Abstract 35: Dispatcher Impression of Stroke: Concordance With Paramedic Identification and Effect on Allocation of Standard Ambulances and Mobile Stroke Units

Background: EMS 911 dispatchers must rapidly assess patient calls and allocate appropriate responding resources including, for stroke, advanced (rather than basic) life support standard ambulances and, in an increasing number of regions, specialty Mobile Stroke Units. While dispatcher accuracy in distinguishing stroke from other high risk conditions is not crucial for selecting ALS rather than BLS response, specialized MSU response necessitates a higher level of accuracy. Methods: We analyzed all transports between January 2018-August 2019 identified as stroke by dispatchers at sendout or by paramedics upon patient assessment in a large urban EMS system -LA County FD EMS, serving 4.1 million lives, 59 cities, 3,000 square miles, via 174 fire stations. Dispatchers are instructed to decide upon the appropriate diagnosis and responding resource within 60s of start of 911 call conversation. “Dispatcher Impression” (DI) and on scene paramedic “Provider Impression” (PI) were compared, and other care metrics analyzed. Results: In total, 12,528 runs were evaluated, of which 79.1% (9913) were dispatched as stroke and paramedics identified stroke in 47.6%. The sensitivity of dispatcher impression of stroke was 56.2% and the positive predictive value was 33.9%. The most common paramedic impressions in non-concordant patients dispatched as stroke were: generalized weakness in 24% (1587), altered mental status in 17% (1111), and syncope in 6% (403). In patients with a dispatcher impression of non-stroke but paramedic impression of stroke, the most common dispatcher impressions were: sick person in 29% (758), unconscious in 28.7% (750), and seizure in 4.6% (121). In dispatcher impressions of stroke, 100% were dispatched at ALS level; and in dispatcher impressions of non-stroke but eventual paramedic impression of stroke, 98.3% were dispatched at ALS level. Patient age over/under 40 & patient sex did not modify rates of dispatcher - paramedic impression concordance. Conclusions: Dispatchers identify more than half of paramedic-identified stroke patients and allocate appropriate ALS standard ambulances in almost all stroke patients. However, for efficient attaching of specialized MSU resources, improved dispatcher identification of stroke is desirable.