The Status of Telestroke in the United States
A Survey of Currently Active Stroke Telemedicine Programs
Background and Purpose Little is known about adoption or success of telestroke networks outside of published or federally funded programs. Under contract to the Health Resource Services Administration, we conducted an environmental scan of telestroke programs in the United States.
Methods An analyst contacted all potential programs identified in comprehensive online searches, interviewed respondents, and collected response data about structural and functional components of currently operating telestroke programs.
Results Among 97 potential programs contacted, 56 programs had confirmed telestroke activity, and 38 programs (68%) from 27 states participated. Hospital and community characteristics of nonparticipating programs were similar to those of participating ones. The top 3 clinical needs met by the telestroke were emergency department consultation (100%), patient triage (83.8%), and inpatient teleconsultation (46.0%). Telestroke programs were in operation a median of 2.44 years (interquartile range, 1.36–3.44 years); 94.6% used 2-way, real-time interactive video plus imaging, but only 44% used dedicated telemedicine consultation software. The mean number of spokes per hub increased significantly from 2007 to 2008 to 2009 (3.78 versus 7.60; P<0.05), and >80% of spoke sites were rural or small hospitals. Reimbursement was absent for >40% of sites. Sites rated inability to obtain physician licensure (27.77%), lack of program funds (27.77%), and lack of reimbursement (19.44%) as the most important barriers to program growth.
Conclusions Telestroke is a widespread and growing practice model. Important barriers to expansion amenable to change relate to organizational, technical, and educational domains and external economic and regulatory forces.
Stroke is the fourth leading killer in the United States and the leading cause of long-term disability. Approximately 800 000 people experience a stroke each year in the US.1 Successful management of acute ischemic stroke is extremely time-dependent. Ideally, the only Food and Drug Administration-approved treatment for acute ischemic stroke should be administered within 3 hours of the onset of stroke symptoms; however, recent guidelines recommend that selected patients can be treated up to 4.5 hours after symptom onset.2–4
The American Heart Association estimates that only 3% to 5% of ischemic stroke patients are treated with thrombolysis.5 There are many reasons that contribute to the uncommon use of intravenous recombinant tissue-type plasminogen activator (rt-PA) in the acute ischemic stroke setting: poor community awareness of stroke symptoms, lack of on-site stroke-care expertise, long distances to tertiary care hospitals, and concerns with the risks of intracranial hemorrhage.6,7 The use of telemedicine in the treatment of stroke has shown great promise for improving patient access to recommended stroke treatments in rural areas and other areas underserved by neurologists.8,9
Using telemedicine technologies, including high-quality videoconferencing for face-to-face interactions and remote viewing of images, the patient and the physician are not required to be in the same room, or even in the same state, for a high-quality care interaction to occur.10,11 Although telehealth programs have existed for many years, over the past decade, telemedicine programs focused on certain diseases have rapidly emerged as a sustainable model of care delivery; this is especially the case in urgent care scenarios as in acute ischemic stroke.12–14
Little is known about the adoption or success of telestroke networks outside the context of federally funded programs or series published in the medical literature.12,14 With the US in the midst of an economic crisis, healthcare and payment reform occurring at both the national and state level, and a growing population of patients needing care for stroke, it is critical that a comprehensive assessment of telehealth adoption in specific diseases be performed.15,16 Our objective was to conduct an environmental scan of telemedicine-based stroke programs in the US and to identify both critical success factors and barriers to the development or sustainability of telestroke programs.
A dedicated project analyst carried out all data collection. She performed Google searches to find potential programs, contacted the programs identified to verify eligibility, secured their participation in the project, and distributed the survey. Once the online surveys were completed, the project analyst conducted in-depth telephone interviews with site representatives.
Online Information Search
A search for telestroke programs was conducted using several methods. A keyword search was performed in Google utilizing the terms telestroke, telemedicine, or stroke. Google alerts were also established so that any recently indexed online content pertaining to the above search terms were delivered by Google via e-mail. In addition, a list was generated from US News and World Reports 2009 Best Neurology Programs in the US, and these programs were searched for the possible presence of any telehealth programs related to stroke.17 When a potential program was identified, searches for contact information at the individual hospital websites were performed. Any personal contacts of the investigators were also used to identify relevant hospitals and individuals. Furthermore, several for-profit telehealth companies with programs in telestroke were contacted, and InTouch Health and REACH Call agreed to provide us with hospital contacts that otherwise were not identified through our search strategy. Finally, a personal list of known telemedicine stroke program directors and industry thought leaders from academia and for-profit companies from 1 of the authors (L.H.S.) and from the Office of Advancement of Teleheath at Health Resource Services Administration was used to complement the initial search. A more detailed description of the search methodology can be found online (Supplemental Methods).
Detailed information on the county population was obtained from the US Census Bureau.18 Rural hub hospitals were defined as nonmetropolitan by Rural-Urban Commuting Area codes, and for spokes as those located in areas outside of metropolitan centers where there is not ready access to specialist, intensive, and/or high-technology care, and where resources, both human and material, are lacking.19
People and Programs Contacted
A list of potential programs was compiled, and multiple attempts were made to contact all identified programs. Contact attempts included use of phone and/or e-mail when available. Each site was contacted repeatedly until the program declined or clarified that it did not offer telestroke services.
Using SurveyMonkey, an online survey tool was developed to assess various aspects of the telemedicine stroke programs. A copy of the final survey tool administered to participants is available online (Supplemental Table S1). The online survey contained 6 general questions (eg, contact information, organization information), specific questions about telestroke, and was distributed to participants along with a letter that explained the purpose of the survey. Before distributing the survey to any outside participants, the survey was piloted internally and was refined using members of the Partners Telestroke team to assess it for content, clarity, and thoroughness. Following the completion of the online survey, telephone interviews were conducted with each of the respondents.
Participants were encouraged to share specific details about their program and were also able to request anonymity if they so chose for attribution of individual comments. The interviews were conducted over the telephone. All of the telephone interviews were recorded for the purposes of transcription. Each interview lasted about 30 minutes (range, 20–40 minutes).
Descriptive statistical analyses of publicly available demographic and health services data on the programs contacted were performed. Discrete variables were analyzed by χ2 or Fisher's exact test as appropriate, and continuous data with Wilcoxon Rank-Sum for nonparametric data or Student t test for normally distributed data as appropriate.
Ninety-seven potential programs across 43 states were identified and contacted to assess eligibility and offer participation. Of the 56 programs with confirmed active telemedicine programs in stroke, 38/56 programs (68%) from 27 different states agreed to participate and are displayed inFigure 1.
Of the 38 programs, 36 programs (95%) were traditional hub and spoke networks and 2 were hubless; there were 33 traditional medical center hubs and 1 spoke site, as well as 2 spokes with neurology coverage from a for-profit company (Figure 2). The person answering the survey was a physician in 52.6%, a registered nurse in 18.4%, a PhD in 2.6%, and an administrator in 26.4%. The questionnaire's answers were not statistically different according to the occupation of the respondent. We therefore report the grouped results.
The hospital and county level characteristics of programs that did not participate were similar to the ones that did, in terms of county level data (total population, % population age ≥25 years, % of individuals below poverty line), annual hub hospital admissions, number of beds, and annual emergency department visits (Table 1).
Respondents identified the internal factors and motivations for implementing telestroke programs, characterized the organizational elements of their networks, identified current reimbursement models, and classified the importance of recognized barriers to broader program adoption.
Internal Factors and Motivations for Implementing Telestroke Programs
Current telestroke programs all support emergency department consultation for specialty care, and many also use the systems for patient triage when determining patient acceptance for transfer to primary or comprehensive stroke center care. Providing a community benefit (97%), improving clinical outcomes (92%), and improving clinical process effectiveness or provider knowledge (76%) were the 3 most common internal factors that prompted the creation of current telestroke programs (Table 2). Notably, revenue enhancement, cost reduction, and bed management were among the lowest-ranking reasons to create a program. External forces, such as competitive forces in the marketplace (32%), and new state legislation (27%) or changes in reimbursement policies (22%), did influence programs.
The hospitals were all asked what the driving force was behind starting up these programs. In several cases, the hub hospitals were already receiving calls for help with patients and they found it difficult to manage cases without being able to visualize the patient or their brain imaging. Because many of the patients would end up being transferred to the hub hospital by default, the hubs found telemedicine was a good way to start managing the patients even before they arrived. Another frequently cited reason for starting telemedicine programs was that the hospital observed health disparities across their service area. A patient who is not living near a major medical center was perceived to be receiving a different level of care from someone who is.
Organizational Elements of Telestroke Networks
Networks varied in the size and scope of their mission and the duration of the program, but adoption has been increasing steadily with a dramatic rise in the last 2 years of the 3-year period reviewed. The mean number of spokes per hub increased significantly from 2007 to 2008 to 2009 (3.78 versus 7.60; P<0.05; range, 0–28). In 65.5% of hubs, >80% of spoke sites were rural, and in 51.7% of hubs, >80% of spokes were small hospitals (0–99 beds). The median telestroke program operating duration since 2000 across all US sites surveyed was 2.44 years (interquartile range, 1.36–3.44 years).
There was variability in terms of how often the spokes had a pre-existing relationship with the hub as part of their corporate network. Although some hubs support remote spoke sites that are only within their hospital's formal organizational network (20%), or that are only outside their hospital's formal organizational network (20%), the vast majority of hubs (60%) support spokes that are either within or outside their hospital's formal organizational network. Formal written contracts or agreements are in place at all spokes in most networks (81%), and in at least some spokes at almost all programs (89%).
Sites use a variety of technologies to support their telestroke work, with essentially all sites (95%) using high-quality videoconferencing (2-way, real-time interactive video with high resolution and frame rates). Single-vendor technical solutions were common (Polycom, 24%; REACH Call, 18%; Tandberg, 11%; InTouch Health, 8%) and hybrid solutions were frequently used (multiple vendors, 21%). More than half of the programs (68%) review brain imaging as part of the consultation process, and almost half of the programs (49%) incorporate telephone-only consultation or store and forward (11%) when appropriate. At 56% of hubs, the systems used can only interact with other systems of the same models by the same manufacturer (closed communication networks). Most sites (68%) use the same basic approach to telemedicine that they used at the initiation of their program. Advice or mentoring from an outside organization before making technology or program design decisions were utilized by 60% of the telestroke programs evaluated.
The most common methods used to store and communicate the documentation of the telestroke consultation to the spoke site are listed inTable 3. Fewer than half of the programs have a dedicated electronic health record system specific to telestroke, and many sites use a combination of paper, dictation, and other means to document the care delivered. Consult multimedia files are archived to a variable degree, with the minority of sites retaining a video recording of the cases (17%).
The composition of the telestroke team varied by site, depending on the resources available at the hub, or by an outside entity if no hub was utilized. Whereas consults were most often (≈75%) provided by specialists employed by the hub hospital's organization, just over a quarter (≈25%) of sites used outside contracted specialists as part of the telestroke team. Telestroke consultants currently practice medicine in the program's state in 97% of the cases, and nearly all include a physician (97%); but, some programs also incorporate other providers, such as registered nurses (38%), advanced-practice nurses (11%), or physician assistants (8%). There is round-the-clock access to dedicated technical support during the consultation process in most sites (86%), and this contributes to operating costs.
Clinical quality of the programs is measured in 91.9% of the programs evaluated, and includes specific case reviews for compliance with protocols (53%), and measurement of processes of care, such as treatment time intervals (44%) and patient satisfaction (6%). Additional services are also provided through the networks that go beyond the specific telestroke encounters including: providing follow-up clinical information (45%), quality metric reports (18%), education and training (9%), service utilization reports (3%), or all of the above (15%).
Current Financial and Reimbursement Models
Third-party health insurance payment for the physician services rendered was uncommon, with no applicable coverage in most networks (43%), followed by some private insurance (30%;Table 4). When available, oftentimes these payments were described as inadequate.
Importance of Recognized Barriers to Broader Program Adoption or Sustainability
Sites rated inability to obtain physician licensure (27.7%), lack of funds (27.7%), and lack of reimbursement (19.44%) as the most important barriers to program growth. The barriers most frequently mentioned during the telephone interviews were: licensing/credentialing challenges, lack of technology support at the spokes, lack of funds, and trouble with physician buy-in. Each site identified and ranked the barriers most likely to have an impact in increasing growth or development of telestroke if they were removed. Although lack of start-up or maintenance funds and lack of reimbursement were at the top of all lists, lack of physician buy-in at the spokes more than at the hubs emerged as an important barrier to growth; the administrative challenges related to licensure and credentialing were identified as a high-priority barrier to remove for telestroke programs.
The results of this environmental scan demonstrate that current telestroke programs are steadily growing over the years evaluated, that all programs support emergency department consultation for specialty care, that most use the combination of real-time interactive video and teleradiology to support the consultation process, and that, unfortunately, reimbursement for telestroke is still limited.
An important part of the application of a new technology depends on human factors and the ability to apply the new technology in a variety of scenarios, from research proof-of-concept environments to real-time acute care delivery interventions and novel practice settings. Modifying the barriers identified by this environmental scan appears a critical next step in realizing the tremendous potential of current and future telehealth applications for all citizens, not just those fortunate enough to live in areas serviced by existing programs.
Key activities that might promote expansion of telehealth services based on our results are: reimbursement for clinical services, streamlined licensure and credentialing, national standards for malpractice determinations, minimal technology performance standards, and incorporation of telehealth programs into comprehensive systems of care delivery models. The majority of the telestroke programs measure clinical quality, but they appear to do so by very disparate methods. National standards and designation-certification process specific for telestroke programs could help to homogenize such evaluations and to create benchmarks to allow appropriate comparisons between programs. In a recent survey of barriers to telemedicine in acute care units, respondents declared that telemedicine program success was still halted by licensing, credentialing, and malpractice protection, as well as costs, billing, and reimbursement issues.20
We found that coverage for telestroke services was uncommon, and that even when there were available payments, they were described as inadequate. Other authors have rated reimbursement as 1 of the most important issues for the success of telestroke systems in the US.10,16,21 Health insurance restrictions on reimbursement need to be modified to accept the validity of this new model of healthcare delivery. Medicare acknowledges the problem of reimbursement in acute stroke care and has created a higher-value diagnostic related group (DRG 559) for treatment of acute ischemic stroke with thrombolytics. Hospital costs have traditionally been greater than Medicare reimbursement for the acute care of patients treated with intravenous rt-PA, which includes not just the cost of the drug, but also advanced imaging, intensive care, and longer lengths of stay. The introduction of DRG 559 has provided more appropriate reimbursement for hospitals and removed the financial disincentive to use intravenous rt-PA if patients are admitted to the hospital where rt-PA was administered. However, for patients who receive rt-PA at a telestroke spoke hospital and then are transferred to the hub hospital, neither site is entitled to the higher DRG 559 payment.21–24 Stroke centers and telemedicine programs are associated with higher rates of intravenous thrombolysis administration within 3 hours of the onset of stroke symptoms. The first US cost-effectiveness analysis of telestroke was recently published and concluded that when a lifetime perspective is analyzed, telestroke seems cost-effective compared with usual care, mostly because telestroke costs are upfront and benefits of better stroke care are lifelong.25,26 More research is needed to demonstrate how these new models can be cost-effective for payers, especially outside the US. Allowing healthcare organizations more flexibility in how payment for services is covered and extending benefits to telehealth transactions would aid in adoption.
Sites rated inability to obtain physician licensure as 1 of the most important barriers to program growth and development. Because there is no national-level medical licensing and credentialing available for physicians, whenever a new spoke gets added to a telestroke network, physicians have to go through the credentialing process again. If the spoke is in another state, the physician must also go through the process of being licensed in that state.10,22 National repositories for physician data or national processes for licensure and credentialing are sorely needed, as is harmonization between regulatory agencies on required processes. Some states have already worked on decreasing the administrative problems that physicians face when they consult via telemedicine. The Model Act to Regulate the Practice of Medicine developed by the Federation of State Medical Boards calls for the issuance of special purpose licenses to physicians with a full license to practice in any state or US territory, as long as there have not been previous disciplinary actions taken against them.10,27 Expansion of telehealth services, especially across state lines, is not scalable without addressing this escalating administrative burden. Recently, The Centers for Medicare and Medicaid Services published a rule that implemented a new credentialing and privileging process for physicians providing telemedicine services. The rule permits hospitals to rely on the credentialing and privileging determinations of another hospital or telemedicine entity, rather than make an individualized decision based on the practitioner's credentials and record. The new rule makes available a more streamlined process if hub and spoke wish to accept it, and will help to remove some unnecessary barriers to the use of telemedicine.28
Although essentially all sites investigated use high-quality videoconferencing to support consultation, storage and technology technical support were still extremely variable among the programs. Technology providers should be required to meet certain minimum specifications for acceptable performance to be certified as legitimate providers. These should be set by independent agencies, with input from key stakeholders. Ideally, the mode of data transmission must provide adequate bandwidth to transmit large amounts of data rapidly and safely; this is so computerized tomography scans can be sent to the hub hospital for concomitant evaluation by the spoke's emergency department staff and the hub's stroke team. Unfortunately, many suburban and rural areas do not have continuous high-speed bandwidth availability to support high-quality video transmission. Therefore, the presence of minimal infrastructure must be confirmed in hospitals that plan to participate in telestroke programs.10 In addition, standards supporting the integration of multimodality communication methods into a single-communication medium would enhance productivity and decrease costs.
A previous review of the published medical literature evaluating telestroke initiatives in 2009 found a total of 20 telestroke networks (12 networks in the United States, 3 networks in Canada, and 5 networks in Europe) and concluded that the long-term sustainability of telestroke practice was threatened by unresolved legal, economic, and market factors.29 Our survey found similar barriers to telestroke growth. The same review urged the need for standardized measurements of telestroke quality of care and acceptable guidelines for telestroke practice.29 In the recommendations for the implementation of telestroke systems of care from the American Heart Association, important gaps to be addressed to warrant broad use of telemedicine included: discussion of licensure and liability legislation, development of admissible rules relating to the confidentiality of information, simplification of the process of requisition and conveying telemedicine appointments, improvement of the instruction of the end users, and development of economic models for reimbursement.10
Our study has some important limitations. First, only 68% of identified programs participated in our survey. Although such a figure constitutes a high response rate for a typical survey, with response rates for physician surveys routinely in the 40% to 60% range, we cannot exclude the possibility of nonresponse bias.30 Nevertheless, programs that did and those that did not participate were similar in terms of hospital- and county-level data, suggesting that they had similar structural characteristics. County-level characteristics, although not the perfect surrogates for the structural characteristics of the programs, were the best available information about the nonparticipating programs. In addition, it is possible that nonrespondent programs might perform worse than would participating ones. Because our objective was not to measure the quality of the programs evaluated, but instead to conduct an environmental scan of telemedicine-based stroke programs in the US, we do not think that our results were biased based on the possibility of different program quality levels. Second, data such as number of teleconsultations and of telemedicine-guided intravenous rt-PA administration per program were not available. When the survey instrument was elaborated, we specifically avoided questions regarding volume of consults and costs, because we were afraid that unwillingness to disclose that information might lead to a low response rate. Finally, despite the interesting information provided by telephone interview, data acquired are subjective and not amenable to a more structured statistical evaluation, and no data were externally audited for accuracy.
The ability to identify and address the organizational, technical, and educational challenges in the context of the current economic, legal, and regulatory environment will determine the pace at which telestroke programs can expand and provide access to the many patients in need of increased health care. Increasingly, inequities in access to information technology will translate into disparities in access to health care, and therefore government and private initiatives will be necessary to ensure ubiquitous access to secure high-quality and high-speed bandwidth for telehealth applications. In conclusion, telestroke is a widespread and growing practice model. Important barriers to expansion relate mostly to organizational, technical, and educational domains internal to organizations, and economic and regulatory forces externally. National standards for licensure, credentialing, and reimbursement could have substantial impact in access to acute stroke care.
The authors gratefully acknowledge the support and constructive feedback provided by the Health Resources Services Administration program officers.
Sources of Funding
This study was supported (Dr Silva, S. Farrell, MBA, Dr Schwamm) in part by a contract from Health Resource Services Administration Requisition 09-HRS9923-AB (Stroke and Traumatic Brain Injury Telehealth Services).
Dr Silva, S. Farrell, MBA, E. Shandra, MBA and Dr Viswanathan report no disclosures. Dr Schwamm is chair of the American Heart Association Get With the Guidelines steering committee (unpaid) and a consultant to the Massachusetts Department of Public Health for Stroke Systems Development. The Massachusetts General Hospital employs Dr Schwamm and Dr Viswanathan and the hospital provides contracted telehealth services to hospitals in New England for telestroke.
The online-only Data Supplement is available with this article at : http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.111.658724/-/DC1.
Jeffrey L. Saver, MD, was the Guest Editor for this paper.
- Received November 22, 2011.
- Accepted March 27, 2012.
- © 2012 American Heart Association, Inc.
- Roger VL,
- Go AS,
- Lloyd-Jones DM,
- Benjamin EJ,
- Berry JD,
- Borden WB,
- et al
- Del Zoppo GJ,
- Saver JL,
- Jauch EC,
- Adams HP Jr.
- Adeoye O,
- Hornung R,
- Khatri P,
- Kleindorfer D
- Kwan J,
- Hand P,
- Sandercock P
- Levine SR,
- Gorman M
- Schwamm LH,
- Holloway RG,
- Amarenco P,
- Audebert HJ,
- Bakas T,
- Chumbler NR,
- et al
- Schwamm LH,
- Audebert HJ,
- Amarenco P,
- Chumbler NR,
- Frankel MR,
- George MG,
- et al
- Garg V,
- Brewer J
- Birns J,
- Bhalla A,
- Rudd A
- Cho S,
- Khasanshina EV,
- Mathiassen L,
- Hess DC,
- Wang S,
- Stachura ME
US News and World Report. Available at: http://health.Usnews.Com/best-hospitals/rankings. Accessed February 15, 2010.
US Census Bureau. Available at: http://censtats.census.gov/usa/usa.shtml. Accessed February 10, 2010.
- Demaerschalk BM,
- Durocher DL
- Nelson RE,
- Saltzman GM,
- Skalabrin EJ,
- Demaerschalk BM,
- Majersik JJ
Report of the Ad Hoc Committee on Telemedicine. Federation of State Medical Boards of the United States. Dallas, TX: Federation of State Medical Boards of the United States. 1996.
Federal register, volume 76 issue 87 (Thursday, May 5, 2011). Available at: http://www.gpo.gov/fdsys/pkg/FR-2011-05-05/pdf/2011-10875.pdf. Accessed January 18, 2012.