目的：研究厦门市细菌性痢疾的流行特征并探讨 SARIMA 模型拟合厦门市菌痢发病趋势预测的可行性。方法利用 SPSS 对厦门市菌痢的发病情况进行流行病学分析,通过取自然对数、差分等方法对菌痢月发病数序列进行平稳化,然后进行模型参数的估计、检验,最优模型的筛选,最后进行预测分析。结果2004-2012年,厦门市菌痢的年平均发病率为15.94/10万,年平均发病率较高的为集美区和思明区,分别为23.69/10万和22.53/10万,其次为海沧区和湖里区,年平均发病率分别为14.19/10万和12.29/10万,而同安区和翔安区的年平均发病率相对较低,分别为7.63/10万和6.81/10万,每年的8-9月为发病高峰,具有明显的周期性,病例主要分布于3岁以下及15~40岁之间,发病数居于前四位的人群包括散居儿童、工人、学生和幼托儿童,占所有病例数的61.12%,SARIMA(0,1,1)(1,1,1)12较好地拟合了厦门市菌痢的月发病数据,预测效果良好。结论厦门市菌痢的发病率较高,可以用 SARIMA 模型进行短期预测,进而指导各项防控措施。%Objective To analyze the epidemic characteristics and test the feasibility of SARIMA model for predicting the incidence of bacillary dysentery in Xiamen city. Methods The incidence of bacillary dysentery in Xiamen was analyzed u-sing SPSS software to acquire the epidemic characteristics. The stabilization of the incidence cases by month was completed by employing difference and natural logarithm. After that, the estimation and test of model parameter and the selection of optimized model was implemented. Results The average incidence rate of bacillary dysentery in Xiamen between 2004 to 2012 was 15. 94 / 100 000. The average incidence rate was higher for Jimei and Siming region, with 23. 69 / 100 000 and 22. 53 / 100 000 respectively, followed by Haicang and Huli region, with 14. 19 / 100 000 and 12. 29 / 100 000 respectively. The average incidence rate in Tongan and Xiangan region was relatively lower, 7. 63 / 100 000 and 6. 81 / 100 000 respectively. The peak time of inci-dence rate was from August to September annually. The cases mainly distributed under age 3 and from 15 to 40 years old. The top four groups of incidence included the scattered children, workers, students and kindergarten children accounting for 61. 12%of all the cases. SARIMA (0,1,1) (1,1,1) 12 was capable of fitting the monthly incidence data of bacillary dysentery. Conclu-sion Due to the high incidence rate of bacillary dysentery in Xiamen, SARIMA would be an appropriate method to perform the short-term forecasting.
展开▼
