Algorithms for rapid outbreak detection: a research synthesis

Abstract

The threat of bioterrorism has stimulated interest in enhancing public health surveillance to detect disease outbreaks more rapidly than is currently possible. To advance research on improving the timeliness of outbreak detection, the Defense Advanced Research Project Agency sponsored the Bio-event Advanced Leading Indicator Recognition Technology (BioALIRT) project beginning in 2001. The purpose of this paper is to provide a synthesis of research on outbreak detection algorithms conducted by academic and industrial partners in the BioALIRT project. We first suggest a practical classification for outbreak detection algorithms that considers the types of information encountered in surveillance analysis. We then present a synthesis of our research according to this classification. The research conducted for this project has examined how to use spatial and other covariate information from disparate sources to improve the timeliness of outbreak detection. Our results suggest that use of spatial and other covariate information can improve outbreak detection performance. We also identified, however, methodological challenges that limited our ability to determine the benefit of using outbreak detection algorithms that operate on large volumes of data. Future research must address challenges such as forecasting expected values in high-dimensional data and generating spatial and multivariate test data sets.

A Survey of Operations Research Models and Applications in Homeland Security

“…Jenkins (2000) used integer programming to identify a small subset of oil spills that are similar to all potential categories of spills to predict the type of pollutant a terrorist group might use. Buckeridge et al (2005) classified bioterrorism outbreak algorithms and found that spatial and other covariate information can improve measures for detecting and evaluating outbreaks. Stuart and Wilkening (2005) used first-and second-order catastrophic decay models to study the impact of degradation of biological-weapons agents leaked into the environment.…”

Enhancing Time-Series Detection Algorithms for Automated Biosurveillance

“…We adopted the approach of evaluating sensitivity at a fi xed 1% alert rate defi ned empirically for each algorithm and dataset, as used by Jackson et al (12). Our approach is in accord with a recent review that recommended basing alert thresholds on empirical data rather than on classical statistical theory (17). A major strength of the study is that BioSense is a national system that provided access to 2 major datasets with differing characteristics and to data from hundreds of facilities in many states.…”

Enhancing Biosurveillance Specificity Using PraedicoTM, A Next Generation Application

“…The Praedico™ ML layer was trained by utilizing hundreds of false positive and true positive syndromic alerts. To guarantee high detection recall, the Praedico™ algorithm leverages many known detection algorithms, including versions of CDC, CUSUM, EWMA, and regression models [2]. The ML model combines these models and uses additional time series features to detect anomalies and user feedback received on previous alerts (high confidence alerts).…”