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

National Collaborating Centre for Chronic Conditions (UK). Stroke: National Clinical Guideline for Diagnosis and Initial Management of Acute Stroke and Transient Ischaemic Attack (TIA). London: Royal College of Physicians (UK); 2008. (NICE Clinical Guidelines, No. 68.)

  • In May 2019, NICE updated and replaced this guideline with NICE guideline NG128 on stroke and transient ischaemic attack in over 16s. Some of the 2008 recommendations have been retained in the new guideline. This 2008 full guideline includes the evidence supporting the 2008 recommendations. Sections of the guideline CG68 that have been updated are shaded in grey in the PDF.

In May 2019, NICE updated and replaced this guideline with NICE guideline NG128 on stroke and transient ischaemic attack in over 16s. Some of the 2008 recommendations have been retained in the new guideline. This 2008 full guideline includes the evidence supporting the 2008 recommendations. Sections of the guideline CG68 that have been updated are shaded in grey in the PDF.

Cover of Stroke

Stroke: National Clinical Guideline for Diagnosis and Initial Management of Acute Stroke and Transient Ischaemic Attack (TIA).

Show details

5The rapid recognition of symptoms and diagnosis

5.1. Pre-hospital prompt recognition of symptoms of TIA and stroke symptoms

5.1.1. Clinical introduction

People who present with acute stroke or TIA need immediate clinical assessment and treatment. Few people have much awareness of the symptoms of stroke, and may delay seeking help as a result; hence the need for the UK Stroke Association’s Act FAST campaign.9 A number of tools have been designed to help paramedics and other healthcare professionals recognise symptoms in the community. Other tools have been developed to improve the speed of diagnosis on arrival in the A&E department to avoid delay in the delivery of specialist assessment and management. It should be noted that some strokes (e.g. those affecting purely balance or cognition) may not be picked up by clinical assessment tools.

The clinical question addressed is whether emergency health professionals are able to use a clinical assessment tool to accurately identify those patients who have had a suspected stroke or TIA.

5.1.2. Clinical methodological introduction

A number of different pre-hospital ‘assessment tools’ were identified for use by paramedics, namely the Face Arm Speech Test (FAST),10,11 Los Angeles Prehospital Stroke Screen (LAPSS),12 Cincinnati Prehospital Stroke Scale (CPSS)13 and Melbourne Ambulance Stroke Screen (MASS).14,15 In addition one assessment tool, Recognition of Stroke in the Emergency Room (ROSIER)16 has been developed for use by ER (A&E department) hospital personnel. Level 2+

Studies varied considerably with respect to patient selection, setting (e.g. hospital versus field study) and outcomes. Table 5.1 below outlines the different pre-admission and emergency assessment tools.

Table 5.1. Pre-admission and emergency assessment tools.

Table 5.1

Pre-admission and emergency assessment tools.

5.1.3. Health economic methodological introduction

No studies were identified.

5.1.4. Clinical evidence statements

1.0. Face Arm Speech Test (FAST)

Two studies evaluated the diagnostic accuracy of FAST by paramedics.10,11 Level 1b+

One study prospectively compared the characteristics and the accuracy of referrals to the acute stroke unit from ambulance staff using the FAST instrument with those of primary care doctors (PCDs) and ER doctors.10 Level 1b+

The positive predictive value (PPV) for ambulance staff was 78% (95%CI 72 to 84%). A stroke/TIA detection rate (diagnostic accuracy) was estimated for the ambulance paramedics by assuming all strokes/TIAs that were taken by ambulance to the ER were referred to the acute stroke unit. This gave an upper estimation of sensitivity of 79%.*10 There were no significant differences between the ambulance paramedics, PCDs and ER personnel on the number of non-stroke cases referred to the stroke service (NS). Overall, the ambulance paramedic diagnosis of stroke was as accurate as that of PCDs or ER doctors, although the strokes they admitted tended to be more severe and may have therefore been easier to diagnose.10 Level 1b+

A further ‘in the field’ study reported acceptable interobserver agreement between neurological signs recorded in the FAST by paramedics and stroke physicians after admission.11 Level 1b+

1.1. The Cincinnati Prehospital Stroke Scale (CPSS)

This study prospectively validated the CPSS used by pre-hospital care providers (paramedics and emergency medical technicians (EMTs)).17 Level 2+

There was high reproducibility among pre-hospital care providers for total score and for each scale item (arm weakness, speech and facial droop). There was a high correlation between the physicians’ total scores and the pre-hospital providers. Agreement on scoring on specific items between physicians and pre-hospital personnel was high for all three items. A single abnormality on the CPSS had a sensitivity of 66% and a specificity of 87% in identifying a patient with stroke when scored by a physician and 59% and 89% respectively, when scored by a pre-hospital provider.17 Level 2+

1.2. Los Angeles Prehospital Stroke Screen (LAPSS)

The diagnostic accuracy of paramedics using LAPSS in the field was compared with that of emergency department and final hospital discharge diagnoses.12 Level 2+

In patients with completed LAPSS forms (corrected for documentation error) the sensitivity was 91% (95%CI 76 to 98%); specificity 97% (93 to 99%); PPV 97% (84 to 99%); and negative predictive value (NPV) 98% (95 to 99%).12 Level 2+

1.3. Melbourne Ambulance Stroke Screen (MASS)

Two studies were identified for this assessment scale.14,15 Level 2+

One study reported that accuracy of paramedics at identifying stroke significantly improved after stroke education and training on the use of the MASS tool from 78% (95%CI 63 to 88%) to 94% (95%CI 86 to 98%). For the MASS paramedics the sensitivity of stroke diagnosis was significantly greater when the MASS tool was used compared with strokes for which there was no documented assessment (95 vs 70%).14 Level 2+

Another study performed an in-field validation of the MASS (N=100 assessments). The MASS showed equivalent levels of sensitivity compared to the CPSS (NS) but was significantly superior to that of the LAPSS (90 vs 78% (95%CI 67 to 87)). The specificity of the MASS was equivalent to that of the LAPSS (NS) but was significantly superior to that of the CPSS (74 vs 56% (36 to 74)).15 The PPV of MASS, LAPSS and CPSS were 90 (95%CI 81 to 96), 93 (95%CI 83 to 98) and 85 (95%CI 75 to 92) per cent respectively, the NPV for MASS, LAPSS and CPSS were 74 (95%CI 53 to 88), 59 (95%CI 42 to 74) and 79 (95%CI 54 to 93) per cent respectively. Level 2+

1.4. Recognition of Stroke in the Emergency Room (ROSIER)

One study prospectively validated the ROSIER in stroke/TIA patients used by ER physicians.16 Level Ib+

For a stroke cut-off rating of 1+ or above, the ROSIER scale had a sensitivity of 93%. The ROSIER scale incorrectly diagnosed 17/160 (10%; 10 false positive, 7 false negative). The diagnostic performance of ROSIER compared with CPSS, FAST and LAPSS in the patients described in this study is presented in table 5.2 below. FAST scores were completed for 49 of 91 (54%) stroke patients taken to ER by ambulance paramedics. For these patients, ROSIER was superior to FAST (sensitivity 92 vs 54%, specificity 96 vs 91%, PPV 96 vs 88%, NPV 92 vs 64%).16 Level 1b+

Table 5.2. The diagnostic performance of ROSIER compared to CPSS, FAST and LAPSS.

Table 5.2

The diagnostic performance of ROSIER compared to CPSS, FAST and LAPSS.

5.1.5. From evidence to recommendations

The use of a validated tool to identify the symptoms and signs of suspected stroke and TIA increases diagnostic accuracy. The GDG noted that a study of ambulance paramedics, A&E and PCDs using the FAST assessment demonstrated a high PPV for accurate diagnosis of stroke and TIA. Immediate diagnosis improves the speed of access to specialist care.

The GDG noted that one study using the MASS validated tool compared the diagnostic accuracy of untrained versus trained paramedics and found that training improved the accuracy of diagnosis.

The ROSIER assessment is validated for use in A&E. It is more detailed than the FAST assessment (it includes blood sugar, visual field assessment and documentation of a history of seizures or loss of consciousness). It is more accurate than CPSS, FAST or LAPSS. The GDG reviewed the evidence and concurred that while a pre-hospital assessment including blood glucose is essential to ensure rapid admission to specialist care, a more detailed assessment tool is required in A&E to exclude common stroke mimics such as hypoglycaemia or Todd’s paresis.

5.1.6. RECOMMENDATIONS

R1.

In people with sudden onset of neurological symptoms a validated tool, such as Face Arm Speech Test (FAST), should be used outside hospital to screen for a diagnosis of stroke or TIA.

R2.

In people with sudden onset of neurological symptoms, hypoglycaemia should be excluded as the cause of these symptoms.

R3.

People who are admitted to accident & emergency (A&E) with a suspected stroke or TIA should have the diagnosis established rapidly using a validated tool, such as Recognition of Stroke in the Emergency Room (ROSIER).

5.2. Early versus late assessment of people with TIA, and identifying those at high risk of stroke

5.2.1. Clinical introduction

Patients with transient neurological symptoms may underestimate their significance. They delay seeking specialist care or may wait days to see a general practitioner (GP). The Intercollegiate Working Party Guidelines in 200118 set a standard for a time to specialist assessment in a rapid access TIA clinic of 14 days, a target that was widely seen at the time as difficult to achieve. By 2004, this target was 1 week. The National Sentinel Audit in 20063 showed that while 78% of Trusts had a designated neurovascular clinic, the average waiting time for a clinic appointment remained high at 12 days (IQR 7–17). Recent data from the Oxford Vascular Study (OXVASC) demonstrate that some patients are at high risk from completed stroke long before this time.

A systematic review of the risk of stroke within 7 days of TIA identified 18 independent cohorts (N=10,126 patients). The outcomes of 15 people were reported at 2 days after TIA and 17 at 7 days. The pooled risk of stroke at 2 days was 3.1% (95%CI 2.0 to 4.1) and at 7 days 5.2% (3.9 to 6.5). Significant heterogeneity was reported between the studies (p<0.0001) reflecting the different study methodologies and clinical characteristics of the patient population.19

Simple clinical scoring systems can identify patients at particularly high risk who require immediate investigation and management. Specialist assessment involves confirmation of the diagnosis of TIA (around 40–50% of all TIA clinic referrals may, after specialist assessment, be diagnosed as non-neurovascular) and its vascular territory, appropriate investigations (including brain and carotid imaging), and assessment and management of vascular risk factors. A number of models of specialist assessment have been developed including ‘rapid access’ TIA clinics, daily in some cases, a 24-hour clinic, and day-case admission to hospital.

The clinical question addressed is whether scoring systems can accurately predict those patients with suspected TIA who need urgent referral for specialist assessment, and whether this early (immediate) assessment improves outcome.

5.2.2. Clinical methodological introduction

Early vs late assessment

One prospective cohort study (EXPRESS)20 and one observational study (SOS-TIA)21 were identified that looked at outcomes in patients with TIA who had undergone urgent assessment and treatment.

One study was a prospective population-based sequential comparison study (EXPRESS) of patients referred to an appointment-based TIA clinic that made treatment recommendations to primary care physicians (April 2002 to September 2004, phase one) (N=310) with a TIA clinic that did not require an appointment and at which treatment was initiated immediately if the diagnosis was confirmed (October 2004 onwards, phase two) (N=281). The mode of access and time of treatment initiation changed but the referral criteria remained consistent throughout. Treatment initiated in the second phase included aspirin 300 mg taken in the clinic, together with a 4-week prescription for any other medication prescribed by the clinic. In addition, clopidogrel 300 mg loading dose was given to all patients initiated on aspirin. In contrast, in phase one, primary care physicians were generally recommended to prescribe aspirin or clopidogrel if the former was contraindicated.20 Level 2++

A prospective observational study (SOS-TIA) (N=1,085) evaluated the impact of a 24-hour rapid assessment clinic for patients with suspected TIA. Clinical assessment occurred within 4 hours of admission. Patients with minor stroke, definite or possible TIA were prescribed antithrombotic medication immediately. The study reported on the risk of recurrent stroke in patients treated in a rapid access clinic compared with that predicted on the basis of signs and symptoms of admission (ABCD2).21 Level 3

Scoring systems

Five studies were identified, all evaluated the accuracy of the one scoring system, namely the ABCD or a derivation of this (ABCD2 score), to predict early stroke risk after TIA.22–25 One study was excluded due to methodological limitations.26

5.2.3. Health economic methodological introduction

No papers were identified. An original cost-effectiveness analysis was conducted for this guideline. Detailed methods are presented in Appendix C, available online at www.rcplondon.ac.uk/pubs/brochure.aspx?e=250.

Aim of the cost-effectiveness model

To evaluate the relative cost effectiveness of assessing patients identified by a GP with a suspected TIA:

For comparison, we also included a strategy of follow-up by the GP without referral for specialist assessment.

We assessed cost effectiveness of each strategy for all patients combined but also compared the strategies for each ABCD2 score group.

General methods

The cost effectiveness of the different strategies was estimated using a simple decision analysis. The NICE reference case was followed. For example:

  • Costs are measured from the perspective of the NHS and personal social services (PSS) including the long-term care costs for stroke patients.
  • Health outcome is measured from the perspective of the patient (not carer or family members).
  • Health outcome is measured in terms of quality adjusted life years (QALYs).
  • A 3.5% discount rate was applied to both costs and effects.

Where appropriate, we have used data and assumptions from the HTA report on the effectiveness and cost effectiveness of carotid artery assessment by Wardlaw et al.27

The model

The decision model sought to capture the following effects:

  • Patients seen at a specialist clinic are more likely to be given appropriate medication and therefore will have strokes averted (in the first 90 days).
  • Patients seen immediately will receive this medication sooner and therefore will have more strokes averted than those seen at weekly clinics.
  • Patients seen at a specialist clinic will receive carotid artery ultrasound imaging (and subsequent carotid endarterectomy if stenosis >=50%), which will reduce the incidence of stroke (over 5 years). Whereas patients followed up by their GP do not receive imaging or surgery.
  • Patients seen at a specialist clinic immediately will be more likely to receive endarterectomy within 2 weeks, when it is more effective, compared with patients who are seen at a weekly clinic. Furthermore, more patients will have a stroke before they have surgery.
  • Carotid artery imaging is not perfectly accurate.
  • Endarterectomy confers a risk of death in the short term.
  • Specialist clinics are more costly than GP assessment. Costs of drugs over the lifetime will be increased. But these costs will be at least partly offset by cost savings from reduced stroke treatment over the lifetime.

The effect of different treatment strategies is first modelled in terms of effect on stroke incidence. Patients are then divided into whether or not the stroke was fatal and whether or not the stroke left them dependent. Long-term quality adjusted life expectancy was estimated for each group and for the patients who do not experience a stroke. Similarly, lifetime healthcare costs are measured for each stroke outcome.

Patients in lower ABCD2 score groups have a lower baseline risk of stroke and therefore have fewer strokes averted compared with patients in higher ABCD2 score groups.

5.2.4. Clinical evidence statements

Early vs late assessment

The EXPRESS study reported a highly significant reduction in the 90-day recurrent stroke rate in phase two compared with phase one for patients referred to the study clinic (12.4 vs 4.4%; p<0.0015).20 Level 2++

At one month a significantly higher proportion of patients referred to the study clinic in phase two compared with phase one were prescribed antiplatelet agents or anticoagulant therapy, aspirin and a 30-day course of clopidogrel, one or more blood pressure lowering drugs. Similarly, a significantly higher proportion were referred to carotid surgery within 7 days or less or 30 days or less.20 Level 2++

There was no significant difference in the delay from the presenting event to seeking medical attention in patients subsequently referred to the study clinic between the two study phases (NS). However, there was a significantly longer delay in seeking medical attention from primary care to assessment in clinic in phase one (median 3 days) compared with phase two (median less than 1 day). A significantly higher proportion of patients were seen within 6 hours or less from first call to medical attention to assessment in the study clinic in phase two than in phase one. Consequently, there were significantly fewer recurrent strokes after presentation to primary care but before assessment in clinic in phase two than in phase one. Median time from seeking medical attention to first prescription of one or the other treatments recommended in the faxed letter from the study clinic to primary care was significantly longer in phase one than in phase two (20 vs 1 day).20 Level 2++

The 90-day risk of recurrent stroke in referrals to the study clinic was significantly greater in phase one than in phase two for patients presenting with TIA, stroke, both men and women and for all age groups (statistical analysis not reported).20 Level 2++

Early treatment (phase two) did not increase the 30-day risk of bleeding events requiring medical attention (NS). No symptomatic intracerebral or other intracerebral haemorrhages were identified in either phase of the study and there was no symptomatic haemorrhagic transformation of infarction (NS).20 Level 2++

The prospective observational study evaluating the impact of a 24-hour rapid assessment clinic for patients with suspected TIA reported:21

  • 701/1085 (65%) patients assessed in the clinic had confirmed TIA or minor stroke. 277/1085 (26%) of patients were admitted to a stroke unit and remainder were discharged home on the day of the examination at the clinic.
  • 824/845 (98%) patients with minor stroke, definite or probable TIA received antithrombotic medication immediately.
  • 44/51 (86%) patients with atrial fibrillation (AF) and definite TIA were prescribed oral anticoagulants.
  • 129 (24%) patients with definite TIA and no evidence of brain tissue damage, and 46 (43%) patients with definite TIA and evidence of brain tissue damage, were started on medication to lower blood pressure or had their medication modified. Level 3

Table 5.3 below reports the 90-day stroke risk (95%CI) recorded in patients attending the rapid assessment clinic compared with that expected on the basis of the ABCD2 score.21

Table 5.3. 90-day stroke risk recorded in patients attending the rapid assessment clinic compared with that expected on the basis of the ABCD2 score.

Table 5.3

90-day stroke risk recorded in patients attending the rapid assessment clinic compared with that expected on the basis of the ABCD2 score.

For all patients seen within 24 hours of symptom onset in the rapid assessment clinic (N=552), the actual 90-day stroke rate was 1.63% (95%CI 0.85 to 3.12) vs 6.49%.21 Level 3

Scoring systems

The ABCD score was derived from the OXVASC study24 where a series of clinical features in people with TIA were related to subsequent stroke risk: age (<60 years=0, ≥60=1); BP (systolic ≤140 mmHg and/or diastolic >90 mmHg=0, systolic >140 mmHg and/or diastolic >90 mmHg =1); clinical features (unilateral weakness=2, speech disturbance without weakness=1, other symptom=0); duration of symptoms (<10 mins=0, 10 to 59 mins=1, ≥60 mins=2). The ABCD score aims to identify individuals at high risk of stroke and who may require emergency intervention.

Due to the different study populations and outcomes, the results of each study are presented separately.

1.0. Rothwell et al. (2005)24

One study derived a score for the 7-day risk of stroke based on a population of patients with probable or definite TIA (N=209). The score was then validated in a similar population cohort (OXVASC, N=190).** The clinical usefulness of the score to ‘front-line’ heath professionals was assessed by using it to stratify all patients with suspected TIA referred to OXVASC (N=378) and to a hospital-based weekly TIA clinic (N=210).24 Level 3

The 7-day risk of stroke for the OXVASC cohort was 5.3% (3.0 to 7.5) and for the hospital-based cohort 5.2% (2.2 to 8.3). In the OXVASC (N=377) population-based cohort 19/377 referrals, 19 (95%) of the strokes that occurred within 7 days of presenting TIA occurred in the 101 (27%) patients with a risk score of 5 or greater. The score was still significantly predictive when the five 7-day strokes that occurred before the patient sought medical attention after the initial TIA were excluded.24 Level 3

In the non-OXVASC hospital-referred TIA clinic (N=206) the median (IQR) time to referral to clinic and the appointment was 9 (4 to 16 days), with 42% seen within the 7 days of referral. Fourteen (7.5%) patients had a stroke before their scheduled clinical appointment. The ABCD score was a significant predictor of stroke before the clinical appointment with no events in patients with a score of less than 4.24 Level 3

2.0. Tsivgoulis et al. (2006)25

This study validated the ABCD score retrospectively by reviewing the emergency room and hospital records of patients with definite TIA (N=226). These patients were followed up prospectively for 1 month to derive a 30-day risk of stroke.25

The 30-day risk of stroke was 9.7% (95%CI 5.8 to 13.6%). The ABCD score was highly predictive of both 7- and 30-day risks of stroke. The multivariate Cox regression analyses revealed that an ABCD score of 5 to 6 was an independent predictor of the 30-day stroke risk. More specifically, an ABCD score of 5 to 6 at the ED was associated with an 8-fold greater 30-day risk of stroke. Furthermore, an ABCD score of 5 to 6 was also independently significantly associated with the 7-day risk of stroke.25 Level 3

Table 5.4. The 7-day risk of stroke stratified according to ABCD score at the first assessment in all referrals with suspected TIA to OXVASC and risk of stroke before the scheduled clinic appointment in all referrals with suspected TIA to the non-OXVASC hospital-referred weekly clinic. Level 3.

Table 5.4

The 7-day risk of stroke stratified according to ABCD score at the first assessment in all referrals with suspected TIA to OXVASC and risk of stroke before the scheduled clinic appointment in all referrals with suspected TIA to the non-OXVASC hospital-referred (more...)

Table 5.5. The ABCD score predictive value of the risk of stroke at both 7 and 30 days.

Table 5.5

The ABCD score predictive value of the risk of stroke at both 7 and 30 days.

Table 5.6. The proportions of strokes occurring by 7 and 90 days stratified by the ABCD score on admission.

Table 5.6

The proportions of strokes occurring by 7 and 90 days stratified by the ABCD score on admission.

3.0. Bray et al. (2007)22

A retrospective study (N=98) evaluated the accuracy of a dichotomised ABCD to predict stroke at 7 and 90 days in patients with TIA presenting to an emergency department.22 Level 3

Dichotomising the ABCD score (4 or less vs ≥5) categorised 48 (49%) of patients at high risk for stroke. This group included the four strokes that occurred within 7 days and six of the seven strokes that occurred at 90 days. See table 5.7 below for the accuracy of the ABCD score (high risk) at predicting the 7- and 90-day risks of stroke.22 Level 3

Table 5.7. The ABCD score (high risk) at predicting the 7- and 90-day risks of stroke.

Table 5.7

The ABCD score (high risk) at predicting the 7- and 90-day risks of stroke.

A large number of patients (79%) were aged ≥60 years and due to the low number of strokes for this variable, a retrospective analysis was performed when this item was removed. This decreased the number of false positives from 44 to 21 at 7 days and from 42 to 19 at 90 days, without changing the scores’ ability to predict stroke.22 Level 3

4.0. Johnston et al. (2007)23

One study validated the ABCD score in four independent groups of patients (N=2,893) diagnosed with TIA in emergency departments and clinics in the USA and UK. From this, a unified score was derived (ABCD2) optimised to predict the 2-day risk of stroke. The ABCD2 score was based on five factors (age ≥60 yrs (1 point); blood pressure ≥140/90 mmHg (1); clinical features: unilateral weakness (2), speech impairment without weakness (1); duration ≥60 min (2) or 10–59 min (1); and diabetes (1)).23 Level 3

In these four groups, 2-day risk was 0% for an ABCD2 of 0 or 1, 1 to 2% for a score of 2, 0 to 3% for 3, 2 to 4% for 4, 3 to 6% for 5, 4 to 14% for 6, and 0 to 50% for 7.

Overall, when the four validation groups were combined 47/4,799 (1%) patients with complete information in the combined cohorts scored 0, 191 (4%) scored 1, 543 (11%) scored 2, 847 (18%) scored 3, 11,165 (24%) scored 4, 994 (21%) scored 5, 852 (18%) scored 6, and 160 (3%) scored 7. Table 5.8 below shows percentage range of risk of stroke at 2, 7 and 90 days for the four validation groups (the sample size of the groups ranged from N=315 to N=1,069).23 Level 3

Table 5.8. The percentage range of risk of stroke at 2, 7 and 90 days for the four validation groups.

Table 5.8

The percentage range of risk of stroke at 2, 7 and 90 days for the four validation groups.

Table 5.9. The data stratified as a low, moderate or high risk of stroke based on the ABCD2 score.

Table 5.9

The data stratified as a low, moderate or high risk of stroke based on the ABCD2 score.

Of 4,746 patients who did not have a stroke during the emergency department evaluation for TIA, 432 (9.1%) were admitted to hospital for the initial attack; this occurred mainly in the validation group from California. Overall, 111 (85%) of 130 strokes occurring within 2 days of the TIA were in patients who were not admitted to hospital, and 45 of these 111 (41%) had an ABCD2 score of greater than 5.23 Level 3

5.2.5. Health economics evidence statement

This section contains the results of an original cost-effectiveness analysis conducted for this guideline. Detailed results are presented in Appendix C, available online at www.rcplondon.ac.uk/pubs/brochure.aspx?e=250

All patients combined

When all patients were assessed in the same way regardless of ABCD2 score, immediate specialist assessment had the least number of strokes and the most QALYs. GP care had the most strokes and least QALYs. Weekly specialist assessment was most costly and GP care least costly.

Immediate specialist assessment dominated weekly specialist assessment, that is to say it was more effective and less expensive than weekly specialist assessment. Immediate specialist assessment was cost effective compared with GP care, costing an extra £3,000 per additional QALY gained.

Prioritising patients by ABCD2 score

The number of strokes averted in the first 90 days after TIA varied greatly by ABCD2 score group. However, immediate specialist assessment dominated weekly clinic for every group. Immediate specialist assessment was cost effective compared with GP care for all groups except 0 and 1.

Limitations

Immediate specialist assessment appears to be cost effective compared to weekly clinics for all ABCD2 score groups, and the results appear to be robust to changes in key parameters.

The results seem to imply that specialist assessment (immediate or weekly) is not cost effective compared with GP care for the lowest ABCD score groups but this cannot be concluded since the model does not incorporate the health gain nor the costs associated with stroke mimics. Furthermore, the model does not capture the health gain attributable to increased uptake of statins and certain other drugs, which are costed in the model.

Although the model includes costs for long-term nursing care for dependent stroke patients, informal care costs were not included since these are not within the NHS perspective. If they had been included, then immediate specialist assessment would have appeared even more cost effective.

The model is a simple representation, looking at only 90 days after the TIA for the effects of medical treatment and extrapolating from this to get long-term outcomes, and thus caution should be applied when using these results. However, the results of this analysis reinforce the conclusions of other studies.

Other related studies

The Wardlaw et al. NHS HTA report indicated that the net benefit of stroke prevention clinics was dependent on the speed with which patients could be investigated or treated.27 As the risk of stroke for TIA patients is high in the first month, treatment strategies which allow patients to be treated within this period appear to be cost effective.

The EXPRESS study, which was published after the development of this model, suggests that the impact of early specialist assessment on stroke risk might be greater still. This before and after cohort study found a relative reduction in stroke risk of about 80% for immediate specialist assessment compared to an appointment-based clinic.

Finally, a forthcoming report for the National Co-ordinating Centre for NHS Service Delivery and Organisation R&D has constructed a similar cost-effective model comparing different assessment strategies.28 Based on the provisional results, it also found that same-day clinics are cost effective compared with weekly clinics for every ABCD2 score group.

Conclusion

Referral of suspected TIA patients for immediate specialist assessment appears to be cost effective because it supports timely prescribing of effective drugs and selection of patients for effective surgery.

5.2.6. From evidence to recommendations

It is clear that scoring systems such as the ABCD and ABCD2 are good clinical predictors and are accurate at identifying patients who are at high risk of subsequent stroke. The level of risk that might be acceptable to patients of completed stroke whilst waiting 7 days for a clinic appointment was discussed with the patient representatives on the group.

The GDG patient view is that any potential risk is a concern to patients. Informing patients of the risk they run whilst awaiting an appointment would cause unacceptable levels of anxiety and distress; they would want appropriate management without delay.

The consensus of the GDG is that high-risk patients need to be immediately identified, assessed and secondary prevention initiated. The GDG considers that high-risk patients are defined as patients with a risk of >4% over 7 days; equivalent to ABCD2 score of 4 or greater. The health economic modelling evidence (please refer to Appendix C for more information, available online at www.rcplondon.ac.uk/pubs/brochure.aspx?e=250) found that the cost effectiveness of immediate assessment declines with ABCD2 score. However, immediate specialist assessment was cost effective compared with weekly specialist assessment even for the lowest ABCD2 score group. This said, the GDG did not feel that immediate specialist assessment was practical for all patients, as it may result in a larger number of non-vascular events (e.g. migrainous symptoms, transient vertigo) being referred urgently to specialist services with the risk that services become overwhelmed. Providers and commissioners concerned about capacity will need to ensure that the highest risk (ABCD2 >3) patients are given highest priority.

Evidence from the EXPRESS and the SOS-TIA studies emphasises the need not only for identification of patients at high risk of subsequent stroke, but also early specialist intervention, including commencement of appropriate secondary prevention treatments and early carotid endarterectomy where indicated. Secondary prevention includes antiplatelet agents, blood pressure management, anticoagulation in selected patients e.g. people with AF, exclusion of diabetes, management of dyslipidaemia including statins, and diet and lifestyle advice, particularly smoking cessation.29

These scoring systems exclude certain populations who may be at particularly high risk such as those with recurrent events and those on anticoagulation who also need urgent evaluation. They also may not be relevant to patients who present late.

Specialist assessment includes:

  • exclusion of stroke mimics
  • identification of vascular territory
  • identification of likely causes
  • appropriate investigation and treatment.

5.2.7. RECOMMENDATIONS

R4.

People who have had a suspected TIA (that is, they have no neurological symptoms at the time of assessment (within 24 hours)), should be assessed as soon as possible for their risk of subsequent stroke using a validated scoring system,*** such as ABCD2.

R5.

People who have had a suspected TIA who are at high risk of stroke (that is, with an ABCD2 score of 4 or above) should have:

  • aspirin (300 mg daily) started immediately
  • specialist assessment and investigation within 24 hours of onset of symptoms
  • measures for secondary prevention introduced as soon as the diagnosis is confirmed, including discussion of individual risk factors.
R6.

People with crescendo TIA (two or more TIAs in a week) should be treated as being at high risk of stroke (as described in recommendation 5), even though they may have an ABCD2 score of 3 or below.

R7.

People who have had a suspected TIA who are at lower risk of stroke (that is, an ABCD2 score of 3 or below) should have:

  • aspirin (300 mg daily) started immediately
  • specialist assessment and investigation as soon as possible, but definitely within 1 week of onset of symptoms
  • measures for secondary prevention introduced as soon as the diagnosis is confirmed, including discussion of individual risk factors.
R8.

People who have had a TIA but who present late (more than 1 week after their last symptom has resolved) should be treated as though they are at lower risk of stroke (see recommendation 7).

Footnotes

*

It was not possible to calculate an accurate diagnostic sensitivity in any referring group because non-referrals to the acute stroke unit were not reviewed.

**

The latter is a subgroup of people in the OXVASC diagnosed by the study neurologist with only possible TIA, made an alternative diagnosis, or could not explain the diagnosis.

***

These scoring systems exclude certain populations who may be at particularly high risk of stroke, such as those with recurrent events and those on anticoagulation who also need urgent evaluation. They also may not be relevant to patients who present late.

Copyright © 2008, Royal College of Physicians of London.

All rights reserved. No part of this publication may be reproduced in any form (including photocopying or storing it in any medium by electronic means and whether or not transiently or incidentally to some other use of this publication) without the written permission of the copyright owner. Applications for the copyright owner’s written permission to reproduce any part of this publication should be addressed to the publisher.

Bookshelf ID: NBK53285

Views

  • PubReader
  • Print View
  • Cite this Page
  • PDF version of this title (5.5M)
  • Disable Glossary Links

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...