Factors related to the sensitivity of emergency medical service impression of stroke.

OBJECTIVES: To examine factors related to sensitivity of emergency medical services (EMS) stroke impression.
METHODS: We reviewed ambulance and hospital records of all patients transported to Long Island College Hospital between January 1, 2009 and January 1, 2011 by the hospital-based EMS with a discharge diagnosis of stroke or a confounding diagnosis, and compared EMS impression to hospital discharge diagnosis. We examined relationships between EMS diagnostic sensitivity and age, gender, ethnicity, NIH Stroke Scale (NIHSS), motor signs, aphasia, neglect, lesion side, circulation, stroke type, EMS provider level, and documented Cincinnati Pre-hospital Stroke Scale (CPSS) with contingency analysis and logistic regression.
RESULTS: Stroke was validated in 18% (56/310) of patients and 50% (28/56) of these were missed by EMS. EMS diagnostic sensitivity was 50% (95% CI: 36-64%), and was related to NIHSS quartile (p = 0.014), with higher sensitivities in 2nd (69%; 95% CI: 44-86%) and 3rd (75%; 95% CI: 47-91%) vs. 1st (20%; 95% CI: 7-45%) and 4th (45%; 95% CI: 21-72%) quartiles, motor signs (62 vs. 14%, p = 0.002), and documented CPSS (84 vs. 32%, p = 0.0002). EMS impression was independently related to NIHSS quartile (1st vs. 2nd adjusted OR = 9.61, 1.13-122.03, p = 0.038) and CPSS (adjusted OR = 12.58, 2.22-111.06, p = 0.003).
CONCLUSION: Stroke was missed more frequently when CPSS was not documented, in patients without motor signs, and in patients with moderate-severe stroke. The sensitivity of prehospital screening for patients with moderate-severe stroke might be improved by including additional non-motor signs and by stressing indications for when screens should be performed.