Article Text
Abstract
Objective Many patients present late after a transient ischaemic attack (TIA). This delays intervention and may partly depend on where patients first present—emergency department (ED) or general practitioner (GP). Studying this behaviour could improve stroke prevention through better targeting of public education and allocation of resources.
Methods Patients with TIA or minor stroke referred to neurovascular clinics in the UK and Canada were studied and the delay from onset to first medical presentation, whether at an ED or GP, was measured. Clinical features, timing and place of presentation were compared.
Results Of 666 patients (469 in the UK and 197 in Canada), only 42% presented on the day of the TIA. The majority (77%) of patients presenting to an ED presented on the same day compared with only 11% of those who presented to a GP. GP delays were longer at weekends. Motor or speech symptoms and prolonged duration were associated with presenting early and to an ED. High-risk patients (ABCD2 score 6–7) in Canada were also more likely to go to an ED. Overall, 65% of Canadian patients and 40% of UK patients went to an ED.
Conclusions Most patients presenting to an ED go urgently, whereas most going to a GP delay, particularly at weekends. Most Canadian patients, particularly those at high risk, go to an ED whereas most UK patients go to a GP. One way to reduce delay, particularly in the UK, would be to direct all patients with TIA to go to an ED rather than to their GP.
- Ischaemic attack, transient
- stroke
- risk reduction behaviour
- time factors
- prevention
- acute medicine-others
- emergency care systems, primary care
- emergency care systems, emergency departments
- neurology, stroke, nursing, pre-hospital
Statistics from Altmetric.com
- Ischaemic attack, transient
- stroke
- risk reduction behaviour
- time factors
- prevention
- acute medicine-others
- emergency care systems, primary care
- emergency care systems, emergency departments
- neurology, stroke, nursing, pre-hospital
Recent trials have shown that effective intervention given as soon as possible after a transient ischaemic attack (TIA) can dramatically reduce the risk of stroke,1 yet most patients delay seeking medical advice. This behaviour is recognised by the American Heart Association Council as ‘the major factor limiting delivery of definitive treatment’. In a US study, only 52% of patients presenting to an emergency department (ED) with a TIA or stroke thought immediate presentation was crucial.2 Similarly, in a recent survey of a UK population, a high proportion of patients suffering a TIA would wait for a recurrence before seeking medical advice: 41–95% depending on the symptom.3 This makes achieving the goal that all high-risk patients should be assessed by a specialist within 24 h, as proposed in the UK government Stroke Strategy (2007), extremely challenging. Patients need clear education as to how to recognise a TIA and its significance, and equally clear direction as to where to turn for assessment and treatment. Patients' current behaviour severely limits the benefits to be gained from recent innovations in stroke prevention such as daily specialist TIA clinics,1 a 24 h specialist assessment and investigation service4 and the adoption of scoring systems in the ED to stratify risk,5 with immediate admission for those at highest risk.6
Given that 12% of high-risk patients and 6% of moderate-risk patients will suffer a stroke within 7 days7 and that up to 32% of referrals to a TIA clinic may suffer a stroke while waiting to be seen,8 a clear change in patient behaviour has to occur if we, as medical professionals, are to initiate time-dependent effective secondary prevention.
We therefore studied the speed of response of patients who had suffered a TIA. Our main points of comparison were where they went (ED or general practitioner (GP)) and when they went (the delay between onset of symptoms and first medical assessment). We compared data collected as part of a prolonged study of patients referred to a specialist clinic in the UK with a snapshot of a similar cohort in Canada using data collected over a 4-month period, sufficiently long to reveal important similarities and differences. Both countries have publicly funded healthcare systems free at the point of delivery, and a choice of where to present in case of sudden illness, principally either to an ED or their GP (termed family physician in Canada). However, whereas the UK has only just started to run a national stroke awareness campaign (February 2009), Canada has been doing so for many years (Heart and Stroke Foundation of Ontario 1999, Alberta Provincial Stroke Strategy 2006). Furthermore, alteplase (recombinant tissue plasminogen activator) in acute ischaemic stroke was approved in Canada 3 years earlier than in the UK, with 10% of eligible patients in Canada receiving thrombolysis in 20049 compared with only 0.2% patients in the UK.10 Studies from North America suggest that patients there present earlier.11 12
We postulate that increased awareness of emergency treatment for stroke in Canada has reduced delays for patients with TIA compared with the UK. We test this hypothesis and analyse the delay in terms of the place of presentation and the clinical characteristics and timing of the TIA. We go on to discuss the implications of this for the design of stroke prevention programmes and the provision of services.
Methods
Subjects
Study subjects were recruited from consecutive patients referred from an ED or GP to two stroke prevention clinics, one in the UK and one in Canada. The UK cohort was drawn from a database of all patients seen between January 2003 and June 2007 in a consultant-run clinic in a district general hospital serving a largely urban population of 250 000 in Hertfordshire, 20 miles from London. The hospital had no hyperacute stroke service until November 2006 when thrombolysis became available during normal working hours only. There was no local campaign to raise public awareness either then or at any time during the period of the study. The comparison cohort was drawn from consecutive patients attending the stroke prevention clinic at the University Hospital in Edmonton, Alberta between May and August 2006. The hospital acts as a general hospital for the population of Edmonton as well as being a tertiary referral centre for stroke, serving a wider population of 1.5 million people, and has provided thrombolysis for acute stroke for the past 10 years. Both clinics offered a rapid response one-stop service with same-day investigation. Inclusion criteria were a definite or probable diagnosis of TIA (transient neurological deficit of presumed vascular origin lasting less than 24 h) or minor stroke (minimal residual disability) as determined by a stroke physician. Exclusion criteria were alternative diagnoses (TIA mimics) or uncertain time of onset.
The Canadian data were extracted retrospectively by chart review. All UK patients were asked specific questions about the time and date (weekday or weekend) of their first symptom, the time they consulted a medical opinion and the time and place where they were first seen, either by a GP or an ED doctor. Dates were checked against written records from ED, GP referral letter and hospital patient administration system. Data collected also included demographic details and duration and clinical features of presenting symptoms. These symptoms were classified into four types as motor (including clumsiness or incoordination), sensory (no motor symptoms), speech (dysphasia or dysarthria without motor symptoms) and visual. The duration of the TIA was recorded where patients were clear about time of onset and recovery. After publication of the ABCD2 score,7 this was calculated retrospectively where data were available. The main outcome measure was the delay between the onset of the first symptom (day 0) and first presentation for medical assessment. The effect of the various demographic and clinical characteristics on this delay was analysed in the two populations and compared.
Statistical analysis
The results between the two populations were compared using a χ2 test for dichotomised data and a t test for continuous data.
Results
A total of 666 patients with TIA or minor stroke referred from an ED or a GP were seen in the study periods in the two countries, 469 in the UK and 197 in Canada. Overall, 38% of UK patients and 54% of Canadian patients presented on the day of the TIA (day 0, p=0.001). In the UK 40% of patients and in Canada 65% of patients presented to the ED.
Speed and place of presentation
UK
A total of 856 new patients (mean age 69 years, range 32–96) with transient neurological symptoms were referred to the UK clinic between January 2003 and June 2007 from an ED or a GP. The number referred rose annually but the ratio of ED to GP referrals remained at 1:4. Patients with stroke/TIA mimics (n=387) were excluded, leaving 469 consultations with definite or probable TIA or minor stroke to be analysed (table 1).
Just over one-third presented on day 0, while two-thirds delayed 1 day or longer. A minority (40%) presented to an ED of whom three-quarters presented on day 0. Eleven percent of patients presenting to a GP presented on day 0, a similar percentage the following day and 77% later still. One-third presented after 1 week; the median delay was 5 days.
Canada
Data on 197 patients (mean age 68 years, range 37–91) with TIA or minor stroke were analysed. One-half (54%) presented on day 0. Two-thirds presented to an ED and, as in the UK, three-quarters of these presented on day 0 (77%). The same proportion of patients as in the UK (11%) presented to a GP on day 0. The median delay was 8 days. In Canada, 40% of the patients presenting to an ED were assessed by a neurologist the same day (this was not available to the UK cohort).
Day of the week
In both countries less than 3% of patients presented to a GP at weekends, and only one patient (UK) presented on day 0. In contrast, numbers presenting to an ED remained consistent across all 7 days of the week. Dividing patients according to when their TIA occurred (weekday or weekend) and taking both countries together, 51% (219/430) of weekday patients and 41% (59/145) of weekend patients presented on day 0 (p=0.03).
In the UK the percentage presenting on day 0 remained constant at 77% if they went to an ED but fell from 14% (30/217) for weekday patients to 2% (1/61) for weekend patients if they presented to a GP (figure 1). Similarly, in Canada none of 19 patients whose TIA occurred at a weekend presented to a GP on day 0 compared with 16% (8/50) of those whose symptoms occurred on a weekday. As in the UK, patients going to an ED showed no difference in time to presentation according to the day on which they presented, with 73% weekend and 79% weekday patients presenting on day 0 (figure 2).
Clinical features
ABCD2 score
The ABCD2 score was calculated retrospectively in 100% of the Canada cohort and 69% of the UK cohort; in 31% incomplete data (most often uncertainty over the duration of a TIA) prevented calculation of the score. The scores were normally distributed (figure 3) and the study populations were comparable in level of risk, the mean score being 4.7 in Canada and 5.3 in the UK (p=0.79; figure 3).
In Canada, patients at highest risk (ABCD2 score 6–7) were more likely to present to the ED than lower risk patients (91% (20/22) vs 61% (107/175), p=0.005). In the UK there was no such difference between higher risk and lower risk patients presenting to the ED (29% (19/65) vs 37% (95/258), p=0.2). High-risk patients in Canada were more likely to present to the ED than high-risk patients in the UK (p=0.00001).
Clinical features
In both countries, patients with speech or motor impairment were more likely to present on day 0 than those with sensory or visual deficit (p= 0.049) and, in the UK, these patients were more likely to present to the ED (p=0.04). In both countries, patients with purely visual symptoms were least likely to present on day 0 (p=0.0004, table 2).
Duration
In both countries patients with prolonged TIAs (>5 min) were more likely than patients with shorter TIAs to go to an ED than to a GP (p=0.00008) and to present on day 0 (p=0.002, table 3).
Discussion
Our study showed that, in both Canada and the UK, patients present late after a TIA. Canadian patients did slightly better than UK patients, but in both countries 77% of those who went to an ED presented on the same day compared with only 11% of those who went to their GP. The difference was more significant at weekends, with only a single patient who presented to a GP being seen on day 0.
The main difference between the two countries is that the majority of patients in the UK go to a GP whereas in Canada the majority go to an ED, despite both services in both countries providing emergency medical services.
Patients with longer TIAs were more likely to present to an ED and to present early in both countries. In Canada, patients at highest risk (ABCD2 score 6–7) and, in the UK, patients with motor or speech TIAs were also more likely to go to an ED.
In the UK the percentage presenting the same day (38%) did not show a consistent rise over the 5 years studied, and compares with the 44% presenting as an emergency in the study by Giles et al13 in Oxford, UK. In that study, only 10% presented to an ED compared with 40% in our study, which may represent local variation or a true change in patient behaviour.
Should all TIA patients be directed to bypass their GP and go straight to an ED?14 Doing so would emphasise the seriousness of TIAs and facilitate hospital admission for those at highest risk, which is cost-effective, facilitating thrombolysis if a stroke occurs.15 Another advantage would be greater consistency of approach in assessment and initial management, and the opportunity for review by a neurologist, as was the case in our study in Canada but not in the UK. The main advantage would be ease and speed of access at all times. Currently, most UK patients present to a GP, a pattern of behaviour which has persisted despite service reconfiguration in April 200416 when 95% of GPs opted not to provide an out-of-hours service themselves but instead to use a deputising service, and simultaneously waiting times in EDs fell in response to a government initiative.17 18 In fact, the majority of patients admitted with acute illness in the UK still present to primary care first unless an ambulance is called on their behalf.19 If access is the problem, then delays incurred by going to a GP might be reduced by extending surgery hours, but it has been calculated that surgeries would have to open from 08.00 h till 20.00 h 7 days a week to increase the percentage of patients calling within 24 h from 34% to 68%.20 We suggest that a more practical solution is to direct patients to go to an ED.
Some stroke educational programmes have successfully increased the likelihood that patients would call the emergency services21 or attend an ED.22 Nonetheless, changing long-established behaviour will be challenging. Only 40% of patients attending an ED after a cerebrovascular event correctly recognised the diagnosis,2 and even patients who have suffered a previous TIA or stroke may be ignorant of the warning signs, both immediately23 and 3 months later.24 Even in the high profile cardiac arena, one comprehensive community intervention programme proved ineffective in reducing time from symptom onset to arrival in the ED in those with chest pain.25
Limitations
Our study did not capture the behaviour of patients who did not present for medical assessment, were not referred or who failed to attend the clinic. These patients may go on to suffer a potentially avoidable stroke. Among patients presenting on the day of the TIA, we were unable to distinguish those who responded immediately from those who initially delayed for a few hours. We excluded approximately one-third of clinic patients (TIA mimics) but, arguably, these patients should have been included if it is appropriate for all transient neurological symptoms to be assessed urgently. We believe our data do reflect patient choice rather than simply the availability of services as, in both countries, EDs and GPs do offer a 24 h service. The reasons why patients delayed and why they chose a GP rather than an ED are the subject of an ongoing study. The periods of study and hence the number of patients included differed in the two sites, but the shorter period of study in Canada was long enough to reveal the similarities and differences between the two equivalent patient groups. Percentages were used for comparisons between the cohorts to make the differences clear.
Conclusion
In both the UK and Canada, the percentage of patients presenting urgently after a TIA was seven times greater if they presented to an ED rather than a GP. In the UK, most patients at all levels of risk go to a GP whereas, in Canada, most patients—and high-risk patients in particular—go to an ED. Many patients, particularly in the UK, incur long waiting times and miss out on assessment and treatment at the time of highest risk. This delay was most prolonged at weekends for patients going to a GP.
We suggest that the behaviour of most patients with a TIA in Canada who go directly to an ED provides a model for UK patients to follow. Public awareness campaigns should emphasise this simple message.
Acknowledgments
Dulka Manawadu was employed as a stroke fellow during the production of this work.
References
Footnotes
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.