Multi-syndrome analysis of time series using PCA: A new concept for outbreak investigation
- 20 December 2007
- journal article
- research article
- Published by Wiley in Statistics in Medicine
- Vol. 26 (29), 5203-5224
- https://doi.org/10.1002/sim.2872
Abstract
To date, despite widespread availability of time series data on multiple syndromes, multivariate analysis of syndromic data remains under‐explored. We present a non‐parametric multivariate framework for early detection of temporal anomalies based on principal components analysis of historical data on multiple syndromes. We introduce simulated outbreaks of different shapes and magnitudes into the historical data, and compare the detection sensitivity and timeliness of the multi‐syndrome detection method with those of uni‐syndrome. We find that the multi‐syndrome detection framework provides a powerful tool for identifying such designated abnormalities in the data and significantly improves upon the detection sensitivity and timeliness of uni‐syndrome analysis. The proposed multivariate framework requires minimal preprocessing of the data and can be easily adopted in settings where temporal information on multiple syndromes are routinely collected and processed, and thus can be an integral component of real‐time surveillance systems. Copyright © 2007 John Wiley & Sons, Ltd.Keywords
This publication has 15 references indexed in Scilit:
- Multivariate scan statistics for disease surveillanceStatistics in Medicine, 2007
- A susceptible-infected model of early detection of respiratory infection outbreaks on a background of influenzaJournal of Theoretical Biology, 2006
- A Space–Time Permutation Scan Statistic for Disease Outbreak DetectionPLoS Medicine, 2005
- Technical Description of RODS: A Real-time Public Health Surveillance SystemJournal of the American Medical Informatics Association, 2003
- Biosurveillance applying scan statistics with multiple, disparate data sourcesJournal of Urban Health, 2003
- Using temporal context to improve biosurveillanceProceedings of the National Academy of Sciences of the United States of America, 2003
- Disease outbreak detection system using syndromic data in the greater Washington DC areaAmerican Journal of Preventive Medicine, 2002
- Use of Automated Ambulatory-Care Encounter Records for Detection of Acute Illness Clusters, Including Potential Bioterrorism EventsEmerging Infectious Diseases, 2002
- A Double-Scan Statistic for Clusters of Two Types of EventsJournal of the American Statistical Association, 1997
- Detection of aberrations in the occurrence of notifiable diseases surveillance dataStatistics in Medicine, 1989