Current methods for influenza surveillance include laboratory confirmed case reporting, sentinel physician reporting of Influenza-Like-Illness (ILI) and chief-complaint monitoring from emergency departments (EDs).
The current methods for monitoring influenza have drawbacks. Testing for the presence of the influenza virus is costly and delayed. Specific, sentinel physician reporting is subject to incomplete, delayed reporting. Chief complaint (CC) based surveillance is limited in that a patient’s chief complaint will not contain all signs and symptoms of a patient.
A possible solution to the cost, delays, incompleteness and low specificity (for CC) in current methods of influenza surveillance is automated surveillance of ILI using clinician-provided free-text ED reports.
Objective
This paper describes an automated ILI reporting system based on natural language processing of transcribed ED notes and its impact on public health practice at the Allegheny County Health Department.