A Statistical Algorithm for the Early Detection of Outbreaks of Infectious Disease

TLDR

Early detection of infectious disease outbreaks is essential for effective control, and automated analysis of large surveillance data can supplement traditional methods when accounting for diverse infection patterns. The study presents a robust system for processing weekly infection reports at the Communicable Disease Surveillance Centre. The system uses a simple regression algorithm to set thresholds and flags organisms that exceed them for further investigation.

Abstract

SUMMARY Outbreaks of infectious diseases must be detected early for effective control measures to be introduced. When dealing with large amounts of data, automated procedures can usefully supplement traditional surveillance methods, provided that the wide variety of patterns and frequencies of infections are taken into account. This paper describes a robust system developed to process weekly reports of infections received at the Communicable Disease Surveillance Centre. A simple regression algorithm is used to calculate suitable thresholds. Organisms exceeding their threshold are then flagged for further investigation.