Abstract
Since 1998, the New York City Department of Health has used New York City Emergency Medical Services (EMS) ambulance dispatch data to monitor for a communitywide rise in influenzalike illness (ILI) as an early detection system for bioterrorism. A clinical validation study was conducted during peak influenza season at six New York City emergency departments (EDs) to compare patients with ILI brought in by ambulance with other patients to examine potential biases associated with ambulance dispatch-based surveillance. We also examined the utility of 4 EMS call types (selected from 52) for case detection of ILI. Clinical ILI was defined as fever (temperature higher than 100 degrees F) on history or exam, along with either cough or sore throat. Of the 2,294 ED visits reviewed, 522 patients (23%) met the case definition for ILI, 64 (12%) of whom arrived by ambulance. Patients with ILI brought in by ambulance were older, complained of more severe symptoms, and were more likely to undergo diagnostic testing, be diagnosed with pneumonia, and be admitted to the hospital than patients who arrived by other means. The median duration of symptoms prior to presenting to the ED, however, was the same for both groups (48 hours). The selected call types had a sensitivity of 58% for clinical ILI, and a predictive value positive of 22%. Individuals with symptoms consistent with the prodrome of inhalational anthrax were likely to utilize the EMS system and usually did so early in the course of illness. While EMS-based surveillance is more sensitive for severe illness and for illness affecting older individuals, there is not necessarily a loss of timeliness associated with EMS-based (versus ED-based) surveillance.