solution to ‘cut down prehospital times’ and improve prehospital stroke recognition and ‘prenotification’ to hospital stroke teams. However, the review concluded that, although prehospital video triage could enhance stroke identification and facilitate advance knowledge of incoming patients, there remained gaps in understanding about 4G coverage, impact on outcomes and cost-effectiveness.53

Reviews published in 2020–21 confirmed these observations, reporting a small but growing evidence base about mobile telestroke technology based on primary studies, but with few employing a randomised design. In their scoping review, Lumley et al.16 found only 15 studies that reported mobile telemedicine using video and audio technology. Most of these (11 of the 15) had limited information on costs, safety and outcomes. While Lumley et al.16 note the relative maturity of telemedicine for stroke, they report:

little robust evidence of impact on patient outcomes . . . Telestroke may expedite time-to-treatment by attenuating hospital-based assessment, but studies to date have shown little evidence of more efficient patient redirection to stroke-specific centres and no impact on health outcomes for specific population groups.

Lumley et al.16

More recently, Guzik et al.93 hypothesised that there may be additional advantages of mobile telestroke in prehospital assessment, particularly in the context of COVID-19, in terms of reducing the need for ‘multiple re-evaluations’ of the patient and directing them to the most appropriate hospital.

Regarding implementation factors, which were also a focus of our systematic review, Rogers et al.48 provided helpful insights into ‘prehospital telehealth utilization’. They identified a variety of health studies, including six studies specific to stroke. Issues concerned bandwidth and download speeds, with an ethnographic study noteworthy for highlighting usability problems. The authors recommended greater input from patients, doctors, and staff in the design of telemedicine systems to support implementation, noting the potential for telemedicine to support remote triage in emergency care. However, they observed ‘a paucity of published studies describing scientifically valid and reproducible evaluations at various stages of telemedicine implementation in ambulances.’48 Although this general observation was confirmed by our up-to-date systematic review, we also identified additional new primary studies on this topic, some of which reported positive outcomes (see Service outcomes).

Intervention characteristics

Information on the types of interventions found in the primary studies included in this systematic review is provided in Appendix 1, Table 11. There has been a gradual evolution in the approaches and technologies adopted over time, and a range of studies conducted, including feasibility pilots and prospective studies. We saw a shift to studying prehospital ‘mobile telestroke’ or ‘mobile telemedicine’. A term found in earlier literature was ‘online medical control’, defined by Verma et al. as instances where ‘paramedics contact the medical control physician before a Code Stroke triage is assigned’:87 a prehospital stroke protocol was implemented to identify patients eligible for tissue plasminogen activator treatment and to expedite transfer to a specialist stroke centre.

Although this study did not include the use of sophisticated videoconferencing technology, the authors did highlight an opportunity to improve diagnostic accuracy and triage by implementing an online approach as such a system would enable paramedics to contact a doctor remotely and seek advice before a final decision is made to triage a patient to a stroke centre. This approach was contrasted with ‘offline’ medical control, whereby paramedics make decisions independently from a hospital doctor, following guidance and stroke protocols.87

More recent studies demonstrated a shift from pre-alerting stroke teams in hospitals to using videoconferencing technology in an ambulance to support two-way communication, alongside the use of recommended stroke protocols and diagnostic tools. However, it became evident during this review (especially when screening titles and abstracts and 288 full texts) that it was not always straightforward to determine the precise nature of communication between ambulance and EMS personnel and hospital doctors. For example, a receiving clinician could be a neurologist, a teleconsultation physician, or an emergency or ‘EMS’ physician. This appeared to result from differing terminology used in published research, and differing acute stroke service configurations employed internationally. In the USA, first emergency responders may be paramedics or firefighters, and therefore there is some mention of firefighting personnel in the EMS response. We included papers discussing ‘EMS physicians’ as well as stroke doctors to avoid missing any important insights about within-ambulance telemedicine systems and communication between EMS teams to external experts who used stroke diagnostic tools to support remote triage and diagnosis, although this did push the limits of our inclusion criteria.

In terms of technological capabilities, although MSUs are highly equipped vehicles and thus excluded from this review, there was evidence that standard ambulances can be equipped with more advanced digital and communication technologies. High-definition cameras and audio equipment were used to support EMS-hospital communication, accurate triage and real-time sharing of patient data [e.g. heart rate and blood pressure (BP)] via stroke teleconsultations. Prehospital ambulance systems showed signs of greater digital maturity over time, including data integration with hospital information systems, as well as more portable elements with the move to 4G and as new technologies come on board (e.g. iPads, handheld devices and computers). The majority of pilots and feasibility studies identified in this systematic review typically involved building telemedicine or ‘mobile telestroke’ systems in ambulances and testing their functionality (e.g. via simulations) before use in the field and on actual cohorts of patients. Noteworthy studies of this kind were undertaken in Germany and the USA. Our systematic review also found several newer studies, including an important prospective and observational study from Germany that suggested that the evidence base on these systems is indeed evolving over time. Eder et al.81 described the ‘Stroke Angel’ programme, an ‘interdisciplinary project, which aims to improve acute stroke management using mobile technologies (handheld computer) for decision-making, documentation, and communication support between EMS and in-hospital stroke staff’.81 This was an example of a well-developed prenotification and acute stroke management system in the prehospital pathway, integrating stroke diagnostic protocols, a handheld device, time stamping, image capture and integration with the receiving hospital’s own information system. Another example, from China, was Wu et al.’s77 study of ‘Green’, which was a prehospital notification ‘real-time communication’ system for the management of acute stroke. This approach used a smartphone application, which underscores the use of more portable and advanced digital approaches over time that support data integration with hospital records and information systems.

Although we found a small number of recent examples of within-ambulance telemedicine in acute stroke care that are particularly relevant to the rapid evaluation, our review generally highlighted a gap around mobile systems used outside the ambulance by emergency clinicians, specifically the use of videoconferencing to bring the neurologist on site, for example in a patient’s home or at a residential care facility. Moreover, some studies (e.g. Mazya et al.84) focused on more traditional hospital pre-notification systems rather than prehospital mobile telemedicine that enables video and audio data transmission in real-time, either in or outside an ambulance. By contrast, recent research has provided examples of more digitally advanced systems.66,77,81 This review therefore confirmed the relative newness of EMS clinicians using videoconferencing and digital technologies in the field to diagnose and triage stroke patients with the support of a remote stroke doctor, alongside the use of established stroke protocols (e.g. the FAST test).

Technological, human and usability factors

Many studies focused on feasibility analyses and piloting prior to their use in the field, owing to the newness of the intervention and a need to ensure technological stability and usability. Such studies also explored interactive elements, such as communication between health professionals, and task and workflow features.

The first ambulance telemedicine system dates to the early 2000s.62 However, a research group from Germany reported ‘the first study that evaluated prehospital teleconsultation including real-time video transmission from an ambulance in real stroke patients’ in 2010.79 In this prospective study, the researchers tested a within-ambulance telemedicine system (the intervention) on patients and compared it against patients treated in the standard way (the control group). The researchers were interested in testing the feasibility of the emergency telemedicine system and its technical functions.79 They found that the telemedicine system worked well in the majority of cases and that ‘neurological co-evaluation’ was feasible. Video streaming was found to be helpful, although there were some cases of loss of audio or video transfer. This represented one of the earliest studies on this topic involving patients and which concluded that the concept of telemedicine system for stroke care was feasible, while noting that technical issues would require resolution. In another paper, the authors suggested that lack of mobile network coverage was a limiting factor.60

Other pilot and feasibility studies confirmed that video image quality and audio transmission was good enough to support remote neurological assessment, with ambulance telemedicine for stroke care performing well technically in feasibility and usability testing with clinicians, including in rural areas, provided there was reliable network or broadband coverage.62,64,67,75 Although it should be noted that these telemedicine set-ups were typically within an ambulance and reliant on mounted ceiling cameras and audio units, the suggestion was that technical failure may be less common than human error. Chapman Smith et al.,62 for example, found in their study of a telestroke platform that ‘91% of the prehospital mobile evaluations were completed without any major technical failure’.

An important human factor is communication; that is, both verbal and non-verbal signalling within consultations. Joseph et al.66 were interested in staff interactions when using an ambulance telemedicine system and applied cognition theory to understand this phenomenon. The research team looked at communication between nurses, paramedics, and neurologists through simulated teleconsultations and structured analysis. They found ‘significant back and forth verbal interactions between the neurologist and the paramedic and the neurologist and the patient, with the paramedic frequently serving as the intermediator between the other two’.66 The researchers also noted potential risks to teleconsultation, including ‘loud background noise from sirens and traffic and poor audio signals’, all of which can make communication problematic. They therefore suggested telemedicine systems should be designed with consideration for ‘nonverbal team communication’.66

Elsewhere, the same US-based research group (Rogers et al.)69 analysed task structure and flow in a remote ambulance telemedicine system for stroke to provide greater understanding of the usability issues and human factors that might influence implementation, again using simulations. They found that not being able to fully understand or clearly hear the paramedic or neurologist was an issue, leading the researchers to observe that better technical equipment (e.g. microphones, camera) in an ambulance can help reduce instances of miscommunication and ensure that the remote neurologist can see the patient and pick up on non-verbal cues.69

It is also important to note the added value of bidirectional video data transmission. In Germany, for example, emergency physicians have often been dispatched on site. However, owing to a shortage of staff, there has been an interest in remote EMS-physician telemedicine, including for diagnosing cases of suspected stroke and other neurological events.85 Although at the boundaries of inclusion for this review owing to use of ‘EMS physicians’ rather than neurologists, the study from Aachen, Germany, was an interesting case of an evolving emergency telemedicine system used for prehospital stroke clinical assessment because the system was modified for neurological conditions to ensure that video became mandatory. Our review thus suggested that two-way audio and video communication may be especially important for clinical assessment of stroke.65

Staff training

It has already been established in the literature that training of emergency medical services personnel can improve stroke screening.90 Several studies have reported training for emergency staff before going live with telemedicine systems, such as by using simulations, protocols and scripts.59,65,79 Chapman et al.62 have noted that doctors needed to be trained about ‘best positioning of tablet and lighting’ in their study of a mobile telestroke platform (the iTreat study), and Bergrath et al.79 have observed that new equipment and workflow processes required doctors to be trained. In one analysis, for example, the researchers discuss an 8-hour staff training programme designed for a prehospital telemedicine system and undertaken prior to implementation.60 The training was for both paramedics and emergency care physicians to ensure that they were familiar with workflow, technical equipment in the ambulance, communication and miscommunication, and processes for emergencies (e.g. via simulated activity).

Service outcomes

Recent studies included in this review indicated that prehospital ambulance telemedicine, which includes prenotification to a hospital, can reduce transfer and treatment times. Studies reported an array of impacts [e.g. feasibility and door-to-needle (DTN) times] depending on their study design, purpose and the development or maturity stage of the specific system under investigation. We did not limit our review to searching for particular outcomes in order to capture this diversity.

Wu et al.78 examined the impact of the ‘Green’ system on DTN times, reporting that, in 2 years of using the system in Beijing, ‘DTN time was significantly reduced from 50 minutes in 2018 to 42 minutes in 2019’.78 The authors concluded that the Green system was impactful in terms of DTN time and highlighted additional time-saving opportunities, such as patients consenting to treatment earlier at the prehospital stage.

Eder et al.81 reported findings from the Stroke Angel initiative, which involved two study cohorts. Cohort II, which followed ‘workflow modification’ after Cohort I, found an impact on door-to-scan times and DTN times compared with standard care, and a higher rate of thrombolysis. Although the study was not a RCT and the authors acknowledge a number of limitations (e.g. use of the system being at EMS clinician discretion), they concluded that ‘Stroke Angel improves the odds of systemic thrombolytic therapy use as compared with a conventional prenotification workflow protocol’.81

Drenck et al.,80 reported a cross-sectional study of an intervention implemented in Denmark, where emergency staff assessed the patient for symptoms and signs of stroke in the field and communicated with a neurologist in a stroke centre to assess eligibility for scanning and potential thrombolysis. Ambulance on-scene time (OST) was a primary outcome, with the authors finding that ‘the median on scene time was 21 min[utes]’ and that ‘neither relatives nor ambulance trainees present on-scene or in the ambulance were found to affect OST when compared to no relatives or ambulance trainees, respectively’.80

Economic and resource outcomes

Our systematic review identified one cost-effectiveness analysis on this topic. Valenzuela Espinoza et al.41 attempted the ‘first cost effectiveness model for in-ambulance telemedicine’, defining in-ambulance telemedicine as ‘live bidirectional audio-video between a patient and a neurologist in a moving ambulance and the automated transfer of vital parameters’.41 The authors compared standard stroke care with in-ambulance telemedicine, applying a decision tree and Markov model. Several cost advantages were identified at the hospital end (e.g. staff) as well as the fact that ambulances can be straightforwardly equipped with telemedicine systems. Therefore, there was overall added value, even when factoring in staff training as an implementation cost. The authors also predicted lower costs as mobile telemedicine is more widely implemented in ambulances. Although the authors acknowledged that their analysis was not based on RCT data, they concluded that ‘in-ambulance telestroke is highly cost-effective from a health-care perspective, resulting in more QALYs and less costs starting from a realized time gain of 12 minutes’.41

Discussion

Principal findings

Our review confirmed a limited, yet growing evidence base about telemedicine use within ambulances for the diagnosis of stroke. This can be viewed as a growing field of research compared with a traditional focus on MSUs, telestroke networks and other hospital-to-hospital telemedicine systems, which are common in the USA and Europe where rural areas require access to neurological expertise that is otherwise unavailable.

This systematic review provided an important update on a scoping review by Lumley et al.16 identified in phase 1. We conducted a systematic review containing more key terms and medical subject headings to hone in on the emerging topic of within-ambulance mobile telemedicine (Lumley et al.16 included a number of other innovations in their scoping review, such as biomarkers, which was not the focus of our study). Despite not including conference proceedings, we were able to identify new studies as our search was highly sensitive and extended beyond 2019 to 2021. As with Lumley et al.,16 a meta-analysis was not attempted owing to the diversity of studies found. We also included existing literature reviews to get a wider overview of the topic and growth in the field and to ensure that we captured all relevant studies.

Our systematic review found limited data on the impact on outcomes, although these, including data showing the impact on treatment and destination times, are emerging. However, there was generally limited information about other service outcomes of importance to understanding prehospital video triage, such as the impact on decision-making, patient destination and avoidance of unnecessary hospital admissions (thus helping reduce demands on hospital services).

It has already been observed that ‘ambulance-to-hospital teleconsultation may result in markedly reduced delivery times for thrombectomy without delaying intravenous thrombolysis’;84 further research and evaluation is required to support this assessment, although there are early indications of positive impacts on thrombolysis and destination times from a small number of observational and cohort studies. Yet, clearly, more insights are needed about not just transfer and treatment times, but destination end points within the stroke care pathway. With regard to the technical specifications and usability of these systems, there was a gap concerning the influence of different professional roles, such as whether there are advantages of ‘paramedic–neurologist’ communication over ‘paramedic–nurse’ communication. Joseph et al.66 provided an especially helpful study into communication between different professional groups, taking a human factors and task-based approach and highlighting teamwork and a number of practical considerations (e.g. background noise).

There was a lack of insight about the overall safety of these systems (although clinical safety is covered by reference to the use of diagnostic scales and stroke protocols, e.g. FAST and NIHSS). We found one study on overall cost-effectiveness that confirmed what some authors have noted elsewhere: that mobile telemedicine in ambulances is a lower-cost option than introducing MSUs. Notably, there were few RCTs, although a number of studies attempted comparisons with standard acute stroke care and emergency pathways. Finally, more needs to be understood about the impact of different health system models and stroke service configurations on implementation. For example, in some cases designated telestroke centres may field all calls from paramedics/EMS clinicians; in other cases, it may be neurologists based within specific hospitals. The impact of different service models, therefore, requires further investigation.

Observations regarding quality and characteristics of evidence

Both of our reviews identified a large number of pilot and feasibility studies, as opposed to mixed-methods service evaluations (i.e. studies including qualitative, quantitative and economic analyses) or comparative study designs (e.g. of standard pathways versus prehospital triage). We found many literature reviews, of variable quality, indicating ongoing interest in the topic of prehospital triage in stroke pathways and its impact on consultation, journey and treatment times. We also noted a smaller number of prospective and retrospective study designs using cohorts of stroke patients. Yet overall, there was a limited amount of empirical evidence drawn from implementation and prospective studies of emergency care and stroke pathways, especially studies reporting national findings as opposed to results from interventions within specific regions or urban geographies. The novelty of the intervention and approach is likely to explain these findings and the tendency in the literature towards ‘proof of concept’ and pilot studies focused on issues of technical feasibility, user satisfaction, system performance and stability. A common example was simulations conducted by researchers and developers concerned with user acceptance and issues of signal quality and data transmission to test whether prehospital triage technology can be used safely in clinical care to support remote diagnostic assessment. Therefore, evidence on prehospital triage systems for suspected stroke that use low-cost telemedicine technologies – specifically, real-time, audio-visual data transmission outside highly equipped ambulances – remains at a relatively early stage of maturity.

Strengths and weaknesses

Strengths of the review included a transparent and explicit search strategy, specifically following PRISMA for Scoping Reviews (PRISMA-ScR) 2020 guidelines in phase 2 where multiple databases were searched by an experienced researcher with expert information specialist input. Collaboration was used to design and execute a systematic search strategy (i.e. with tailoring to the different databases). We developed and validated a screening tool to operationalise the inclusion and exclusion criteria with a second reviewer sense-checking findings and final inclusion decisions through a process of using software (Rayyan.ai’s collaboration function) and meeting to discuss interpretations. The wider team and co-investigators also provided ongoing advice and feedback into the design and findings.

Owing to the timeframes of this rapid evaluation, no risk-of-bias assessment or detailed critical appraisal was performed, although this will take place prior to publication of the systematic review in a peer-reviewed journal, following PRISMA-ScR 2020 recommendations.24 Given the limited evidence on this topic, the review could be further strengthened by undertaking additional steps that were not feasible within the timeframe of a rapid evaluation (e.g. additional citation searching and checking pre-print servers and recent conference proceedings). In this chapter, we have summarised the main findings with a focus on the issues most relevant to the empirical findings and conclusions of this report, and the most recent empirical evidence found.

Implications

There were several implications from this search of the literature, especially from the findings from the systematic review. First, mobile telemedicine was found to involve some examples of well-equipped ambulances with high-definition cameras and audio equipment, and there are suggestions in the literature that these may mitigate background noise and capture non-verbal cues, which are important in stroke teleconsultations. Given a small but growing evidence base in this area, there is need for further evaluations and clinical trials of these systems that take into consideration the specific equipment used, its functionality, its costs and the implementation factors that shape usage, for example understanding which contexts are optimal for performing remote triage and assessment and which are not, and why. Second, only one cost-effectiveness analysis was identified in the systematic review, suggesting a need for more economic studies, especially given the variations found between health systems internationally and regionally and different system set-up costs. Finally, there was a lack of detailed exploration about the patient experience, with most studies focusing instead on the perspectives of health-care professionals and advantages for health providers.

Overall, we found that less is known about stroke teleconsultations that are on scene (such as in a person’s home), away from a more highly equipped ambulance, and involve emergency clinicians and remote stroke neurologists. In particular, more knowledge is needed about how prehospital telemedicine systems affect stroke destination decisions (including the identification of non-stroke patients or ‘stroke mimics’), and this may help to ensure that patients are directed to the most appropriate point of care or service. This is also necessary given the potential of mobile telemedicine to help with managing high levels of demand on stroke and emergency services, and the possibility of costs savings (e.g. by reducing unnecessary ambulance journeys and diagnostic scans). Very few studies discussed in any detail bypassing certain hospitals as a result of remote triage decision-making between ambulance and stroke clinicians.

More understanding is also required about the contribution of mobile telemedicine over prehospital stroke notification systems that involve a stroke alert to the hospital and data transmission (e.g. vital signs). Future studies may also wish to consider which types of service outcomes are most appropriate for measurement when evaluating pilot services. For example, there may be only marginal gains of using mobile telemedicine for prehospital triage if onset-to-treatment times are already excellent and occur within a well-specified destination pathway. As one paper80 noted, there has been much attention on studying and improving onset-to-treatment times for decades, yet only in recent years has there been a focus on prehospital stroke care, including how much time is spent on scene by emergency clinicians.

Copyright © 2022 Ramsay et al. This work was produced by Ramsay et al. under the terms of a commissioning contract issued by the Secretary of State for Health and Social Care. This is an Open Access publication distributed under the terms of the Creative Commons Attribution CC BY 4.0 licence, which permits unrestricted use, distribution, reproduction and adaption in any medium and for any purpose provided that it is properly attributed. See: https://creativecommons.org/licenses/by/4.0/. For attribution the title, original author(s), the publication source – Journals Library, and the DOI of the publication must be cited.
Bookshelf ID: NBK584537
Contents

Views

  • PubReader
  • Print View
  • Cite this Page
  • PDF version of this title (1.4M)

In this Page

Other titles in this collection

Recent Activity

ClearTurn OffTurn On

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...