Chapter 5: Telestroke

Mr. Taylor is a 45-year-old man enjoying a summer vacation camping in rural southwest Virginia. He has been enjoying hiking and swimming with his wife and two young daughters for the last two days. Today at 5:25 P.M., his wife finds him in their tent “slumped over” with incomprehensible speech. She had just talked to him no more than 10 minutes prior and he seemed fine. She sees that his right side is paralyzed and immediately knows he needs emergency help. Emergency medical services (EMS) are summoned by the clerk at the park entrance. Mr. Taylor is lifted into the ambulance to be transported to the nearest hospital, which is a small 16-bed critical access hospital serving the surrounding rural area. During ambulance transport, the EMS provider identifies the patient as having a potential “stroke” using a validated prehospital stroke scale. Under a new protocol, the EMS provider connects with a telestroke neurologist at a regional academic medical center using a mounted tablet with 4G LTE connectivity. The neurologist reviews the history, medications, and vital signs with EMS and performs the NIH Stroke Scale (NIHSS). The neurologist's assessment suggests a severe stroke that may be eligible for thrombolytic treatment and perhaps a large vessel occlusion that may warrant endovascular therapy with clot retrieval. This information is relayed to the local emergency department physicians.

Upon arrival to the hospital, Mr. Taylor is rapidly taken to the computed tomography (CT) scanner, and the images are digitally transferred to the hub hospital's radiology server. A videoconference telemedicine cart is placed at the foot of the bed, and the telestroke consulting neurologist rapidly repeats the NIHSS to confirm the deficits. She discusses her findings with the patient's wife, explaining that he is a good candidate for the clot-busting drug, intravenous tissue plasminogen activator (tPA), and recommends immediate treatment. The tPA is administered at 6:35 P.M., 80 minutes since Mr. Taylor was last seen normal. The neurologist has notified the on-call interventional neuroradiologist at the hub hospital. Using multipoint videoconferencing, they discuss the patient's neurovascular imaging, including a CT angiogram that demonstrates an abrupt “cut-off” sign indicating an embolus in the left middle cerebral artery. The emergency department (ED) physician, the neurologist, and the interventionalist discuss with the patient's family by live videoconferencing and recommend that Mr. Taylor be air-lifted to the hub stroke center for immediate clot retrieval. Mr. Taylor successfully undergoes the procedure and is able to discharge to home with minimal neurological deficits several days later. He is independent and back to work at 3 months.

The paradigm of acute stroke care shifted in the 1990s with the approval of intravenous (IV) tPA, rendering stroke a time-sensitive condition requiring rapid availability of neurological expertise. ¹ For many hospitals and communities unable to provide continual stroke expertise, particularly in rural and underserved areas, ^2,3 the emerging practice of telemedicine offered a virtual bridge to support rapid tPA decision making between centrally based consultants and remote-access patients and emergency providers. This concept was crystalized with the term telestroke in 1999 by Levine and Gorman. ^4,5

Over the past two decades, telestroke systems of care have expanded to take a prominent role in the practice of stroke medicine in the United States, Europe, and around the world. According to a 2009 nationwide survey, there were over 50 active telestroke programs in the United States, predominantly serving small and rural hospitals, but this number has undoubtedly grown over the past 5 years (Figure 5-1). ⁶ Additionally, numerous for-profit telestroke companies have cropped up in recent years, producing a nationwide network of independent teleconsultants and blurring the traditional geographic demarcations of regional stroke care. Moreover, advances in acute stroke treatment, including the recently established prominence of endovascular therapy (EVT) for severely affected patients with large vessel occlusions (LVO), ⁷ have placed a renewed premium on the role of telemedicine to help restructure our stroke systems of care. In this way, telestroke models are being explored to facilitate better interfacility access to endovascular services, and in the prehospital setting, to guide earlier diagnosis and triage decisions.

In this chapter, we highlight both the established aspects of telestroke care and burgeoning concepts in its ongoing evolution. We will explore various telestroke models and modalities, traditional and emerging technology, the underlying evidence base, financial and regulatory considerations, quality of care and limitations, and relevant guidelines and statements on the practice of telestroke. Of note, we will not cover applications of telestroke in subacute or chronic care, including stroke telerehabilitation and preventive telestroke services in the ambulatory setting, both of which could be expounded upon in their own form.

From this chapter, we hope that readers will not only learn the practical tenets of stroke telemedicine, but also appreciate the unique and pivotal role of telestroke in the wider world of telehealth.

Performing a high-quality telestroke consultation requires a real-time virtual partnership between the consultant neurologist at the hub and the emergency room staff at the spoke facility. Although the technology to support this interaction must be reliable and secure, it is the personal connection between the teleprovider, telepresenter, and the patient that serves as the foundation for a successful telestroke encounter. The Star Telehealth Alliance offers certification training for bedside telepresenters assisting during acute telestroke encounters and can be accessed at the following link: http://www.startelehealth.org/.

The principal features of the acute telestroke consultation include:

• – Establish a remote audiovisual (AV) link

• – Obtain a brief history, including key elements such as “time of last known well”

• – Perform the NIHSS with the assistance of a local telepresenter (Figure 5-2)

• – Review initial brain imaging, typically a noncontrast head CT via teleradiology

• – Determine eligibility for IV tPA or other acute stroke treatments

• – Offer guidance for further management or transfer to a higher level of care (eg, drip and ship)

• – Provide feedback to the spoke facility and emergency medical team

Although the telestroke evaluation has proven reliable for performing the standardized NIHSS, it is not validated in nonstroke diagnoses or as a substitute for the full neurological exam. For instance, in the absence of a neurologically trained telepresenter, visual observation alone is not reliable for complex visual field testing, manual assessment of tone and motor strength, obtaining reflexes, distinguishing fine sensory modalities, and additional bedside maneuvers (eg, vestibular testing, discerning organic from nonorganic findings).

In some large series, up to 20% of patients treated with IV tPA receive a nonstroke discharge diagnosis. ⁸ However, the challenge of treating stroke mimics was no different between telestroke and face-to-face encounters. ⁹ More common stroke mimics include migraine headaches with neurological accompaniments, seizures, metabolic/toxic/infectious encephalopathies, mass effect, reactivation of old stroke symptoms/signs, and psychogenic etiologies. In an attempt to minimize treating mimics during complex, time-critical telestroke evaluations, the TeleStroke Mimic Score (TM-Score) was devised. ¹⁰ The score is based on factors independently associated with stroke mimic status, including age, medical history (eg, atrial fibrillation, hypertension, seizures), facial weakness, and NIHSS >14. The TM-Score performed well on receiver operator curve (ROC) analysis (internal validation area under curve = 0.71, external validation area under curve = 0.77) and perhaps should be utilized more often in acute telestroke evaluations.

There are a variety of telestroke clinical care models, including hospital-based hub and spoke, for-profit telemedicine company based, hub and spoke, hubless private practice physicians, and multipoint telestroke consults (Figure 5-3). ⁶

The “hub and spoke” model is the traditional telemedicine model, with the hub consisting of the providing consulting hospital, and the “spoke” the network of hospitals that the providers service. The hub is either a Joint Commission–certified Primary Stroke Center or a certified larger Comprehensive Stroke Center tertiary care medical center. This model is well suited for either nonprofit or for-profit medical systems of care.

The for-profit telemedicine company-based model is in essence a spoke and hub model, with the hub being the company but the provider pool decentralized and consisting of physicians across the United States with service also rendered from homes, offices, or other site. Another model, the “hubless private practice physicians,” has been gaining in popularity, particularly with physicians who are contracted by a hospital system or are practicing in an urban setting. Leaving a busy office practice in the middle of the day to cover an emergency stroke care call is not practical for patient care. Instead, in this model the practice uses an office-based telestroke program to efficiently see stroke consultations in the ED. These four models can be used for either acute or nonacute consultation, in a variety of settings, from the ED, to neurological intensive care units (ICUs), to inpatient wards, or community health centers.

Focusing on acute telestroke consultation in the hub and spoke model, the primary goal is to provide neurological support for the local ED physicians in determining eligibility for IV thrombolysis or IV tPA. In this scenario, acute stroke patients are often treated locally and then transferred to the hub or another primary stroke center for further inpatient or ICU management. This model has been dubbed “drip and ship,” and observational studies suggest improved outcomes for patients transferred to a higher level of care at certified stroke centers compared to those who remain at the initiating facility. ^11–13

Emerging Models

A burgeoning model of telestroke advances care to the prehospital setting, facilitating remote neurological consultation with EMS providers prior to hospital arrival. The 2009 AHA/ASA Review of the Evidence for the Use of Telemedicine Within Stroke Systems of Care addressed the possibility of ambulance-based telestroke by stating, “It is clear that existing technology can provide some degree of interactive video and audio communication with preshoptial units in transport, although current published applications have unacceptably low frame rates, and broad application of this technology to large fleets of EMS vehicles is not yet practical … The usefulness of this intervention in real practice is uncertain, and further research is required.” ¹⁴

Although the concept of mobile or ambulance-based telestroke dates to the late 1990s with the TeleBAT pilot studies, ^15–17 feasibility has improved in the last several years with the rapid evolution in mobile, or m-health, technologies and high-fidelity 4G cellular connectivity to support real-time AV streaming, ^18–22 Efforts to make prehospital telestroke more practical using tablet-based endpoints and low-cost, off-the-shelf components have been demonstrated in both rural and urban EMS networks, including the Improving Treatment with Rapid Evaluation of Acute Stroke Using Mobile Telemedicine (iTREAT) Study. ^23–26 Pilot trials of ambulance-based telestroke suggest the possibility of faster stroke treatment times, ²⁷ and future trials will seek to translate technical feasibility into improved clinical outcomes.

The prehospital telestroke model has been taken a step further with the development of so-called “mobile stroke units,” customized ambulances equipped with a portable CT scanner and other advanced diagnostics, with the capability of administering IV thrombolysis in the field. ^28–31 Although effective in significantly reducing stroke onset to treatment times, these units have primarily been deployed in urban or suburban EMS networks with limited coverage areas, and cost utility remains a barrier to widespread adoption and generalizability across health systems. ³² Nonetheless, telemedicine and teleradiology may help streamline efficient mobile stroke unit deployment. ^33–35 Mobile stroke units could also enhance prehospital diagnosis and select triage for patients with LVO to facilities capable of providing EVT, ^36,37 but further research and systems-of-care considerations are needed.

At its inception in the 1990s, a telestroke consult required that the physician use a fixed workstation and in some cases a second computer for documentation of the encounter or reviewing films. However, because the “time is brain” concept mandates instant access to the patient for a rapid evaluation for potential therapy, this was not an ideal 24/7 telestroke model.

Telemedicine workstations continue to exist in the hospital or physician's office, but are complemented by more mobile technology options such as laptops, tablets, and smart phones. Cloud-based technology (circa 2006) facilitated doing telestroke consultations using these mobile technologies. Consultations can be provided wherever there is a secure Internet connection and via “hot spot”–generating cellular connections. These advances allow the provider significant access to web-based tools for documentation, imaging, and electronic medical records review.

Furthermore, telestroke-tailored, software-bundled applications allow for live video consultation (VC) encounters, with the consultation documentation and the HCT imaging for review, all into one convenient screen. The development of “multipoint” capacity, having more than two terminals connected by a single communications channel, allows for two or more physicians to join the VC encounter and confer on the clinical case. For example, the neurointerventionalist can enter the call to discuss the candidacy of a patient with the stroke neurologist “live” and directly with the ED physician, the patient, and/or their family.

Brain/Vascular Imaging Transfer (Teleradiology)

First conceived in 1982, Picture Archiving and Communication Systems (PACS) is a medical imaging-based technology that provides convenient access to images from multiple modalities. PACS allows electronic images and reports to be rapidly transmitted digitally via a PACS link. In the United States, the University of Kansas developed one of the first PACS, and the University of California Los Angeles deployed pilot PACS in their pediatric radiology department. The first filmless fully digital radiology department using PACS was in the Danube Hospital in Vienna in 1991. The current PACS is critical for fast interpretation of head CT scans in the evaluation of the stroke patient who may qualify for the time-sensitive thrombolytic TPA. Today, most PACS can be accessed securely using cellular or Internet connections.

Telestroke patient imaging includes an emergency noncontrast CT brain scan to identify any intracranial hemorrhage (which is a contraindication for thrombolytic treatment). In selected cases, CT angiography of the head and neck, brain magnetic resonance imaging (MRI), and/or magnetic resonance (MR) angiography are obtained. A newer modality, a multiphase CT, has been used to assist in the selection of patients who may benefit from thrombectomy. This technique generates time-resolved cerebral angiograms of brain vasculature from the skull base to the vertex in three phases after contrast material injection. By evaluating the potential for collateral vessels at the later phases, the teleprovider can better predict outcome.

VC equipment has advanced tremendously since the first “peer to peer” VC device proposed by AT&T at the New York's World's Fair in the 1960s. The technology improved with the introduction of the Network Video Protocol, and in the 1980s, the availability of digital transmission of voice and video by electrical pulses, through the Integrated Services Digital Network. In 1991, IBM introduced the first personal computer VC product, followed by Multipoint VC by Apple Macintosh. In 1995 the first VC occurred internationally between North America and Africa. During this time frame, high-capacity broadband services and more powerful computing processors followed high-definition systems, “icloud” applications, and lower-cost video-capturing technologies, leading to today's highly sophisticated VC equipment and systems. A critical shift occurred from company-based proprietary equipment and software to standards based technology. The Unified Communications Interoperability Forum (UCIF), a nonprofit alliance between communications vendors, was launched in May 2010. The UCIF created VC standards and protocols to ensure interoperability regardless of the platform.

Telestroke VC equipment requires the ability to perform an observational bedside assessment that includes components of alertness and cognition, vision, speech, motor, sensory, and coordination (ie, NIHSS). Optimally the system has a high-definition, remote-controlled video camera with a “pan-tilt” and zoom capability, or “PTZ,” camera. In addition, omnidirectional microphones are essential for quality sound transmission.

Current Internet speeds can easily handle transmission of most radiological images required to evaluate the patient. Cellular 4G (50 Mbps download/350 Mps upload), broadband wireless (600 Mbps – 2007), and LAN/fiber-optics (10 Gbs – 2010) will undoubtedly continue to advance in the future. Initiatives to expand the footprint for telemedicine include federal programs, such as the FCC's Connect America Fund (https://www.fcc.gov/encyclopedia/connecting-america) and HRSA's regional funding of Telehealth Resource Centers (http://www.telehealthresourcecenter.org).

The evolution of mobile media, or m-health, will also shape the future application of mobile technologies in telestroke. ³⁸ Telestroke assessments have been validated across a range of mobile endpoints, including smart phones and tablet-based devices. ^39–41 Advances in mobile imaging may lead to the ability to evaluate brain hemorrhage using handheld devices in the future. ⁴²

Regarding security, telestroke providers must ensure that telemedicine networks are secure and encrypted. Although this is crucial, doing so without compromising the real-time flow of voice and video is challenging. The American Telemedicine Association has issued guidelines for a variety of subspecialties, and currently, FIPS 140-2, known as the Federal Information Processing Standard, is the U.S. government security standard used to accredit software encryption standards and lists encryption techniques such as AES (Advanced Encryption Standard) as providing acceptable levels of security. ⁴³ All providers must use extreme caution and restraint in using personal devices to provide clinical care.

To date, only a few prospective randomized and comparative trials have assessed the clinical efficacy of telestroke intervention. The Stroke Team Remote Evaluation Using a Digital Observation Camera (STRokE DOC) trial randomized acute stroke patients at four remote spoke sites (n = 222) to AV telestroke (intervention) versus telephone-only consultation (control), with the primary outcome of accuracy for thrombolysis treatment decision as assessed by masked central adjudication. The trial was stopped early for superiority of video-based telestroke with an absolute difference in correct decision making of 16% favoring the intervention (98% versus 82% correct; odds ratio [OR], 10.9; 95% confidence interval [CI], 2.7–44.69; p = 0.0009). ⁴⁴ Although the trial failed to show any difference in clinical outcomes between the two groups, the study was not powered for these secondary measures.

The STRokE DOC AZ TIME trial was a follow-up study conducted in a single-hub, two-spoke system in rural Arizona. Using similar methods and patient selection (n = 54), there were no differences between AV telestroke and telephone-only consultation regarding thrombolysis decision making; however, the intervention was limited by technical issues (74%). ⁴⁵ A pooled analysis from the STRokE DOC trials further supported the efficacy of AV telestroke versus telephone-only consultation to enhance decision making for thrombolysis eligibility (96% versus 83% correct; OR, 4.2; 95% CI, 1.7–10.5; p = 0.002), but results were driven largely by data from the original trial. ⁴⁶

From the Telemedic Project for Integrative Stroke Care (TEMPiS) network in Bavaria, Germany, a nonrandomized, open-label study found that patients treated in network hospitals were less likely to have a poor outcome (moderate to severe disability) at 90 days compared to out-of-network community hospitals (OR 0.62, 95% CI 0.52–0.74; p < 0.0001). ⁴⁷ Review of the longitudinal experience in this network demonstrated sustained benefit in increased proportion of acute stroke patients receiving IV thrombolysis, reduction in median treatment times, and improved clinical outcomes. ^48,49 Similar observational results have been reported from numerous retrospective cohorts and case series. ⁵⁰

Supplementing the limited number of comparative telestroke trials is a robust body of literature supporting the feasibility and generalizability of telestroke in practice. Fundamentally, numerous studies have confirmed the interrater reliability of performing the NIHSS via remote, real-time AV consultation, although some components are less reliable than others. ^51–53 NIHSS reliability has further been confirmed across multiple telemedicine endpoints, including smart phone and tablet devices as previously mentioned. ^39,40,41,54,

Several studies have assessed the level of agreement for head CT interpretation in the telestroke encounter, essential to acute management and determination for thrombolysis eligibility. ^55,56 A pooled analysis of data from the STRokE DOC trials (n = 261) found excellent agreement in head CT interpretation between spoke radiologists and hub neurologists (95.4%, kappa = 0.74, 95% CI 0.59–0.88). ⁵⁷ Additionally, at least one other observational study found no difference in final diagnosis of stroke, confirmed by brain imaging, in thrombolysis-treated patients at the telestroke hub versus spoke hospitals. ⁵⁸

In addition to studies supporting the practice of telestroke in a hub and spoke model, emphasis has focused on telestroke systems of care to improve access to treatment and neurological expertise in rural and underserved areas. ^59–62 In 2014, the Health Resources and Services Administration (HRSA) announced grants worth more than $22 million to support health care in rural areas, including funds to improve emergency services. The HRSA “Evidence-Based Tele-Emergency Network Program” was created to determine the effectiveness of tele-emergency care for rural patients and providers, including telestroke care.

In the Rural “EQUITe” Project, designed to evaluate the effectiveness of telestroke to improve access and timeliness of treatment for rural patients in Virginia, data were pooled from 46 rural hospitals encompassing 3,257 emergency telestroke consults. ⁶³ Nineteen percent of patients were treated with tPA, and the majority (73%) were treated within 3 hours from the time of last known normal. In the majority of cases, the median “door-to-needle” time was 83 minutes, and about a fourth of the patients received tPA in less than or equal to 60 minutes from their arrival to the ED. Although this was in many cases an improvement from pretelestroke benchmarks, it revealed that systems improvements were needed to augment the efficiency of the telestroke encounter at many of these rural hospitals.

As the treatment of acute ischemic stroke has evolved with evidence supporting EVT for select patients, there is a dire need to improve stroke systems of care capable of rapid identification and triage between first-arrival spoke hospitals and endovascular-capable facilities. ⁶⁴ Although observational experience suggests these processes may be enhanced by telestroke systems, ⁶⁵ further research is needed to examine the impact of telestroke in the endovascular era.

Telestroke shares the same barriers in reimbursement, regulation, and licensing as with other telemedicine subspecialties. ^66–68

At the time of this writing (2017) not all 50 U.S. states have parity insurance reimbursement with the same “in-person” insurance reimbursement schedule for the same service. Strides are being made in rural telestroke reimbursement from the Centers for Medicare & Medicaid Services (CMS), but many needful urban-based encounters remain unreimbursable, which is a disincentive for hospitals to develop the service. For Medicaid, the majority of states cover telemedicine services statewide without distance restrictions or geographic designations. ⁶⁹

For Medicare, the originating site—where the patient is located—must be a facility located within a primary care health professional shortage area (HPSA) and/or outside of a metropolitan statistical area (MSA). By definition, these are sites with fewer than one primary care physician per 3,500 people and cities with no more than 50,000 residents. ^70,71 Twenty-two states and Washington, DC, have private insurance telemedicine parity laws providing statewide coverage, with no provider, technology, or patient setting restrictions. ⁶⁹

Federal legislation has been introduced in the United States to increase access and reimbursement for telemedicine services. Specific to telestroke, the bipartisan bill “Furthering Access to Stroke Telemedicine,” or the FAST Act (H.R. 2799 and S.1465), was introduced in 2015 to broaden the definition of the originating site to include both rural and urban spoke facilities for CMS reimbursement. At the time of this writing, the FAST Act legislation, widely supported by professional bodies, including the American Telemedicine Association, American Heart Association/American Stroke Association (AHA/ASA), and American Academy of Neurology, has subsequently been referred for committee review in the House and the Senate and awaits final determination.

Beyond coverage reimbursement, the economic sustainability of telestroke networks requires buy-in and ongoing support by key stakeholders ranging from hospital leadership to nurses and first responders at the point of care. Additional financial considerations include contract negotiations between hub and spoke sites, adequate salary support or incentives for telestroke providers, capital budgeting for technology startup and maintenance, return on investment, and marketing strategies to promote the telestroke program in the surrounding community. ⁶⁶

Numerous studies have also assessed the cost effectiveness and cost utility of telestroke programs in both centralized and decentralized health care systems. Most analyses have concluded a favorable incremental cost-effectiveness ratio (ICER) well below societal thresholds for reimbursement, particularly when considering outcomes across the lifespan. ^72–78 Whereas most of the attendant costs in a telestroke system are in startup, societal return on investment is substantial when considering long-term savings in reduced lifelong disability for stroke patients.

From a business model perspective, hospitals can realize a return on investment incurred by an increased number of patients treated with tPA (ie, higher diagnosis-related group billing) at the spoke, transfers to the hub for EVT, and patients discharged from both hub and spoke. ⁷⁹

Since the inception of telestroke systems in the early 2000s, there have been several evidence-based guidelines and policy statements guiding practice and implementation for acute stroke care. Given limited data from prospective randomized trials, the literature supporting the implementation of telestroke primarily reflects analyses of practical feasibility, observational experience, and recommendations from a systems-of-care perspective (Table 5-1). ^80,81

A recent scientific statement from the AHA/ASA suggested new recommendations for monitoring quality and outcomes in telestroke programs. ⁸² Specifically, these recommendations include measuring processes such as time to treatment and hospital transfers, documenting patient outcomes and safety events, and technical quality metrics of the telestroke system. Many of these recommendations overlap with previous stroke guidelines and hospital-reporting requirements for certified stroke centers; however, a number of additional metrics may serve as new quality standards and benchmarks for telestroke practice going forward (Table 5-2 and 5-3).

In 2017, the American Telemedicine Association drafted Practice Guidelines for Telestroke “to assist practitioners in providing assessment, diagnosis, management, and/or remote consultative support to patients exhibiting symptoms and signs consistent with an acute stroke syndrome, using telemedicine communication technologies.” This comprehensive document covers a broad scope, including clinical, administrative, and technical guidelines, and seeks to highlight “the unique aspects of delivering collaborative bedside and remote care through the telestroke model.” Although these guidelines are currently under review at the time of this writing, they represent another important step in the effort to standardize high-quality telestroke care for patients and communities (http://hub.americantelemed.org/resources/telemedicine-practice-guidelines).

By any account, the establishment of telestroke is a true success story in the annals of telehealth. Throughout much of the world, telestroke partnerships are now an integral component of emergency care and have helped ameliorate geographic disparities in access to acute stroke treatment. Despite widespread acceptance of telestroke in clinical practice, much work remains to broaden the evidence base and integrate quality improvement metrics across telestroke systems of care. As many hospitals explore telestroke programs for the first time, efforts are underway to utilize telestroke for novel means, particularly with innovations in mobile technology for prehospital stroke care and as a tool to facilitate enhanced regionalization of care in the endovascular era. Financial and regulatory considerations will continue to shape the expansion of telestroke, as proponents await the outcome of important legislation such as the FAST Act to guide reimbursement in years to come.

Despite much progress, stroke persists as one of the most disabling and devastating medical conditions, with stroke incidence estimated to more than double by 2050 as the population ages. ⁸³ Thus, great opportunity remains to further utilize telemedicine to improve the outcomes for stroke patients and lessen the burden of stroke on society.