Assessing the Effectiveness of Cumulative Sum Normal- and Poisson-Based Tests for Detecting Rare Diseases
Abstract
The early detection of a rare disease outbreak is of vital importance in public health and national defense. The six Category A biological agents designated by the Centers for Disease Control and Prevention are causal agents of rare diseases. The Francisella tularensis is one of these and is the causal agent of the tularemia disease. Tularemia is used as the motivating problem to evaluate and compare the effectiveness of the normal and Poisson-based CUSUM in the early detection of an outbreak using simulated rare disease data based on its theoretical behavior at varying occurrence outbreak means. In this study, a mean relationship between the nonoutbreak 0 and outbreak 1 means was found, 1 = 7.40 0. Simulations were run to study the mean relationship in three cases: one theoretical case, where the normal- and Poisson-base CUSUM are exactly equal; and two other extreme cases. The computational results show that when 1 is very close to 0 the normal-based CUSUM system behaves improperly resulting in early detection delays, when 1 is equal or greater than 0 the normal- and Poisson-based CUSUM behave almost equally if the threshold for the normal-based CUSUM is selected properly. Methods to determined proper thresholds are also given.
Document Details
- Document Type
- Technical Report
- Publication Date
- Dec 01, 2010
- Accession Number
- ADA536104
Entities
People
- Manuel E. Gamaza
Organizations
- Naval Postgraduate School