Measures aimed at controlling epidemics of infectious diseases critically benefit from early outbreak recognition [1]. SSS seek early detection by focusing on pre-diagnostic symptoms that by themselves may not alarm clinicians. We have previously determined the performance of various Case Detector (CD) algorithms at finding cases of influenza-like illness (ILI) recorded in the electronic medical record of the Veterans Administration (VA) health system. In this work, we measure the impact of using CDs of increasing sensitivity but decreasing specificity on the time it takes a VA-based SSS to identify a modeled community-wide influenza outbreak. Objective This work uses a mathematical model of a plausible influenza epidemic to test the influence of different case-detection algorithms on the performance of a real-world syndromic surveillance system (SSS).
Case Detection
Objective
To demonstrate how natural language processing (NLP) of clinical records can contribute to case detection and characterization in biosurveillance.
The Real-time Outbreak and Disease Surveillance system collects chief complaints as free text and uses a naïve Bayesian classifier called CoCo to classify the complaints into syndromic categories. CoCo 3.0 has been trained on 28,990 manually clas-sified chief complaints. The free text chief com-plaints are challenging to work with, due to problems caused by linguistic variations such as synonyms, abbreviations, acronyms, truncations, concatenations, misspellings and typographic errors. Failure to correct these word variations may result in missed cases, thereby decreasing sensitivity of detection.
Objective
To determine whether preprocessing chief complaints before automatically classifying them into syndromic categories improves classification performance.