Markup Language) format, and monitoring of the ... monitoring selected clinical illnesses in the emer- ... or terms identifiedby keyword searches of "chief.
Syndromic Surveillance For Bioterrorism Using Automated Collection of Computerized Discharge Diagnoses Lober WB', Trigg LJ', Karras BTV, Bliss D", Ciliberti Jd, Stewart L', Duchin JSbc aClinical Informatics Research Group, bSchool ofMedicine, University of Washington, Seattle WA cPublic Health - Seattle & King County, Seattle, WA, dOverlake Hospital Medical Center, Bellevue, WA The Syndromic Surveillance Information Collection (SSIC) system aims to facilitate early detection of bioterrorism (BT) attacks through automated data collection of visit level discharge diagnoses from heterogeneous clinical information systems, integration of those data into a common XML (Extensible Markup Language) format, and monitoring of the results to detect unusual patterns of illness in the population. The system, operational since January 2001, collects, integrates, and displays data from three emergency and urgent care (ED/UC) departments and nine primary care clinics, by automatically gathering data from the information systems of those facilities. MeSH: Bioterrorism, Biological warfare, Sentinel Surveillance, Informatics, Information Systems, Database, Data Collection Introduction: Experience with a manual, clinicianbased (versus automated, information system-based) Emergency Department (ED) surveillance system employed during the 1999 World Trade Organization (WTO) Ministerial in Seattle suggested that syndromic surveillance is a potentially useful means of monitoring selected clinical illnesses in the emergency department setting'. Objectives: Optimal response to BT attacks is predicated on the reliability of early detection methods. The goal of the SSIC project is to develop and support an automated, countywide surveillance system that will facilitate early detection of BT attacks. Specifically, we: 1) employ automated collection of data from heterogeneous source systems; 2) normalize clinical syndromic data and store it in a centralized database; 3) provide secure, remote access to this data for public health staff using aberration detection software; and 4) characterize baseline frequencies of certain diagnoses and diagnostic clusters. Methods: Several candidate agents for potential BT attacks are characterized by aerosol dispersion resulting in acquisition of infection by inhalation. We monitor data on patient visits to the participating emergency departments and primary care clinics for the occurrence of either "sentinel" ICD-9 diagnoses, or terms identified by keyword searches of "chief complaint" fields. System Overview: The Syndromic Surveillance Information Collection System currently collects data
AMIA 2002 Annual Symposium Proceedings
on sentinel events from four clinical information systems, transmits it to a centralized cluster of servers, where it is normalized to a common format, and then stored in a relational database. These uniform, multisite data are then made available for secure queries from specific workstations, on which aberration detection software is run.
Data are coDlected, stored, and can be queried centrally
Results: The system has been running since June 2001. We currently collect daily data from all sites and have obtained historical data, which at two sites extends back 4 years. The database contains records of 60,000 patient visits resulting in a report of a sentinel event, from all sites, though July 2002. Conclusions: We have made substantial progress in building an infrastructure to automate syndromic surveillance, demonstrated a series of technologies to collect data from heterogeneous information systems, and have over a year's worth of experience with this deployed system. Evaluation ofthese systems remains a challenge, but we are presently exploring evaluation based both on descriptive criteria2 and on proxy disease detection. Acknowledgements: Centers for Disease Control and Prevention State BT Preparedness Grants (B2 section) U90/CCUO17010-02. Further information:
lobera,washington.edu. ieff.duchinametrokc.gov Www.cirg.washington.edu 1
Plough A. WTO enhanced surveillance project local and national collaboration leads to success. EPI-LOG Communicable Disease and Epidemiology News. Dec. 1999;(39)12. 2 CDC. Updated guidelines for evaluating surveillance systems: recommendations from the guidelines working group. MMWR 2001;50 [no. RR-13]
1088
