Conflict of interest: None declared.
What is stroke symptom knowledge?
Article first published online: 19 MAR 2013
© 2013 The Authors. International Journal of Stroke © 2013 World Stroke Organization
International Journal of Stroke
Special Issue: Global Stroke Statistics Edition
Volume 9, Issue 1, pages 48–52, January 2014
How to Cite
Mosley, I., Nicol, M., Donnan, G., Thrift, A. G. and Dewey, H. M. (2014), What is stroke symptom knowledge?. International Journal of Stroke, 9: 48–52. doi: 10.1111/ijs.12024
Funding: This study was supported by a grant from the National Health and Medical Research Council (NHMRC) ‘Centre of Clinical Research Excellence (Neurosciences)’, Australia. A. G. T. was supported by a fellowship from the NHMRC (438700).
- Issue published online: 19 DEC 2013
- Article first published online: 19 MAR 2013
- National Health and Medical Research Council. Grant Number: 438700
- public awareness;
- response to symptoms;
- stroke recognition;
- stroke symptoms
No commonly agreed definition exists for ‘stroke symptom knowledge’ among members of the general public. Recalling at least one correct stroke symptom has been used in the past. However, this criterion was not associated with rapid presentation to hospital. Rapid presentation is vital in order to provide effective acute stroke treatment.
Aims and/or hypothesis
We sought to identify a base level of community stroke symptom knowledge associated with stroke recognition when symptoms occur, an immediate ambulance call, and ‘stroke recognition and immediately calling an ambulance’ as a single sequence of events.
For six-months in 2004–2005, we identified all patients with stroke living in a defined region of Melbourne and who were transported by ambulance to one of the three hospitals. The person who called the ambulance (caller) was interviewed.
One hundred ninety-eight patients were identified and 150 callers interviewed. Symptoms reported most frequently were limb weakness (67%), speech problems (57%), and facial weakness (24%). Reporting at least two of the symptoms – facial weakness, limb weakness, or speech problems (62% of callers) – was associated with stroke recognition (P = 0·004), immediately calling an ambulance (P = 0·065), and both ‘stroke recognition and immediately calling an ambulance’ (P = 0·053).
Knowing at least two of the symptoms – facial weakness, limb weakness, and speech problems – appears to be an appropriate indicator of stroke symptom knowledge as it is associated with stroke recognition and appropriate action. Recognizing stroke symptoms and immediately calling an ambulance increase the potential to reduce prehospital time delays and improve eligibility of acute stroke patients for rapid treatment.
Stroke is the third most common cause of death in the Australian community and the single greatest cause of disability [1, 2]. However, the general public appears to know little about the signs and symptoms of stroke and what to do when symptoms occur [3, 4]. Furthermore, there is limited knowledge about the relationship between stroke symptom knowledge and time to hospital presentation [5, 6].
Time is the most critical factor impacting on the appropriate performance of acute stroke interventions , with many patients excluded from treatment with alteplase because the delay to presentation exceeds the time when treatment is efficacious [8, 9].
Public awareness messages, such as ‘Stroke Chain of Survival’ (United States) and ‘Face, Arms, Speech, Time’ (FAST) (Australia and United Kingdom) [10, 11], have been developed to reduce presentation delays. These programs are based on the premise that better knowledge of stroke symptoms will lead to improved recognition of stroke and an immediate call for an ambulance.
No commonly agreed standard exists to define an appropriate level of ‘stroke symptom knowledge’ among the general public. Some authors have used recall of at least one stroke symptom as a measure of ‘stroke symptom knowledge’ following stroke education [12-15]. However, recalling at least one symptom has not been associated with shorter times to hospital presentation .
Developing community stroke symptom knowledge is vital, but it is unclear what minimum level of stroke symptom knowledge is required to reduce delays to treatment and improve outcomes for acute stroke patients. The impact of community awareness campaigns on the level of community stroke symptom knowledge is only truly applicable to the clinical environment if it is associated with recognition of stroke when it occurs and a rapid call for ambulance assistance .
In this current study, we aimed to quantify ‘stroke symptom knowledge’ in an applied way that may be used to evaluate community awareness interventions aimed at reducing time to hospital presentation. Specifically, we aimed to identify a base level of stroke symptom knowledge associated with both recognition of stroke and an immediate call for ambulance assistance.
This was a prospective observational study of patients from a geographically defined region (population of 383 000) in metropolitan Melbourne who presented by ambulance to one of the three public hospital emergency departments with a final emergency department diagnosis of ‘stroke’ or ‘transient ischemic attack’ (TIA).
The study region was selected for several reasons. First, Melbourne Metropolitan Ambulance Service (now Ambulance Victoria) records for the previous 12 months indicated that approximately 90% of all ambulance-transported patients (n = 762) from this geographic region were transported to one of the three hospitals, namely, Austin Hospital (60%), Northern Hospital (30%), and Royal Melbourne Hospital (RMH) (10%). Second, surveillance of this region was expected to yield a sample of approximately 250 patients over a six-month period, a reasonable snap shot of current practice. Third, the included hospitals provided different stroke services. Austin Hospital and RMH both have large comprehensive stroke services offering intravenous thrombolysis to eligible patients. Northern Hospital offered stroke unit care with a multidisciplinary team but, at the time of the study, did not provide thrombolysis and there was no on-site access to neurological specialist or neurosurgical expertise.
Emergency department (ED) computer records were used to identify potential patients for inclusion in the study. Patients were eligible for inclusion in the study if they were ≥18 years, resident within the study region, transported to hospital by ambulance, and diagnosed by ED staff as having had a stroke or TIA. The person who called for ambulance assistance (‘the caller’) was identified for each case. ‘Caller’ participants were excluded from the study if they were unable to be identified, aged <18 years, or were medical practitioners or members of the police or other emergency services. All callers identified were followed up immediately, often in the emergency department, to minimize any loss of knowledge over time. ‘Callers’ were interviewed using a structured face-to-face questionnaire to obtain demographic data and their description of the stroke event. Participants were asked questions regarding stroke symptoms, their assessment of the patient's problem, and action taken prior to the ambulance call.
‘Stroke recognition’ was defined as callers who stated at interview ‘I thought the problem was stroke’ (or similar response).
‘An immediate call for ambulance assistance’ was defined when the caller reported that their first response to symptom onset was to call an ambulance.
Ethical approval for the study was obtained from each participating hospital. Informed consent was obtained from the patient and caller before any interview was conducted.
Univariable logistic regression was undertaken to explore the associations between different levels of stroke symptom knowledge and (a) stroke recognition and (b) immediate ambulance call. A multivariable model was adjusted for confounding factors (age and sex of the caller) and all other variables with a two-sided P value ≤0·1. Backward stepwise elimination was used to remove all variables with a P value >0·1. Each eliminated variable was reentered into the final model to confirm its exclusion or inclusion.
One hundred ninety-eight patients were recruited into the study (n = 198). One hundred fifty callers (n = 150) were interviewed, seven callers refused to participate, four could not be located, and 27 were excluded under the study exclusion criteria.
In total, 43% of callers were male and the mean age was 63 years (Table 1). Stroke recognition was reported in 93 cases (62%), while 85 callers (56%) reported immediately calling an ambulance. In 80 cases (53%), stroke was recognized and an immediate ambulance call was made.
|Variables||n (%)||Recognizing the problem as stroke||Immediate call for ambulance assistance|
|OR (95% CI)||P value||OR (95% CI)||P value||OR (95% CI)||P value||OR (95% CI)||P value|
|Vision problems||11 (7)||1·07 (0·30–3·86)||0·91||0·91 (0·26–3·13)||0·88|
|Dizziness, loss of balance, fall||25 (17)||0·74 (0·31–1·77)||0·50||0·66 (0·28–1·56)||0·34|
|Numbness||14 (9)||1·60 (0·47–5·35)||0·45||2·03 (0·61–6·80)||0·25|
|Weakness, paralysis of limbs||101 (67)||1·98 (0·98–3·97)||0·055||1·73 (0·84–3·54)||0·135||1·24 (0·63–4·47)||0·54|
|Weakness, paralysis of the face||36 (24)||2·63 (1·11–6·29)||0·029||2·62 (1·08–6·31)||0·032||1·74 (0·79–3·08)||0·17|
|Loss of consciousness||26 (17)||0·66 (0·28–1·56)||0·35||0·41 (0·17–0·97)||0·043||0·41 (0·17–0·98)||0·045|
|Difficulty speaking or understanding||86 (57)||1·52 (0·78–2·97)||0·21||1·43 (0·75–2·76)||0·27|
|Headache||16 (11)||0·76 (0·26–2·18)||0·62||0·74 (0·26–2·10)||0·57|
|Difficulty swallowing||2 (1)||*||*|
|Male sex||65 (43)||0·68 (0·35–1·33)||0·26||0·72 (0·35–1·46)||0·36||0·91 (0·48–1·75)||0·78||0·98 (0·49–1·98)||0·96|
|Caller age <45 years||34 (23)||0·66 (0·38–1·83)||0·19||0·66 (0·28–1·52)||0·33||0·60 (0·28–1·31)||0·21||0·57 (0·26–1·27)||0·17|
|Family stroke history||58 (39)||1·27 (0·65–2·53)||0·70||1·08 (0·52–2·26)||0·83||1·28 (0·66–2·49)||0·72||1·13 (0·56–2·29)||0·73|
|Aware of stroke in the media||55 (37)||1·85 (0·91–3·76)||0·089||1·49 (0·98–5·82)||0·29||1·77 (0·89–3·53)||0·10||1·78 (0·89–3·58)||0·10|
|Aware of the NSF||41 (27)||2·35 (1·05–5·27)||0·038||2·33 (1·03–5·29)||0·043||1·46 (0·70–3·07)||0·31||1·53 (0·69–3·42)||0·29|
The only demographic factor associated with stroke recognition was awareness of the National Stroke Foundation (NSF) (P = 0·043). No demographic variable was associated with an immediate call for ambulance assistance (Table 1).
Stroke symptoms recalled
Callers most frequently recalled symptoms of facial weakness (droop) (24%), arm weakness (67%), and speech problems (57%) (Table 1). Facial weakness (P = 0·029) was the only single symptom associated with stroke recognition (Table 1). No individual symptom was positively associated with an immediate call for ambulance assistance (Table 1).
Knowledge of more than one stroke symptom
When clustering was used to assess factors associated with stroke recognition, we found that recalling ‘limb weakness and face weakness’ (P = 0·007), ‘limb weakness and speech problems’ (P = 0·022), and ‘facial weakness and speech problems’ (P = 0·063) was associated with stroke recognition (Table 2). Recalling at least two of the three symptoms was independently associated with stroke recognition (P = 0·004).
|Symptoms recalled||n (%)||Recognizing the problem as stroke||Immediate call for ambulance assistance|
|OR (95% CI)||P value||OR (95% CI)||P value||OR (95% CI)||P value||OR (95% CI)||P value|
|Limb weakness and speech problems||68 (45)||2·23 (1·22–4·42)||0·022||2·0 (1·03–4·23)||0·042||1·83 (0·95–3·54)||0·072||1·82 (0·92–3·60||0·084|
|Limb weakness and facial weakness||28 (19)||4·61 (1·51–14·09)||0·007||4·82 (1·53–15·19)||0·007||2·19 (0·90–5·35)||0·085||2·38 (0·94–6·00)||0·067|
|Facial weakness and speech problems||21 (14)||2·96 (0·94–9·30)||0·063||3·08 (0·95–10·01)||0·061||2·78 (0·96–8·05)||0·059||3·04 (1·01–9·14)||0·047|
|Recall of two or more of the following: limb weakness, facial weakness, and speech problems||93 (62)||2·96 (1·49–5·58)||0·002||2·86 (1·41–5·81)||0·004||1·91 (0·99–3·67)||0·05||1·89 (0·96–3·73)||0·065|
Each pair of the three frequently reported symptoms exhibited a strong but borderline association with an immediate ambulance call [‘limb weakness and face weakness’ (P = 0·067), ‘limb weakness and speech problems’ (P = 0·084), and ‘facial weakness and speech problems’ (P = 0·047)] (Table 2). When adjustment was made for demographic variables, the ability to recall at least two symptoms was associated with seeking immediate ambulance assistance (P = 0·065) (Table 2).
Each pair of the three frequently reported symptoms exhibited a borderline association with ‘stroke recognition and an immediate ambulance call’ (Table 3). Furthermore, the ability to recall at least two symptoms was associated with ‘stroke recognition and an immediate ambulance call’ (P = 0·029) (Table 3). After adjustment for demographic variables, the ability to recall at least two symptoms (P = 0·053) and awareness of stroke in the media (P = 0·045) were associated with ‘stroke recognition and an immediate ambulance call’ (Table 3). More than one-third of callers (37%, n = 55) reported stroke being mentioned in the media with 32 callers (21%), stating that they had gained knowledge of stroke from television programs, including news and advertisements.
|OR (95% CI)||P value||OR (95% CI)||P value|
|Limb weakness and speech problems||68 (45)||1·87 (0·97–3·59)||0·061|
|Limb weakness and facial weakness||28 (19)||2·11 (0·89–5·03)||0·092|
|Facial weakness and speech problems||21 (14)||2·46 (0·89–6·74)||0·08|
|Recall of two or more of the following: limb weakness, facial weakness, and speech problems||93 (62)||2·08 (1·08–4·02)||0·029||1·93 (0·99–3·77)||0·053|
|Aware of stroke in the media||55 (37)||2·20 (1·11−4·36)||0·025||2·04 (1·01–4·01)||0·045|
We found that people recalling at least two of the three most commonly recognized symptoms of stroke, i.e., facial weakness, limb weakness, and speech difficulties, was associated with a much greater likelihood of recognizing the problem as stroke than any one of these symptoms. These symptoms are those specified in the FAST stroke assessment tool because of their positive predictive values for stroke [18, 19].
Stroke symptom knowledge has been described in previous studies as the ability to recall one or more stroke symptoms [13, 15]. This categorization separates those with some stroke symptom knowledge from those with no knowledge. However, this cutoff may not be meaningful in reducing delay times for stroke. An appropriate level of stroke symptom knowledge should be associated with not only the ability to recognize stroke when it occurs but also the capacity to know to call an ambulance .
The observation that an ability to recall at least two of the FAST symptoms was significantly associated with stroke recognition and the action of immediately calling an ambulance demonstrates that knowledge can lead to action. The ability to develop and transfer prior stroke symptom knowledge into recognition and action remains the foundation element of stroke community awareness campaigns [10, 11].
We specifically aimed to determine a definition of stroke that was associated with both recognition and action during a real stroke event. Recalling at least two of the FAST symptoms (62% of all callers) met these criteria. This level of recognition is approximately double that reported by Hickey et al. in their study of the general public . Others have reported that between 37% and 47% would call an ambulance for stroke [3, 15, 21]. Using simulated research recall tasks, as used by Billings-Gagliardi and Mazor , may only go part of the way to replicating real world emergency events. To address this issue, we sought out the person who made the ambulance call during a real stroke event. No previous study has uniformly included interviews with callers. Callers provided an in-depth account of the events as they were active participants in the process of recognizing symptoms, the decision to seek care, and calling an ambulance . While every effort was undertaken by the investigators to reduce any recall bias, some discrepancies may exist in the recall of information emanating from such a stressful event.
A limitation of this study is that we only included patients with stroke who arrived by ambulance. No information was available on those who presented by private transport or did not present to hospital. This study does, however, provide robust data that reflect community awareness and response to stroke symptom onset over a six-month period in a large metropolitan city. The analysis undertaken in this study directly mirrors the symptoms currently in use by NSF in their FAST community awareness campaign . The additional finding that awareness of stroke in the media was associated with both stroke recognition and an immediate ambulance call supports the assertion that mass media advertising can change both knowledge and behavior following the onset of stroke symptoms .
In conclusion, there is no right approach to defining an intangible such as community stroke symptom knowledge. However, setting a clear benchmark for stroke symptom knowledge that is easily applied in a clinical environment is important if we wish to understand the impact of interventions on the knowledge, behavior, and actions of the public following the onset of stroke symptoms. Furthermore, having uniformly agreed definitions may assist comparing the results of different community awareness interventions around the world.
Our findings provide evidence that knowing at least two of the FAST symptoms is associated with both understanding what is wrong (stroke recognition) when stroke symptoms occur and knowing what to do (immediately call an ambulance). Further research is required to better understand care-seeking decisions following symptom onset and the causes of delay to hospital presentation for patients with acute stroke.
- 9Challenges for cerebrovascular disease. Front Neurol 2010; 1:3. Epub 2010/12/29., .
- 20Knowledge of stroke risk factors and warning signs in Ireland: development and application of the Stroke Awareness Questionnaire (SAQ). Int J Stroke 2012; 7:298–306., , , , .