Current Practices in Stroke Systems of Care: An International Survey of Progressive Stroke System Processes

Introduction

Stroke remains the 5^th leading cause of death in the United States, with increasing prevalence, and 2^nd leading cause of death globally.^1,2 Recent Centers for Disease Control (CDC) data also indicates that despite nearly four decades of decline in stroke death rates, this decline has begun to stagnate and even reverse in many “Stroke Belt” states in the Southeast US.^2,3 Further, recent advances in endovascular therapy (EVT) for acute ischemic stroke caused by large vessel occlusion (LVO)^4–10 have magnified the need for rapid access to advanced specialty providers and therapeutics, which are not widely available. In light of this rapid paradigm shift in acute stroke care, health systems must quickly evolve. This will require large-scale regionalization approaches. Recently published American Heart Association/American Stroke Association (AHA/ASA) guidelines for the treatment of acute ischemic stroke recognized this with a high level recommendation (1A) that “regional systems of stroke care should be developed”.¹¹

In response to these needs, especially in a region of the United States with the highest rates of stroke mortality³, we developed a comprehensive regional approach to improve stroke care through the application of best clinical practices utilizing a systems of care model quality improvement program known as Implementation of Best Practices for Acute Stroke care – Developing and optimizing regional systems in stroke Care (IMPROVE Stroke Care). IMPROVE Stroke Care utilizes implementation science approaches that have previously been demonstrated to be highly effective in other time-sensitive vascular emergencies such as ST elevation myocardial infarction (STEMI) and cardiac arrest.^12–14 The primary vision for the IMPROVE Stroke Care project is to reduce disability and mortality from acute stroke in the region by identifying optimal processes of stroke care delivery through sharing data on performance and best practice, allowing redesign of systems within the partnered networks, leveraging existing local hub-and-spoke model hospital systems. The methods for this program have been previously published.¹⁵

Aims

The first phase in the development of the IMPROVE Stroke Care program was to evaluate the current state of advanced stroke systems worldwide, and identify best practice processes to inform and provide evidence for recommendations included within the program’s Manual of Operations. High-level clinical guideline statements exist from organizations such as the AHA/ASA and the European Stroke Organization (ESO). Additional practice guidance can be found in large clinical trials, and big-data reviews from large data repositories such as Get With The Guidelines – Stroke. However, cutting-edge approaches are often not available from these sources, simply because they are too new to have big-data backing, or large clinical trials to support it, which take considerable time and funding. Additionally, expert consensus opinion level data on current practice controversies and innovative science is often lacking in published literature, or difficult to locate and summarize. We developed this international survey of high-performing and progressive stroke centers to obtain this data and inform advanced practice recommendations for the IMPROVE Stroke Care Program Manual of Operations.

Methods

An 81-question survey was developed with input from team members with backgrounds in stroke, neurocritical care, emergency medicine, health services research, quality improvement and nursing. Questions in the survey tool included center demographics, patient volumes, staffing protocols, acute stroke code processes, Endovascular therapy (EVT) processes, Emergency Medical Systems (EMS), stroke IT and data collection, telestroke use, patient transfer processes, post-acute patient management and system feedback methodology. Data was collected from January 2017 to May 2017. The entire survey tool can be found in appendix 1. Twenty advanced stroke systems worldwide were identified based on publications, conference presentations from investigators at these institutions, as well as colleague input, representing a convenience sample. Physician leaders of their respective stroke programs at those institutions were contacted via email and invited to participate in the study. The survey and database were built, and study data were collected and managed, using RedCap¹⁶ electronic data capture tools hosted at Duke University. Statistical analysis was conducted using RedCap internal statistical tools. Basic counts and percentages, averages (mean or median as appropriate) and range were calculated as appropriate for individual survey element responses.

All study procedures were deemed exempt from review by the local Institutional Review Board. IMPROVE Stroke Care is funded by grants from the Medtronic Foundation, Daiichi Sankyo, and Chiesi. The funding sources had no involvement in the design or conduct of the study, data management or analysis, manuscript preparation or authorization for submission.

Results

Center demographics

Of 20 invited stroke centers, 9 responded with complete surveys (45%) (Figure 1). Seven of these centers self-identified as “Comprehensive Stroke Center” as defined by the US Joint Commission of Hospital Certification, or equivalent certifying agencies in Canada and Europe; 2 identified as “Intervention capable”, defined as primary stroke center with endovascular capability but without the official credentials of “comprehensive stroke center”. The survey pre-dated official separate certification by The Joint Commission for Thrombectomy-Capable Stroke Centers. On average (median) the 9 responding centers cared for 750 ischemic stroke patients per year (range 350-1444) and 125 primary intracerebral hemorrhagic strokes per year (range 100-501); in total, 81.4% percent ischemic and 18.6% primary intracerebral hemorrhage. Additionally, these centers reported a median 80 (range 47-177) primary subarachnoid hemorrhages (SAH) cared for yearly. Participating centers reported a median of 150 acute ischemic stroke patients treated with IV alteplase (tissue plasminogen activator -tPA) (range 60-440) and an average of 86 endovascular thrombectomy procedures per year (12-130).

• Download figure
• Open in new tab

Figure 1. Map of participating centers

Discussion

We developed this international survey of advanced and progressive stroke centers in order to obtain expert opinion-level data in acute stroke care, and inform advanced practice recommendations for the IMPROVE Stroke Care Program Manual of Operations.

The survey results showed several areas of current controversy are managed variably even in these advanced centers, though trends can be identified. For example, slightly over half of responding centers indicated that transferred-in LVO patients bypass their ED and are transported directly into the neurointerventional suite. Additionally, in large systems, destination protocols to transport patients with suspected LVO directly to thrombectomy-capable centers was routinely practiced in only about half of the responding centers. Only 4 of the 9 responding centers noted the routine use of perfusion imaging modalities for patient selection for intervention; notably, this survey was completed before the release of the DAWN and DEFUSE 3 trial data. Similarly, the level of admission (ICU, step-down unit, or ‘floor’ acute care unit) was split amongst the responding centers, which may indicate the location is less important than the processes and practice in those locations.

Other practices that were expected to be found at some progressive centers were in fact not in use at all, such as the utilization of pre-hospital mobile apps for tracking and communicating stroke patient data. Similarly, some practices that were expected to be universal, were not. The majority, but not all responding centers, gave quick feedback to their respective EMS, ED, and provider teams regarding stroke patient processes and outcomes. The majority, but not all centers, also participate in large-scale data repositories such as Get With The Guidelines or equivalent.

There were several limitations to this study. The response rate was 45%, however as the design of this study was not intended to be a representative sample of hospitals, and was essentially a convenience sampling methodology, it is unlikely this introduced significant bias. Likewise, non-responding centers were fairly evenly distributed across the globe, including 3 European, 1 Canadian and 6 centers in the United States (3 Midwest, 4 ‘stroke belt’/Southeast). More Southeast US sites were invited to participate than other locations. Further, language barrier may have contributed to non-response from some European sites. No Asian sites participated. Another limitation is lack of available outcome data from all participating centers. While this information could be used to more objectively define high-performing centers, it is rarely widely available without pre-existing and time consuming data use agreements. Further, we are not able to verify any responses.

Additionally, some practices varied significantly between North America and Europe. Namely, transportation of patients to specific centers is managed differently in compact and high-density European city centers. For example, in London, all acute stroke patients are transported directly to the advanced centers regardless of suspected LVO status. As another example, in many UK centers, stroke patients are managed by ‘stroke physicians’ who may not be neurology trained and boarded, which affected responses to staffing survey questions.

Overall, this international survey of advanced and progressive stroke systems revealed useful practice patterns and processes, but also highlights that a one-size-fits-all approach is unlikely to be useful. Instead, we suggest regional collaboration to evaluate area resources, geography and hospital capabilities, to develop local best-practices. Processes from other high-functioning stroke systems such as those outlined in our survey should be used as a guide or menu of possible approaches that have a documented history of success in specific settings. The findings of this survey were used to supplement AHA/ASA guideline statements as well as results of big-data and large clinical trial publications in the development of the IMPROVE Stroke Care Manual of Operations recommendations. Practices that were common among the advanced systems surveyed are listed as lower level-of-evidence recommendations for potential changes in practice in local stroke systems when resources or improvement goals align. Controversial practices remain differentially managed by different centers and should be considered in the context of a particular hospital’s resources and overall system of care.

Disclosures

ME Ehrlich: Dr. Ehrlich discloses salary support from grant funding by The Medtronic Foundation, and Daiichi Sankyo.

S Shah: Dr. Shah discloses salary support from grant funding by The Medtronic Foundation, and Daiichi Sankyo.

BJ Kolls: Dr. Kolls discloses salary support from grant funding by The Medtronic Foundation.

M Roettig: no disclosures

L Monk: no disclosures

C Graffagnino: Dr. Graffagnino discloses salary support from grant funding by Medtronic Foundation; medical consultancy and clinical trial support from Daichi Sankyo; medical consultancy for Portola and research funding from Chiesi.