Abstract WP243: Identification of Common Confounders in the Prehospital Identification of Stroke in Urban, Underserved Minorities

Background: Pre-hospital providers report positive predictive values (PPV) of 40-85% for the diagnosis of stroke, typically using the Cincinnati Pre-hospital Stroke Scale. In New York City 911, both emergency medical technicians [Basic Life Support (BLS)] and paramedics [Advanced Life Support (ALS)] care for/transport stroke patients. Hypotheses: (1) In an urban, underserved, minority population, ALS would have greater PPV than BLS for the confirmed diagnosis of stroke, and (2) Hypoglycemia and seizures would be frequent confounders, particularly for BLS providers. Methods: After IRB approval, medical records were reviewed of patients (85% Carribean Blacks) transported by EMS to participating Brooklyn facilities from 1/1/10-12/31/11. We compared EMS impressions of the presenting problem with AHA/ASA “Get with the Guidelines” (GWTG) databases - serving as the gold standard for diagnosis of stroke/TIA. Medical records of patients thought to have stroke by EMS but without a GWTG stroke diagnosis were queried for confounders (e.g. hypoglycemia, seizures). Patients who were not transported by EMS were excluded. Differences in confounders between ALS and BLS were analyzed using chi squared and Fisher’s exact test. Results: Of 671 patients suspected of having stroke by EMS, 628 patients (94%) had complete records (n=445 BLS, 183 ALS) available for review; 392 were confirmed strokes (275 ischemic, 46 TIA, 66 ICH, 5 SAH). There was no difference in PPV between ALS 59% (95% CI, 52% to 66%) and BLS 64% (95% CI, 59% to 68%) (P=0.277). Of the remaining 236 who did not have confirmed strokes, seizure was the greatest single confounder (n=72, 31%) followed in frequency by syncope (n=24, 10%), infections (n=23, 10%) and hypoglycemia (n=18, 8%). ALS more frequency diagnosed confounders than BLS (chi squared 8.64, df=3, p = 0.034), driven by seizure (2 tailed p < 0.0001). Conclusions: Our data suggest EMS called seizures a stroke more commonly than the other 3 leading confounders (hypoglycemia, syncope, and infection) combined and this was true more frequent for ALS than BLS, suggesting an area for focused EMS education. There was no difference between ALS and BLS in PPV. Further study is planned to determine the accuracy of ALS vs. BLS providers for the diagnosis of stroke.