• 
• 

Three Part Question

In [patients with suspected stroke in the pre-hospital setting] is [the FAST screening tool more accurate than the ROSIER tool] at [correctly identifying stroke]?

Clinical Scenario

You are tasked to assess a 42-year-old solider who has developed sudden onset slurred speech and weakness in the right arm whilst on exercise. The National Institute for Health and Care Excellence (NICE) guidelines recommend the use of a validated stroke screening tool such as “FAST (Face Arm Speech Test)” in the pre-hospital setting, or “ROSIER (Recognition of Stroke in the Emergency Room)” in the hospital setting. Recognising different screening tools are recommended between settings, you wonder whether there is a difference in accuracy between the screening tools.

Search Strategy

The Health Database Advanced Search interface was used to search the PubMed, EMBASE and Medline databases. The search terms used were: “prehospital”, “pre-hospital”, “stroke screening”, “CVA screening”, “cerebrovascular accident screening”, “ROSIER”, “recognition of stroke in the emergency room” and “FAST”.

Search Outcome

1,233 papers were found and screened, of which 17 were potentially relevant. Following full text review of these papers, 5 were chosen for inclusion, consisting of 4 systematic reviews and 1 retrospective study.

Relevant Paper(s)

Author, date and country	Patient group	Study type (level of evidence)	Outcomes	Key results	Study Weaknesses
Zhelev et al. 2019	Non-comatose, no trauma adults presenting within 24 hours of a suspected TIA or stroke, who were assessed with a stroke recognition scale in a pre-hospital or Emergency Room setting	Cochrane review	Sensitivity and specificity in the pre-hospital and hospital settings	No statistically significant difference in sensitivities and specificities between ROSIER and FAST in the ER. ROSIER: 2895 patients included Sensitivity= 84%- 91% Specificity= 18%-93% FAST: 1894 patients included Sensitivity 64%-97% Specificity	No comparisons made between FAST and ROSIER in the field pre-hospital environment. Small number of studies per test per setting. Heterogeneity between studies. High risk of bias for two of the FAST studies.
Antipova et al. 2019	Adults presenting with ischaemic stroke, acute haemorrhagic stroke, stroke mimicking conditions and transient ischaemic attack in the pre-hospital and hospital settings	Systematic review	Sensitivity, specificity, PPV, NPV	ROSIER: Sensitivity= 79% Specificity= 76% PPV= 61% NPV= 88% FAST: Sensitivity= 84% Specificity= 44% PPV= 32 % NPV= 90%	Only focused on large vessel occlusion, excluding more minor strokes. No differentiation between use of the tools in the pre-hospital or hospital setting
Rudd et al. 2016	Adults presenting with suspected stroke, assessed with a stroke recognition tool prospectively applied face-to-face by a clinician	Systematic review	Sensitivity, specificity, PPV, NPV	ROSIER: 2445 patients included. Sensitivity 83%-97% Specificity 18%-93% PPV 62%-94%) NPV 33%-88%. FAST: 1841 patients included Sensitivity 79%-97% Specificity 13%-88% PPV 62%-89% NPV 48%-93%	Heterogenous study design. Wide variation in the sensitivity and specificity of each tool between studies. Limited recognition of false negative rates, especially in studies where test-negative patients were not transported to the study centre.
Brandler et al. 2014	Adults presenting with suspected stroke, who were assessed by a paramedic or emergency medical technician, with the use of a pre-hospital stroke scale	Systematic review	Stroke prevalence, sensitivity, specificity, LR + and LR- of each tool.	ROSIER: 295 patients included from 1 study Sensitivity= 97% Specificity=18% LR+= 1.17 LR-= 0.19 FAST: 295 patients included from 1 study Sensitivity= 97% Specificity= 13% LR+= 1.10 LR-= 0.26	Large confidence intervals. Only 1 study compared FAST and ROSIER
Purrucker et al 2015	All patients presenting with potential stroke, attended by EMS paramedics and emergency physicians in one emergency department	Retrospective study	Sensitivity, specificity, PPRV, NPV	ROSIER: Sensitivity= 80% (73%-85%) Specificity= 79% (75-83%) PPV= 59% (53%-66%) NPV 91% (88%-94%) FAST: Sensitivity 85% (78-90%) Specificity 68% (64-72%) PPV= 50% (44%-55%) NPV 92% (89%- 95%)	Retrospective study design

References

1. Zhelev et al. Pre-hospital stroke scales as screening tools for early identification of stroke and transient ischemic attack (Review). Cochrane database of systematic reviews 2019 Cochrane database of systematic reviews 2019 (4).
2. Antipova et al. Diagnostic accuracy of clinical tools for assessment of acute stroke: a systematic review. BMC Emergency Medicine 2019;19:49.
3. Rudd et al. A systematic review of stroke recognition instruments in hospital and prehospital settings. Emerg Med J 2016;33:818-822.
4. Brandler et al. Prehospital stroke scales in urban environments. American Academy of Neurology 2014:2241-2249.
5. Purrucker et al Comparison of stroke recognition and stroke severity scores for stroke detection in a single cohort. J Neurol Neurosurg Psychiatry 2015;86:1021-1028.