Swenson DJ, Ed M, Taylor C, Southworth J, Default K
Surveillance investigation tool development targeted for results
Advances in Disease Surveillance
This paper details the development of electronic surveillance tools by Communicable Disease Surveillance, which have increased detection and investigation capabilities. BACKGROUND The Automated Hospital Emergency Department Data (AHEDD) System is designed to detect early indicators of bioterrorism events and naturally occurring public health threats. Four investigatory tools have been developed with drill-down detail reporting: 1. Syndromic Alerting, 2. Chief Complaint Data Mining 3. ICD9 Code Disease, and 4. Influenza-Like-Illness (ILI) Tracking. All analysis processing runs on the server in seconds using ORACLE PL/SQL stored procedures and arrays. METHODS Seven NH hospitals transmit real-time Emergency Department data securely using HL7 format via Virtual Private Network (VPN) to the NH Division of Public Health Services data repository. Vendor created syndrome alerts were developed with the RODS syndrome classifier CoCo , standard statistics, and CUSUM and Zhang , methodologies) to detect mid to large disease clusters. Alerts occur when counts exceed 3 standard deviations over baselines for the same month previous year/s. Chief complaint data text mining detects over 140 infectious diseases using queries with exclusions, array groupings, and data values within these groupings. Likewise, non-reportable Influenza-Like-Illness (ILI) encounters are detected similarly and charted as a percent of total encounters by MMWR week. Figure 1 – The chart above compares ILI non-reportable query encounters against Flu Sentinel Reports as a percentage of total hospital visits for the 2006-7 Influenza season.