01 Development and validation of a pragmatic prehospital tool to identify stroke MIMIC patients. Issue 1 (14th January 2019)

• Record Type: Journal Article
• Authors: Flynn, Darren; Rodgers, Helen; Price, Chris
• Is Part Of: Emergency medicine journal. Volume 36:Issue 1(2019)
• Page Start: e1
• Page End: e1
• Publication Date: 2019-01-14
• Subjects: Emergency medicine -- Periodicals; 616.02505
• DOI: 10.1136/emermed-2019-999.1
• Languages: English
• ISSNs: 1472-0205

Abstract

Abstract : Background: Stroke mimics (SM) are non-stroke conditions producing similar symptoms to stroke. Prehospital stroke identification tools prioritise sensitivity over specificity, therefore >25% of prehospital suspected stroke patients are SM. Failure to identify SM Results in inefficient use of ambulances and specialist stroke services. We developed a pragmatic tool for paramedics, using information often available in the prehospital setting, to identify SM amongst suspected stroke patients. Methods: The initial tool was developed using a systematic literature review to identify SM characteristics, a survey of UK paramedics to explore the acceptability of SM identification and regression analysis of clinical variables documented in ambulance records of suspected stroke patients linked to their primary hospital diagnoses (n=1, 650, 40% SM). The initial tool was refined using two focus groups with paramedics (n=3) and hospital clinicians (n=9) and analysis of an expanded prehospital dataset (n=3, 797, 41% SM) to produce the final STEAM tool. Results: STEAM scores six variables: 1 point for Systolic blood pressure <90 mmHg 1 point for Temperature >38.5°C with heart rate >90 bpm 1 point for seizures or 2 points for seizures with diagnosed Epilepsy 1 point for Age <40 years or 2 points for age <30 years 1 point for headache with diagnosed Migraine 1 point for FAST–ve A score of ≥2 on STEAM predicted SM diagnosis in the expanded derivation dataset with 5.5% sensitivity, 99.6% specificity and positive predictive value (PPV) of 91.4%. STEAM was validated using an external dataset (n=1, 848, 33% SM) of prehospital suspected stroke patients where STEAM was 5.5% sensitive, 99.4% specific with a PPV of 82.5%. Conclusions: STEAM uses common clinical characteristics to identify a small number of SM patients with a high level of certainty. The benefits of reducing SM admissions to specialist stroke services should be weighed against delayed admission for the small number of stroke patients identified as a SM.