An Improved EWMA-Based Method for Outbreak Detection in Multiple Regions

Authors

  • Sesha K. Dassanayaka
  • Joshua French

DOI:

https://doi.org/10.5210/ojphi.v8i1.6517

Abstract

We present a simple, fast, and easily interpretable procedure that results in faster detection of outbreaks in multiple spatial regions. Disease counts from neighboring regions are aggregated to compute a Poisson CUSUM statistic for each region. Instead of controlling the average run length error criterion in the testing process, we instead utilize the false discovery rate. Additionally, p-values are used to make decisions instead of traditional critical-values. The use of the false discovery rate and p-values in testing allows us to utilize more powerful multiple testing methodologies. The procedure is successfully applied to detect the 2011 Salmonella Newport outbreak in Germany.

Downloads

Published

2016-03-24

How to Cite

Dassanayaka, S. K., & French, J. (2016). An Improved EWMA-Based Method for Outbreak Detection in Multiple Regions. Online Journal of Public Health Informatics, 8(1). https://doi.org/10.5210/ojphi.v8i1.6517

Issue

Section

Poster Presentations