移动平均模型预警传染病的建立与应用研究  

Study on Establishment and Application of Moving Average and Linear Regression Model for Predicting Communicable Diseases

作　　者：熊新平[1] 龙术国 唐雨新[1] 李竹清 张宇[1] 

机构地区：[1]湖南省株洲市疾病预防控制中心,湖南株洲412000

出　　处：《实用预防医学》2012年第9期1336-1338,共3页Practical Preventive Medicine

摘　　要：目的探索新的传染病预警模型,评价预警效果。方法联合运用移动平均和线性回归模型,使用株洲市2004-2009年流行性腮腺炎发病数据,建立以周为时间序列的发病模型,计算预警阈值,制作预警控制图,并用灵敏度、特异度、及时性评价预警效果。结果该模型全部预警到株洲市2010-2011年流行性腮腺炎的暴发和流行,灵敏度100%、特异度91.67%,较监测系统早2~6周预警暴发和流行,预警阈值范围在均数加1.6~2.3倍标准差之间。结论联合移动平均和线性回归模型,具有较高的预警灵敏度、特异度和及时性,适合预警有周期性和季节性的常见传染病。Objective To explore a new early warning model for communicable diseases, and to evaluate the effect of the early warning model. Methods Moving average model and linear regression model were combined applied in the study. The data concerning mumps occurred in Zhuzhou from 2004 to 2009 were used to establish time sequence analysis model based on weekly - onset. The warning threshold value was calculated. Early warning and control chart was designed, and the effect of ear ly warning model was evaluated through assessing its sensitivity, specificity, and timeliness. Results The early warning mod- el warned the outbreaks and epidemics of mumps in Zhuzhou during the period of 2010 - 2011, its sensitivity and specificity were 100% and 91.67%, respectively. It warned the outbreaks 2 - 6 weeks earlier than that of disease monitoring reporting system, and the warning threshold value ranged between the average plus 1.6 - 2.3 times of standard deviation. Conclusions Moving average and linear regression model has the advantages of higher sensitivity, specificity and timeliness in early warning. It is suit- able for early warning periodic and seasonal communicable diseases.

关 键 词：传染病 预警 移动平均 

分 类 号：R181.8[医药卫生—流行病学]