The ‘time is brain’ concept and emergence of hyperacute treatments, such as thrombolysis, means that treatment of stroke is increasingly being regarded as an emergency.1 Patients with transient ischaemic attack (TIA) are less likely than those with stroke to receive emergency admission or timely specialist assessment,2^–4 and some patients with TIA never seek medical attention.5 There is now clear evidence, however, that the risk of recurrence in patients with TIA is highest within the first 24 hours.3 In addition, it is possible to identify patients at greatest risk of recurrent stroke by using validated scores.6 Appropriate secondary prevention—including antiplatelet administration, antihypertensive therapy, lipid lowering therapy, anticoagulation therapy and surgical intervention for significant carotid stenosis—can reduce the risk of subsequent events.7^–10 Rapid access schemes that combine prompt assessment and investigation with early initiation of secondary prevention have proved effective in reducing recurrent stroke.7 9

Despite these advances in management of both stroke and TIA, many patients are slow to seek medical advice after a vascular event.11 Numerous studies have investigated demographic and clinical factors relating to delay in presentation after stroke. More recently, perceptual and behavioural responses have also been assessed.12 Systematic reviews have consistently demonstrated that failure to contact the emergency medical services (EMS) is the single most important contributor to delay in assessment in acute stroke, with social and demographic factors being less important.11 13

Although some of these stroke studies included patients with TIA, to date there has been little work exploring delays and potential reasons for delay after TIA. We sought to perform a systematic review of the literature to examine potential factors associated with delay in seeking medical review in patients with TIA.

METHODS

We sought observational studies assessing delay in presentation in patients with TIA. Electronic searches were performed in the MEDLINE, EMBASE and Science Citation Index databases for papers published between December 1995 and September 2008. The reference lists of all relevant articles were also checked to identify further suitable studies. The search strategy was developed for MEDLINE and adapted for the other databases. Only studies published in the English language were included. Studies were selected for inclusion if they assessed delay in seeking medical attention after TIA. Delay was taken as the time between symptom onset and first seeking medical advice, whether that be via the EMS, at the emergency department or from general practitioner (GP). Time from presentation to specialist (stroke or neurology) assessment was not assessed. Disagreement regarding study inclusion was resolved by discussion (NS, CM, AW, TGR), with data extracted and checked by two authors (NS, CM) independently.

RESULTS

The initial electronic search yielded 1,854 titles, all of which were reviewed in abstract form; 107 articles were subsequently reviewed in full (fig 1). A total of 55 publications assessed delay in presentation after stroke or TIA; however, the majority of these studies excluded patients with TIA. Of the 22 studies that included TIA patients, nine of these did not present data for the TIA cohort. In a further four studies, although some TIA cohort data was present, no information on presentation delay in the TIA cohort was available. Due to study heterogeneity, pooling of results or meta-analysis was not possible. Median delay to presentation is presented for each study in table 1.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Flow chart of the systematic review search strategy. TIA, transient ischaemic attack.

Study design

Nine of the studies we identified provided data on presentation delay in patients with TIA (table 1), with large variations in study size, population and methodology. One study included TIA patients only (n = 241),14 whereas the remaining eight studies recruited both stroke and TIA patients, with TIA patients (n = 821) making up only a small proportion of the total patient number (n = 3,202). The majority of studies used a structured interview format to evaluate patients’ behaviour and perception immediately after symptom onset.14^–19 One study ascertained whether patients recognised symptoms as stroke/TIA in addition to collecting demographic data.20 Two studies were population based,14 19 whereas the remaining studies recruited patients from hospital-based stroke registries. Two studies did not perform interview and provided only demographic data from stroke registries.18 21 Four studies stated that patients were recruited consecutively.14 16 19 22

Presentation via emergency medical services

Referral pattern was recorded in all studies, depending on whether patients first contacted the EMS, attended an emergency department (ED) or visited a primary care physician. Across all studies, less of a delay was consistently seen in patients (stroke and TIA) who used the EMS, although patients with TIA were less likely than stroke patients to use EMS.15–17 21

Presentation to an emergency department

Length of delay varied, but in general the delay was shorter in patients attending an ED than in those calling a GP. In one study,22 for example, 75% of patients with TIA arrived at the ED in less than 3 hours, with those who recognised their symptoms as stroke/TIA more likely to reach the ED within 3 hours. In another study,18 a third of patients with TIA presented within 6 hours and the majority between 6 hours and 24 hours, although a fifth of patients waited more than 24 hours until presenting.

Presentation to a general practitioner

In studies conducted in the UK, the vast majority of patients first presented to their GP,14 20 with only a minority (10–26%) attending an ED. Half of all patients attended within 24 hours of symptom onset, whereas a quarter of patients waited 2 days or more to seek attention.14 The longest delays were seen when symptom onset occurred at the weekend.14 The time to call a GP was significantly increased in patients who had events outside practice opening hours (median 12.9 hours vs 4.0 hours in patients who had events during practice opening hours).20 Delay was not increased in patients who attended an ED out of GP opening hours (0.9 hours vs 0.7 hours in patients who attended an ED during GP opening hours).

Comparison with ischaemic stroke

The length of delay in presentation for patients with TIA compared to patients with stroke was inconsistent. One study21 demonstrated shorter delays in TIA patients than in stroke patients (median 4.9 hours vs 5.7 hours), a finding that was replicated in three other studies.16 19 22 In contrast, another study found that patients with TIA had longer delays than did patients with stroke (median 3.3 hours vs. 2.8 hours).15

Recognition of symptoms

In one study, delay was reduced when patients recognised their symptoms as those of stroke or TIA.22 In another study (median delay after TIA 4 hours), half of all patients with TIA failed to recognise their symptoms as being stroke related, and this factor was associated with increased delay. A fifth of patients in Wester et al.’s study did recognise their symptoms as stroke related, yet still failed to seek urgent medical attention.16 Failure to contact anyone when symptoms developed accounted for the majority of delay time in TIA patients in this study.16 In the studies conducted in the UK, recognition of symptoms as stroke did not reduce delay14 20 or increase the proportion of patients who attended an ED.20

Other factors

The presence of a witness at symptom onset15^–17 reduced delay across stroke and TIA patients, but no data was available on the influence of a witness in patients with TIA alone.

In the UK study by Giles and co-workers, patients with motor symptoms and patients with symptom duration of greater than 1 hour were both less likely to delay in seeking medical attention than other patients.14

DISCUSSION

Prompt recognition of TIA with urgent assessment and treatment is increasingly being recognised as a priority;23 however, many patients delay in seeking medical attention after TIA. In this review, we sought to explore further the nature of these delays. To date, most research has focused on delays after stroke. Although some of these studies included patients with TIA, many fail to report data separately for the TIA cohort, making it difficult to draw conclusions for delay after TIA.

Despite the limited amount of data available on patients with TIA, our analysis shows that many patients did delay in seeking medical attention after TIA. As in acute stroke, delay is generally longer in patients attending primary care rather than those utilising the EMS or attending an ED. Recognition of TIA symptoms was not good,14 16 and even those patients who recognised their symptoms as those of TIA did not always seek emergency medical attention,14 20 a phenomenon that has also been documented after stroke.17 24 25 Furthermore, recognition of symptoms did not influence the patients’ decision to attend an ED or contact an out of hours GP after TIA, resulting in delay and potentially preventable stroke.20 Taken together, these findings reflect what many of us suspect from clinical practice—that public awareness of TIA and stroke remains inadequate.

Reduced delay in seeking medical help among patients with severe stroke, haemorrhagic stroke or a decreased level of consciousness26 has been attributed to a witness calling the emergency medical services.15 17 Schroeder et al. reported that patients with TIA were less likely to call the EMS;17 thus, it is perhaps surprising that some studies found shorter delays in TIA patients than stroke patients. It is possible that rapid clearance of symptoms in TIA may allow patients to react faster than those with a severe deficit that may limit help seeking behaviour.19

Population surveys have demonstrated that public perception and knowledge of TIA is poor; for example, only 8% of people in a US study were able to recognise the definition of TIA or the symptoms.27 In a Swiss study,28 87% of people surveyed had never heard the term TIA, less than 10% recognised TIA symptoms as being related to stroke and more than half would wait days or weeks before contacting their family physician for advice. Even in an at risk clinical trial population who have received related education, reporting of TIA events was poor.29

Mass media educational campaigns do appear to work. Multi-level interventional programs have been effective at improving service delivery of thrombolysis,30 although as yet it remains unclear which type of educational interventions are the most effective.31 An observational study assessing community stroke education using mass media campaigns observed increased ED attendance among patients with stroke and TIA.32 Interestingly, the campaign appeared to have greatest effect on ED attendance in patients with TIA, suggesting that the media campaign was more effective in reducing assessment delay after TIA than after stroke.

Not only the public needs educating. In the UK studies,14 20 only a very small number of patients with TIA or minor stroke called NHS direct, a national health help line, and some patients who used the service were given suboptimum advice. Similarly, in two linked US studies, both emergency telephone dispatchers and ambulance paramedics had suboptimum levels of recognition of stroke symptoms.33 34 These findings support the need for ongoing training to enable telephone dispatchers and paramedics to recognise TIA and stroke symptoms and thus triage patients appropriately.

Given that only one study focused solely on TIA and that there were variations in study populations and methodology in the trials we assessed, analysis and interpretation of the data regarding delays after TIA is limited. Notably, most of the studies reviewed here were set up to assess acute stroke and were hospital based. Patients with TIA who present to an EMS or ED may represent a different group of patients to those who present to primary care. Furthermore, models of care varied: in-patient management of TIA in Europe and North America is very different to outpatient based care in the UK, where patients are often seen after some delay.35 Most studies used a structured interview to assess perception and behaviour. Delayed interviews might be subject to recall bias in patients diagnosed with stroke or TIA, however, as subsequent investigation and management may influence patient recollection. All studies excluded patients with unknown time of symptom onset or excessive delay, again leading to bias and potential underestimation of delay. Finally, these studies cannot be extrapolated to provide explanation for the TIA patients who fail to seek any medical attention at all.

Despite limited numbers, this review highlights that many patients do not recognise stroke/TIA symptoms, and even when they do, many fail to seek emergency medical attention. The majority of delay in seeking assessment is due to lack of response by patient. Patients presenting to general practice delay longer than those that attend an ED. To date, research in delay has focused on stroke patients and on nonmodifiable demographics. Future work should focus on social, cognitive and emotional factors that contribute to delay in seeking medical assistance11 and may help in developing educational programs to overcome such factors.12

Although the development of 24 hour TIA assessment services is important, more work is needed to educate the public to seek medical review immediately after TIA in an attempt to avoid recurrent stroke and its devastating consequences. On the basis of current evidence, treatment delay is minimal if patients contact the EMS or attend an ED immediately.