Estimation of Serial Interval and Reproductive Number with Geographic Information： Applications of Enterovirus Syndromic Surveillance

• Main Authors: Shih-Wan Chou, 周詩婉
• Other Authors: I-Feng Lin
• Format: Others
• Language: zh-TW
• Published: 2012
• Online Access: http://ndltd.ncl.edu.tw/handle/76555330353097121984

碩士 === 國立陽明大學 === 公共衛生研究所 === 100 === Background and Objectives During the past years, effective reproduction number(Rt ) has been commonly used as a measure of severity of an outbreak of infectious diseases. A previous study integrating temporal and geographical information which modified the Wallinga and Teunis’s (2004) method assumed the serial interval a known parameter. The Rt has seldom been used in smaller areas. The purposes of this study were to estimate the serial interval using empirical data and to estimate the Rt integrating the temporal and geographical information. The method was applied in an Enterovirus syndromic surveillance and regional Rt were plotted on Taiwan map. Methods The empirical study used Enterovirus syndromic surveillance data from Real-time Outbreak and Disease Surveillance (RODS) database in 2009. The serial interval was estimated using maximum likelihood estimator assuming a Gamma or a Weibull distributions using the SAS NonLinear Programming (NLP) Procedure. Then a weight was assigned to each of relative likelihoods according to the distances between pairs of cases, where the weights decreased as the distances between cases increased. The Rt using the estimated serial intervals then were incorporated onto Taiwan map using the SAS GMAP procedure. Results and Conclusion Our empirical study from the syndromic surveillance has shown that the mean (standard deviation, SD) of the estimated serial interval assuming a Gamma distribution were 5.5(2.2) days, which was slightly greater than those assuming a Weibull disgtribution with mean (SD) as 4.1(2.2) days. Both estimates were slightly greater than the mean (SD) as 3.7(2.6) days from a previous study (Chang 2004) which was based on a 2001 household survey on Enterovirus71. After integrating the calculated serial intervals and geographical distances between each pair of possible infected individuals, the estimated small-area Rt were plotted in the Taiwan map. This graphical presentation was more sensitive and more informative than the single value of Rt summed from the whole Taiwan.