The burden of stroke

Stroke is one of the leading causes of acute medical admissions to hospitals. It is among the most common causes of death and is a major cause of disability and poor health outcomes (Box 3.1). Worldwide, 17million people suffer a stroke each year and stroke is the third most common cause of death around the globe, accounting for 12% of all deaths, and exceeded only by heart disease and cancer (Reference FeiginFeigin et al., 2009; Reference ThriftThrift et al., 2014; GBD 2013 Mortality and Causes of Death Collaborators, 2015). While its management involves a series of specific responses by the hospital, the principles underlying them – including the importance of coordinated multispeciality and multiprofessional care, speed of response in the acute episode, the importance of prevention (of both the initial episode and any recurrence), and a model of care that follows the patient along the entire pathway, from the onset of illness to recovery and rehabilitation – apply equally to many other common medical conditions, such as acute myocardial infarction, gastrointestinal haemorrhage, or the acute and chronic complications of diabetes.

to 26 million (Reference FeiginFeigin et al., 2015). Essentially, changes in population demographics are outpacing improvements in stroke prevention, resulting in an increasing burden of stroke on populations and health systems, particularly in lower and middle income countries.

In addition to mortality, stroke causes a wide range of disabilities and impairments and has long-term implications for the health and well-being of survivors. These include neurological impairments such as muscle weakness or paralysis, impaired vision and impairments of speech and language skills. Up to 50% of patients will develop major depression in the years after stroke (Reference AyerbeAyerbe et al., 2013). Cognitive impairment is common and cerebrovascular disease is a major risk factor for dementia. More subtle cognitive problems, such as perceptual impairments, a change in personality, and profound fatigue are common, often lasting for years after the stroke (Reference WolfeWolfe et al., 2011). These problems are often referred to as “hidden deficits” but they account for a significant proportion of the suffering and costs that stroke causes. Effective risk reduction and high quality treatment should not only result in improvements in physical health but also reduce the future burden of dementia and mental health problems.

The financial costs of stroke are large and diverse: to the individual and their family in terms of health care and time off work; to governments in terms of medical and social care; and to wider society in terms of lost productivity. Other non-monetary costs are harder to calculate but are equally important – such as the emotional cost to family and friends of caring for a loved one who can no longer live independently.

Stroke accounts for between 2% and 4% of the total health care expenditure in developed countries. Moreover, stroke incurs substantial costs outside the health care system, reflecting survivors’ high rates of disability and dependence. In 2008 the total direct and indirect costs associated with stroke were approximately £8.9 (€9.7)billion per year in the United Kingdom (Reference SakaSaka et al., 2009). Most costs are incurred in the initial months and years after the patient has been discharged from hospital (Reference SakaSaka et al., 2009). Studies from Italy (Reference BottacchiBottachi et al., 2012), Denmark (Reference JennumJennum et al., 2015) and France (Reference SchmidtSchmidt et al., 2015) have produced similar estimates of the costs of stroke in Europe, at €7000–20000 per stroke. Almost any intervention that reduces the incidence of stroke or reduces the likelihood of long-term disability will be cost-effective in countries with expensive health and social care systems.

Age-adjusted incidence and mortality rates for stroke have fallen significantly over recent decades, thought to be due to improvements in stroke prevention through improved management of risk factors for stroke, especially hypertension and tobacco control, and, in some places, improved acute care. Between 1990 and 2013 the age-adjusted mortality rate for stroke in developed countries fell from 113 to 67 per 100000 (Reference FeiginFeigin et al., 2015). However, because of increasing longevity and the strong association between stroke risk and age, the absolute numbers of people having stroke are rising year on year, from an estimated 4.3 million globally in 1990 to 6.9million by 2013 (Reference FeiginFeigin et al., 2015). This is leading to increasing numbers of people dying from stroke (2.1 million in 1990 to 3.3million deaths in 2013 from ischaemic stroke), increasing disability adjusted life years and an almost doubling in the prevalence of stroke between 1990 and 2013, from 14million

Evidence-based stroke care

Historically, stroke was considered a condition for which little could be done, but there is now an extensive evidence base for interventions that are effective in improving outcomes after stroke: reducing disability, improving survival and reducing the risk of stroke recurrence. Some of these interventions (such as stroke unit based care) are applicable to almost all patients with stroke, while others are limited to selected patient groups (Figure 3.1).

Figure 3.1Number needed to treat for the main evidence-based interventions in acute ischaemic stroke

Source: Authors’ compilation

The stroke care pathway

Hospitals are central to stroke care but exist as components of a pathway of care that spans pre-hospital emergency medical services, acute hospital care, rehabilitation and primary care (Figure 3.2). For many of the elements of the patient pathway, there is now good quality evidence about how to organize health services to optimize patient outcomes. At the same time, we also know that in the real world there are wide variations both across Europe and within individual health economies in how stroke care is delivered.

Figure 3.2

An example of a typical stroke care pathway in high income settings

Source: Authors’ own

Workforce

Optimal stroke care is highly multidisciplinary, with a core stroke service requiring specialist doctors, nurses, physiotherapists, occupational therapists, speech and language therapists, dieticians and psychologists. Coordinating the work of the team is essential and formal MDT working (such as regular MDT meetings to discuss individual cases) is one of the components of stroke unit care (Reference Langhorne and PollockLanghorne, Pollock & Stroke Unit Triallists Collaboration, 2002).

Stroke medicine has traditionally not existed as a medical speciality in its own right, and as a result there is variation between countries in the specialty background of the lead physician. In most countries acute stroke care is largely provided by neurologists, but in some countries (such as the United Kingdom) stroke care is mainly led by stroke specialist physicians with a background in geriatric medicine.

Other medical specialties with important roles in the stroke pathway include neuroradiology (both diagnostic and interventional), neurosurgery, vascular surgery, intensive care, emergency medicine, rehabilitation medicine, and primary care.

Nursing care is an essential aspect of acute stroke care and it is likely that good quality nursing is one of the key mechanisms for the beneficial effect of stroke units (Reference Middleton, Grimley and AlexandrovMiddleton, Grimley & Alexandrov, 2015). In addition to general nursing skills, nurses need specific skills and knowledge in managing patients with stroke, such as screening and managing dysphagia, the positioning and mobilization of patients with muscle weakness or paralysis, prevention of pressure sores, and communicating with patients with language impairment after stroke (aphasia). Because many patients with stroke die as inpatients, nurses also need skills in providing end of life and palliative care. In some countries (such as the United Kingdom) nurses have taken on extended roles in prescribing, diagnostics, and assessing patients for thrombolysis.

In addition to the general evidence concerning nurse staffing levels and patient outcome (Reference NeedlemanNeedleman et al., 2011), there is specific evidence in stroke care that nursing-to-patient staffing ratios are associated with patient outcomes, with higher mortality rates for patients admitted at weekends to units with lower numbers of trained nursing staff (Reference BrayBray et al., 2014).

Rehabilitation is typically carried out by physiotherapists, occupational therapists, and speech and language therapists (speech pathologists). This includes carrying out assessments of the extent of a patient’s impairments and the impact of these on functioning, and planning treatment goals. Describing the full range of assessments and therapies provided by stroke therapists is beyond the scope of this chapter, but a wide range of methods may be used, from relatively simple mobilization techniques to more sophisticated interventions requiring the use of specialized equipment and aids. Therapists may also carry out additional diagnostic tests requiring additional skills and equipment. Therapists are central in planning patients’ discharge from hospital and implementing adaptations or the installation of equipment in patients’ homes, and have a key role in communicating with and providing psychological support for patients and their families and carers.

Stroke services also involve a variety of other allied health professionals. Dietetics is a core component of an acute stroke service, since many patients require nutritional support or assisted feeding. Problems with cognition, memory, mood or executive functioning are common after stroke and access to a clinical psychologist enables more detailed neuropsychological assessments to be carried out and appropriate information, support and therapy to be provided. Some patients with persistent physical or visual impairments may also require the provision of support aids from prosthetics and orthoptics specialists. As most patients will be discharged home on new or changed medications, pharmacists have an important role in ensuring safe prescribing, medicines reconciliation and in providing information to patients and family members about medications and side effects.

Networks of stroke care

In parallel with the development of organized stroke care in individual hospitals, many health systems have developed regional and network models of stroke care. In much of Europe and the USA a distinction is made between primary stroke centres and comprehensive stroke centres. Primary stroke centres are those with the necessary staffing, infrastructure and expertise to provide treatment for most stroke patients, but which may not have the capability to manage patients with more complex problems. Comprehensive stroke centres provide the same core stroke service but also the high technology and resource-intensive elements of care, such as interventional neuroradiology or neurosurgery, and play a greater role as centres for research and education. The European Stroke Organisation has produced guidelines setting out in detail the facilities and staffing required by comprehensive stroke centres (“ESO Stroke Centre”) in Europe (Ringelstein et al., 2013), which includes 24/7 provision of advanced imaging and interventional neuroradiology. In many countries these levels of care are formally accredited through certification schemes (for example, in the USA and Germany) or through quality registers (for example, in the United Kingdom and Sweden). There is evidence from the USA (Reference XianXian et al., 2011), Japan (Reference IiharaIihara et al., 2014), Finland (Reference MeretojaMeretoja et al., 2013) and the United Kingdom (Reference BrayBray et al., 2013a) that hospitals with higher levels of organized stroke care provide better outcomes for patients, suggesting that formal mechanisms to ensure stroke quality standards are important.

Networks of hospitals are frequently used in stroke care to provide access to the higher technology care offered in comprehensive stroke centres. These may act as the central referral centre for “hub and spoke” networks of hospitals, taking referrals from a number of primary stroke centres. Such networks have become increasingly important with the advent of more sophisticated diagnostic and interventional innovations, such as advanced brain imaging and thrombectomy, which would not be feasible or cost-effective to provide in smaller hospitals. In the United Kingdom the concept of the comprehensive stroke centre and primary stroke centre is more frequently defined in terms of hyperacute stroke units (HASUs) (providing acute care for the first 72 hours after stroke) and stroke units (for post-72 hour care).

There is evidence that these types of network can lead to better patient outcomes. In 2010 health care providers in London carried out a major reorganization of stroke services, reducing the number of acute admitting hospitals from 28 to 8 centres, each serving a population of approximately 1–1.5 million people. These eight hospitals were designated as HASUs and formal pre-hospital protocols were established so that all patients with suspected stroke would be transferred to a HASU. These HASUs provide acute care for up to 72 hours, and patients requiring ongoing inpatient treatment and rehabilitation are then transferred to a stroke unit closer to their home. The network is supported by agreed protocols for patient transfers, minimum standards for training, facilities and staffing, a common framework for payment and reimbursement from funders, and regular audit of quality and performance. Since these changes were established, there have been large improvements in the quality of stroke care in London, with stroke case fatality rates falling faster in London than in other urban areas in England (Reference MorrisMorris et al., 2014). One of the key aspects of stroke care in London that is different from many other “hub and spoke” models of care is the concept of providing higher level acute care to all patients and not just to selected patients; the majority of patients therefore have the opportunity to benefit from early intensive acute stroke care in an HASU.

Some models of care have emerged to help tackle the issue of providing specialist stroke care at scale by providing regional systems for transferring patients to specialist centres or providing specialist input remotely. Telemedicine is widely used in stroke care and many areas have implemented telemedicine systems to transmit video, audio and imaging data so that stroke specialists at home or working in another hospital can help in assessing patients presenting with acute stroke (Reference Hess and AudebertHess & Audebert, 2013). This has particular uses in delivering thrombolysis in rural areas where it may not be feasible to provide specialist stroke services in areas of low population density.

Telemedicine models may also be augmented by pathways that provide initial triage and assessment of patients in local hospitals, initiate thrombolysis if appropriate, and then transfer the patient to a hospital with specialist stroke care provision. These “drip and ship” models have been used particularly in the USA, where one in four patients treated with thrombolysis is now managed this way (Reference ShethSheth et al., 2015).

Variation in quality

Within Europe there is wide variation in the organization of stroke services and in the use of policies aimed at increasing care quality, such as clinical audit, financial incentives, clinical guidelines, accreditation and regulations (Reference Di CarloDi Carlo et al., 2015). The quality of care delivery across Europe is hard to measure consistently, since even when data are available, differences in the choice and definition of quality indicators make comparisons difficult (Reference WiedmannWiedmann et al., 2012). Nonetheless, wide variation exists even for aspects of stroke care with the strongest evidence base and most consistent inclusion in guidelines and audits, such as admission to a stroke unit or treatment with thrombolysis (Reference AyisAyis et al., 2013). For example, in 2011 only 33% of stroke patients in France were admitted to a stroke unit (Reference SchmidtSchmidt et al., 2015), compared with 62% in Scotland (Reference TurnerTurner et al., 2016). A survey of 25 European countries in 2005 found evidence of extremely wide variation in the provision of acute stroke care, with particularly poor provision of stroke unit care in Estonia, France, Greece and Portugal (Reference LeysLeys et al., 2007). Only 49% of the 886 hospitals included in the survey provided the minimum level of care to be considered a primary or comprehensive stroke centre. Stroke outcomes also vary significantly across Europe. For example, one comparative study of stroke outcomes between six European cities (in France, Italy, Lithuania, the United Kingdom, Spain and Poland) found three-fold variation in the risk of death after stroke (Reference HeuschmannHeuschmann et al., 2011).

Poor provision of acute stroke care occurs even in higher income European countries. For example, a survey of neurology services in Italian hospitals found that only 28% provided stroke unit based care, and large numbers of patients with stroke were admitted to hospitals without stroke units (Reference de Falco, Leone and Beghide Falco, Leone & Beghi, 2009). Organized stroke care was also relatively slow to develop in France, with only two hospitals out of 121 surveyed in 2005 providing stroke unit care (Reference LeysLeys et al., 2009). Policy-makers in France have subsequently prioritized stroke care and developed a national strategy for improvements, with a particular focus on developing stroke care networks (Reference LebrunLebrun et al., 2011).

Information on stroke care quality is collected systematically in some European countries (Reference WiedmannWiedmann et al., 2012), but most countries in Europe lack or have only fragmentary systems of data collection for quality improvement (Reference Di CarloDi Carlo et al., 2015). Where data are available, there is evidence of widespread variation in care quality within countries, not just between countries. For example, stroke care quality is measured in England and Wales by the Sentinel Stroke National Audit Programme (SSNAP) (Sentinel Stroke National Audit Programme, 2018) and there are wide geographical variations in a variety of care quality indicators (Figure 3.3). National clinical audits and registries in other countries show similar variation in stroke care quality, including Scotland (Scottish Stroke Care Audit; SSCA), Sweden (RiksStroke) and Germany (Reference WiedmannWiedmann et al., 2014). Some of these variations reflect broader geographical inequalities in the provision of health care services, such as the relative under-provision of stroke units in southern Italy compared to northern Italy (Reference GuidettiGuidetti et al., 2013).

Figure 3.3

Geographical variation in admission to a stroke unit within four hours of admission in England and Wales

Source: January–March 2015, SSNAP

Specialist resources

The main essential component of a stroke service is a stroke unit (Figure 3.4), and stroke unit beds should be available 24 hours a day for new admissions. Stroke units are more defined by staffing than by physical infrastructure, but stroke units do have a few specialist environmental and equipment considerations. Acute stroke units need to have the equipment to provide continuous (or regular) physiological monitoring. There is evidence that care bundles of nursing interventions that focus on physiological monitoring and detecting complications are effective in improving outcomes after stroke (Reference Middleton, Grimley and AlexandrovMiddleton, Grimley & Alexandrov, 2015). Appropriate facilities and space need to be available to mobilize patients and provide rehabilitation. This might include a gym for physiotherapy and areas (such as a therapy kitchen) for occupational therapy assessments.

Figure 3.4Specialist resources for a stroke service

Source: Authors’ compilation

Brain imaging is the core diagnostic requirement of a stroke service. CT imaging needs to be available 24 hours a day for the assessment of new patients and in carrying out imaging on patients who develop neurological deterioration after stroke. Although non-contrast CT imaging is adequate for most acute treatment decisions, increasing use is being made of more advanced CT imaging modalities (CT angiography and perfusion) and MRI.

A small proportion (less than 5%) of patients will require neurosurgical intervention (Reference VahediVahedi et al., 2007). Managing these patients requires access to neurosurgical infrastructure (theatres, specialist surgical, anaesthetic and nurse staffing, critical care facilities) either on site or after transfer to a referral centre. Similarly, some patients with stroke or TIA require vascular surgery for carotid endarterectomy. Patients with strokes that result in reduced consciousness may need to be managed in an intensive care unit as part of their admission. As thrombectomy services become more widely available, stroke care in high income countries will increasingly require access to neurointerventional facilities and the staffing required to provide them for patients suitable for this treatment.

Vascular ultrasound and echocardiography are also recommended as part of the diagnostic work-up for many patients (European Stroke Organisation Guidelines). All of these investigations require appropriate equipment and staff skilled in carrying out and interpreting these tests.

The requirements from stroke services of laboratory and pathology services are limited largely to common blood tests and in the diagnostic work-up of rarer causes of stroke. POCT is widely used in the emergency room for patients being assessed for thrombolysis, where rapid results are important in achieving fast door to needle times (Reference RizosRizos et al., 2009).

Stroke is not an area with a high use of specialist pharmaceuticals. The main pharmaceuticals that are required for an acute stroke service are alteplase (the only agent licensed for stroke thrombolysis in Europe) and secondary prevention agents (anticoagulants, antiplatelets, statins, anti-hypertensives).

Barriers to delivering optimal care

Training sufficient numbers of specialist staff has been a challenge in many countries. In the United Kingdom, for example, despite training pathways for physicians developing a specialty interest in stroke and the existence of stroke-specific professional organizations to support trainees and specialists, there is a shortage of physicians specialized in stroke medicine, with 25% of consultant posts remaining unfilled (Sentinel Stroke National Audit Programme, 2018). The reasons for this are unclear but might include the increasing requirement for out-of-office-hours working as stroke care has become higher in intensity. There are also fewer options for private practice for stroke specialist physicians than many other procedure-based or office-based specialties, which may affect its perceived attractiveness as a specialty. Lack of staffing resources, particularly of therapists and nursing home staff, was also identified as being one of the main barriers to improving stroke care in France (Reference GacheGache et al., 2014). There is potential for tackling staff shortages by expanding the roles and skills of existing clinical staff, such as empowering nurses to manage thrombolysis calls and take on leadership roles in stroke services.

One of the key barriers to providing optimum care has been the difficulty of closing the gap between evidence and widespread implementation into practice. Here the barriers are not merely lack of financial or other resources, but also contextual and behavioural factors such as culture, organization and leadership. Indeed, implementation gaps in stroke care involve not only the high technology and resource-intensive elements of stroke care, but also the key evidence-based components of care (such as stroke units and secondary prevention): even “getting the basics right” can be difficult. For example, even 20 years after the publication of the first trial to demonstrate the effectiveness of thrombolysis for ischaemic stroke, rates of use of thrombolysis vary widely even in highly developed health economies. In particular, the uptake of thrombolysis was initially very poor in the United Kingdom: when the National Audit Office reported on the quality of stroke care in England in 2005 it found that fewer than 1% of stroke patients were receiving thrombolysis. By contrast, during the same period 3–4% of stroke patients were treated with thrombolysis in Sweden (Reference ErikssonEriksson et al., 2010). The low rate of implementation in the United Kingdom occurred in the context of underdeveloped stroke services and highly variable care between centres, with only 60% of stroke patients being cared for in a stroke unit and many patients waiting more than two days for a brain scan (National Audit Office, 2005). This report prompted the development of a national improvement strategy, financial investment in stroke care, an expansion in training for stroke specialists and new resources allocated to quality improvement and audit. National clinical audits have since demonstrated significant improvements in the quality of stroke care in England and an acceleration in the uptake of thrombolysis, with 11–12% of patients (of all ages) now treated with thrombolysis: rates that are comparable with other high performing health systems in Europe (Sentinel Stroke National Audit Programme, 2018).

Although stroke care has been transformed by evidence-based medicine, there are still many areas of stroke care with little evidence to guide practice. One of the reasons for this may be relative underfunding of stroke research: in the United Kingdom for every £10 of health and social care costs attributable to stroke, it received only £0.19 in funding, compared to £1.08 for cancer and £0.65 for coronary heart disease (Reference Luengo-Fernandez, Leal and GrayLuengo-Fernandez, Leal & Gray, 2015). The areas of stroke care with the poorest evidence base are generally the less acute components of care, such as therapy and rehabilitation. Even fundamental questions about rehabilitation, such as when physiotherapy should commence after stroke, and at what intensity, are only now being addressed in randomized controlled trials (AVERT Trial Collaboration Group et al., 2015). Lack of evidence makes it difficult to define what optimal care in these areas should be, contributing to variations in practice. For example, there are wide variations across Europe in the amount of therapy provided to patients after stroke, which are not explained by differences in patient characteristics and likely reflect variation in access and availability (Reference WolfeWolfe et al., 2004; Reference WellwoodWellwood et al., 2009).

One of the biggest challenges for stroke medicine in high income countries in the next few years will be in implementing access to thrombectomy. Current provision is largely concentrated in relatively small numbers of specialist hospitals, and even in these hospitals there may not be round-the-clock access. The main barrier to implementation is insufficient numbers of trained neurointerventionists; increasing capacity will take time and there are likely to be resource challenges in maintaining a 24/7 acute thrombectomy service that may only be used relatively infrequently, with only a minority of acute stroke patients being appropriate for this treatment. Another risk is that a focus on developing thrombectomy services will distract attention and resources away from the wider challenge of implementing good quality stroke unit based care and post-stroke rehabilitation.

The future

The challenge for the future involves the twin tasks of implementing an ever-growing evidence base on new interventions and innovations and in improving the availability and quality of the elements of stroke care that we already know to work. As has already been described, wide variations in care quality exist both across and within European countries and these will not be reduced if the focus of clinicians, funders and managers is solely on implementing the “new”. Indeed, it is worth emphasizing that from a global perspective most people with stroke do not even receive the most core elements of stroke care such as stroke unit based care. By far the greatest reduction in the future burden of stroke on populations will come about not through new technologies but as a result of public health efforts to reduce stroke incidence through tobacco control, public health programmes to reduce cardiovascular risk factors (such as hypertension, obesity, alcohol and physical inactivity), increasing access to stroke unit based care and rehabilitation, and effective use of secondary prevention.

There are also examples of interventions that are still in use, despite evidence of ineffectiveness or even harm. One of the most prevalent of these is the use of antiplatelet agents in patients with AF. Historically, antiplatelet agents such as aspirin were often used as an alternative to anticoagulants to reduce the risk of stroke in people with AF (particularly in older people), but it is now known that antiplatelets provide much less benefit and are no safer than oral anticoagulants (Reference Aguilar, Hart and PearceAguilar, Hart & Pearce, 2007). Current guidelines therefore recommend that antiplatelet agents are not used for stroke prophylaxis in AF. However, many patients with AF are still prescribed antiplatelets, and oral anticoagulants remain underused. In England, for example, 31% of eligible patients known to be in AF in primary care were not prescribed an oral anticoagulant in 2013/2014 (NHS Quality and Outcome Framework, 2015), resulting in many thousands of avoidable strokes per year. Newer oral anticoagulants have become available in recent years that offer similar reductions in stroke risk but may have reduced risks of major complications than treatment with warfarin (Reference Gómez-OutesGómez-Outes et al., 2013).

As already discussed, the innovation most likely to change stroke care in the next five years in high income countries is thrombectomy for ischaemic stroke. The challenges of implementing this at scale, though, are significant and it is not certain how quickly this will become widely available experience from thrombolysis suggests that it is likely to be slow and highly variable between settings. There are other emerging areas of research that are still at early stages but may lead to significant impacts in the future. One of the most intriguing ideas is of reducing delays in thrombolysis by installing brain CT scanners in ambulances, allowing pre-hospital diagnosis of stroke type and administration of thrombolysis if appropriate. The concept has been demonstrated in a small number of centres (Reference WalterWalter et al., 2012; Reference ParkerParker et al., 2015), and although the real-world feasibility and cost-effectiveness of this model of care remain unproven, using new diagnostic technologies to facilitate pre-hospital stroke diagnosis could transform stroke care pathways. Imaging is an area of fast-moving innovation and development – for example, it is now possible to non-invasively image areas of unstable atherosclerotic plaque that are the source of the majority of strokes, and identify at an early stage the patients at highest risk of new or recurrent stroke (Reference Tarkin, Joshi and RuddTarkin, Joshi & Rudd, 2014). Further off, there is the prospect that stem cell technologies may allow the repair of established brain damage occurring as a result of stroke; early-stage clinical trials in stroke patients are already ongoing (Reference BanerjeeBanerjee et al., 2014). Rehabilitation is also increasingly making use of new advances in robotics to provide therapy and augment motor functioning in patients with limb paralysis after stroke (Reference BurgarBurgar et al., 2000).

Although exciting, most of these innovations are likely to be applicable only to a minority of stroke patients. The implementation of these new resource-intensive interventions therefore needs to be linked to efforts to develop models of delivery that can provide these in the most clinically and cost-effective way: for many aspects of acute stroke care this is likely to mean further development of networks of care and centralization of specialist services into hub hospitals.

Perhaps of greater medium-term significance to population health will be innovations that are applicable to all patients with stroke: the shift towards increased engagement of patients in managing their own health through shared decision-making and self-management, and in the increasingly sophisticated use of data to support research, quality improvement and new models of care. For example, there is good evidence that helping patients to manage their own blood pressure leads to better blood pressure control than the typical model of clinic-based management (Reference McManusMcManus et al., 2014); it is likely that health care services will increasingly take the role of supporting stroke survivors (and their carers) in managing stroke as a long-term condition. Similarly, health care in the future will make much more sophisticated use of real-world data such as electronic health records and clinical registries (Reference KrumholzKrumholz, 2014). For example, use of such data to generate more accurate predictions of prognosis, or to generate patient-specific estimates of the harms and benefits of interventions, can help in making better decisions about treatment and support patients in shared decision-making (Reference SpertusSpertus et al., 2015).

Stroke care in the hospital of the mid-21st century

Stroke care has changed dramatically over recent decades, driven by the development of organized multidisciplinary care and an increasing emphasis on acute intervention. The dependency on advanced medical imaging, resource-intensive multidisciplinary care and acute treatments that can only feasibly be administered in large hospitals means than hospitals are likely to remain the cornerstone of acute stroke care with most patients being admitted for inpatient care. The hospital of the future, if it is to provide comprehensive care for patients with stroke, will therefore need to be organized and designed to deliver:

• round-the-clock access to advanced imaging, diagnostics and neurointervention facilities that are geographically located within the hospital to optimize speed of access;

• stroke units to which patients with acute stroke are admitted without delay and which are the setting for multidisciplinary stroke specialist care;

• the appropriate environment and equipment to enable optimal provision of therapy and to support rehabilitation and recovery; and

• organized pathways of care that reduce treatment delays and support the provision of good quality therapy not just in hospital but also in the community.

In many places this means that some acute hospitals should no longer attempt to treat stroke. Rather, there is a need to find alternative models of care whereby those suffering a stroke will be taken, at least for definitive treatment, to a hospital that can provide a comprehensive care package, including rapid diagnosis and intervention where required. This will often not be the nearest facility. This could have profound implications for the organization of hospitals in a defined area, especially where they have had a high degree of autonomy. It will often be extremely challenging, politically and legally, to tackle this and each solution must be tailored to the particular context.

The critical component of stroke care services will remain not physical assets and medical devices but the MDTs of people that are the core of organized stroke unit care. Maintaining and developing this resource will require long-term investment in the training of the stroke workforce (medical, nursing and allied health professions). Providing ongoing education and training will be of increasing importance in helping clinicians keep up to date with the accelerating pace of new medical knowledge and evidence.

It will be disappointing if the next few decades do not see the development of new, high-impact drugs and devices that improve recovery after stroke, reduce complications, or help survivors manage the long-term consequences of stroke. The development of therapies that facilitate brain repair (for example, through stem cells) could be a real paradigm shift, but the brain is vastly complex and still contains many mysteries; progress in the development of new “brain regeneration” therapies is hard to predict. For patients with permanent impairments after stroke, assistive technologies (such as robotics) are likely to become much more mainstream and sophisticated, and allow more stroke survivors to live independent lives. The challenge for the future will be providing these innovations at scale in a cost-effective way and in speeding up the diffusion of new evidence into widespread clinical practice.

The evidence of current variation in care quality and outcome points to the importance of prioritizing and developing quality improvement in stroke care. This includes supporting and developing current systems of clinical audit (SSNAP, SSCA, Riks-Stroke, Danish Stroke Register) and increasing the capacity of health care systems to deliver continuous quality improvement. As the sophistication and scope of health care data collection increase, this is likely to lead to a growing emphasis on the development of new ways of using data in stroke care as part of clinical care, research and quality improvement. It is hard to foresee in much detail what this new, data-aware world of health care will look like in practice, but it may have a transformative effect on the delivery and organization of stroke care in the next few decades.