Early Aberration Reporting System Modelling of Korean Emergency Syndromic Surveillance System for Bioterrism.
- Author:
Jae Bong CHUNG
1
;
Moo Eob AHN
;
Hee Cheol AHN
;
Ki Cheol YOU
;
Hyun KIM
;
Jun Whi CHO
;
Young A CHOI
;
Eun Kyeong JEONG
Author Information
1. Department of Emergency Medicine, Hallym University College of Medicine, Chunchon, Korea. skyahn@hallym.or.kr
- Publication Type:Original Article
- Keywords:
Emergency;
Surveillance;
System;
Bioterrorism
- MeSH:
Bioterrorism;
Data Mining;
Emergencies*;
Humans;
Humidity
- From:Journal of the Korean Society of Emergency Medicine
2003;14(5):638-645
- CountryRepublic of Korea
- Language:Korean
-
Abstract:
PURPOSE: This study were designed to supply the opportunity to make a base of emergency syndromic surveillance warning system to detect the bioterrors through the construction of predictive models which were made by reported patients in 'Emergency Syndromic Surveillance System' who were diagnosed as waterborne contagious diseases. METHODS: On this study, we used the neural network analysis methods among the data mining to analyze the reliable variables which was extracted from the reported data bases in the Emergency Syndrome Surveillance System. RESULTS : In this study, we were using the patients data pools from 13th May 2002 to 13th May 2003 in Emergency Syndrome Surveillance System. So we could get the reliable variables - clinical symptoms, severity of patient, humidity and temperature - to predict the waterborne infections. This study shows the successful predictation rate of 96% in error rate of 0.4 with sensible variables through Chisquare analysis and the construction of one hidden layer which is near linearity. CONCLUSION: Early emergency syndromic surveillance warning models made by the neural network in Emergency Syndrome Surveillance System could make the early detection of waterborne infections, could also stop the transmission of waterborne infections in early stage, and furthermore could be used as the preventive and detective methods of bioterror attacks.