-
摘要:
目的 对我国布鲁菌病(简称布病)月发病率进行预测,为了解我国布病流行趋势、制定防控策略提供数据支持和决策依据。 方法 以国家人口与健康科学数据共享平台为数据来源,使用2004年1月-2016年12月全国布病月发病率数据建立历史序列,应用R软件构建自回归滑动平均混合模型(autoregressive integrated moving average model,ARIMA)并进行数据拟合和预测。 结果 本研究构建乘积季节ARIMA(2,1,2)(2,1,1)12模型各项参数都有统计学意义(均有P < 0.001),模型很好的拟合了全国布病月发病率的变化规律,预测值与实际值之间的平均相对误差为21.77%;预测2017年、2018年、2019年和2020年布病的月平均发病率分别为0.399 5/10万、0.423 8/10万、0.445 6/10万、0.471 2/10万,呈逐渐增高趋势(χ2=14.244,P < 0.001),在4-7月份出现发病峰值。 结论 在自然状况下,我国人间布病的月发病率将逐年增高,应采取相应措施进行控制。 Abstract:Objective The aims is to predict the monthly incidence of brucellosis in China, in order to understand the epidemic trend of brucellosis in China, to formulate prevention and control strategies, and to provide data support and decision-making basis. Methods The national population and health science data sharing platform was used to collect the national incidence of brucellosis from January 2004 to December 2016 by month. The data were fitted and predicted using ARIMA model with R software. Results In this study, the parameters of the product season ARIMA (2, 1, 2) (2, 1, 1)12 model had statistical significance (all P < 0.001). The model fitted well the monthly incidence of brucellosis in China. The average relative error between the predicted value and the actual value was 21.77%. The monthly average incidence of brucellosis in 2017, 2018, 2019 and 2020 were predicted to be 0.399 5/100 000, 0.423 8/100 000, 0.445 6/100 000 and 0.471 2/100 000 respectively, showing a gradually increasing trend (χ2=14.244, P < 0.001), with a peak incidence from April to July. Conclusion Under natural conditions, the monthly incidence of human brucellosis in China will increase year by year, and corresponding measures should be taken to control it. -
Key words:
- Brucellosis /
- Time series /
- ARIMA model
-
表 1 ARIMA(p, d, q) (P, D, Q)s模型的组合与最优模型的选取准则检验
Table 1. Combination of ARIMA (p, d, q) (P, D, Q)s model and selection criteria test of optimal model
模型指标 AIC RMSE MAE MAPE ARIMA(1, 1, 2)(1, 1, 1) -0.023 7 0.042 3 0.021 7 12.597 ARIMA(2, 1, 1)(2, 1, 1) -0.003 0 0.041 5 0.021 1 12.199 ARIMA(2, 1, 2)(2, 1, 1) -0.003 8 0.041 5 0.021 1 12.197 ARIMA(2, 1, 1)(1, 1, 1) -0.003 3 0.042 2 0.021 6 12.467 表 2 ARIMA (2, 1, 2, ) (2, 1, 1)12模型对全国布病发病率的拟合结果
Table 2. Fitting results of ARIMA (2, 1, 2, ) (2, 1, 1)12 model on incidence of brucellosis in China
时间 实际值 预测值(95%CI)值 2015年1月 0.303 6 0.406(0.283, 0.530) 2015年2月 0.293 8 0.321(0.190, 0.453) 2015年3月 0.409 9 0.413(0.277, 0.550) 2015年4月 0.493 1 0.461(0.322, 0.599) 2015年5月 0.565 7 0.510(0.371, 0.649) 2015年6月 0.532 6 0.477(0.337, 0.616) 2015年7月 0.456 8 0.407(0.267, 0.546) 2015年8月 0.350 0 0.328(0.188, 0.468) 2015年9月 0.228 4 0.233(0.093, 0.373) 2015年10月 0.198 2 0.219(0.079, 0.358) 2015年11月 0.184 7 0.210(0.071, 0.350) 2015年12月 0.165 9 0.196(0.056, 0.336) 2016年1月 0.251 6 0.395(0.252, 0.538) 2016年2月 0.264 2 0.340(0.197, 0.484) 2016年3月 0.356 6 0.433(0.289, 0.577) 2016年4月 0.415 5 0.477(0.333, 0.621) 2016年5月 0.461 3 0.525(0.381, 0.670) 2016年6月 0.416 5 0.489(0.345, 0.633) 2016年7月 0.365 7 0.418(0.274, 0.563) 2016年8月 0.295 9 0.347(0.202, 0.491) 2016年9月 0.187 1 0.253(0.109, 0.397) 2016年10月 0.159 9 0.240(0.095, 0.384) 2016年11月 0.150 3 0.235(0.091, 0.379) 2016年12月 0.135 2 0.221(0.076, 0.365) -
[1] De Massis F, Di Girolamo A, Petrini A, et al. Correlation between animal and human brucellosis in Italy during the period 1997-2002[J]. Clin Microbiol Infect, 2005, 11(8): 632-636. DOI: 10.1111/j.1469-0691.2005.01204.x. [2] Marianelli C, Graziani C, Santangelo C, et al. Molecular epidemiological and antibiotic susceptibility characterization of Brucella isolates from humans in Sicily, Italy[J]. J Clin Microbiol, 2007, 45(9): 2923-2928. DOI: 10.1128/JCM.00822-07. [3] Memish ZA, Balkhy HH. Brucellosis and international travel[J]. J Travel Med 2004, 11(1): 49-55. DOI: 10.2310/7060.2004.13551. [4] Zhong Z, Yu S, Wang X, et al. Human brucellosis in the People's Republic of China during 2005-2010[J]. Int J Infect Dis, 2013, 17(5): e289-e292. DOI: 10.1016/j.ijid.2012.12.030. [5] Box GE, Jenkins GM. Time series analysis: forecasting and control rev. ed[J]. Journal of Time Series Analysis, 2010, 31(4): 303-303. DOI: 10.1111/j.1467-9892.2009.00643.x. [6] 陈田木, 刘如春, 谢知, 等. 流行性感冒数学建模研究进展[J]. 中国热带医学, 2014, 14(7): 890-894. https://www.cnki.com.cn/Article/CJFDTOTAL-RDYX201407042.htmChen TM, Liu RC, Xie Z, et al. Advances in influenza mathematical modeling[J]. Chinese Tropical Medicine, 2014, 14(7): 890-894. https://www.cnki.com.cn/Article/CJFDTOTAL-RDYX201407042.htm [7] 易丹辉. 时间序列分析: 方法与应用[M]. 北京: 中国人民大学出版社, 2011.Yi DH. Time series analysis: methods and applications[M]. Beijing: Renmin University Press, 2011. [8] 张蔚, 张彦琦, 杨旭. 时间序列资料ARIMA季节乘积模型及其应用[J]. 第三军医大学学报, 2002, 24(8): 955-957. DOI: 10.3321/j.issn:1000-5404.2002.08.026.Zhang W, Zhang YQ, Yang X. ARIMA seasonal product model of time series data and its application[J]. Journal of Third Military Medical University, 2002, 24(8): 955-957. DOI: 10.3321/j.issn:1000-5404.2002.08.026. [9] 赖圣杰, 李中杰, 金连梅, 等. 传染病暴发早期预警系统评价内容及其指标[J]. 中华流行病学杂志, 2009, 30(6): 637-640. DOI: 10.3760/cma.j.issn.0254-6450.2009.06.025.Lai SJ, Li ZJ, Jin LM, et al. Evaluation contents and indicators of early warning system for infectious disease outbreaks[J]. Chin J Epidemiol, 2009, 30(6): 637-640. DOI: 10.3760/cma.j.issn.0254-6450.2009.06.025. [10] 陈秋兰. 我国布鲁菌病流行时空分析与患者健康相关生命质量研究[D]. 北京: 中国疾病预防控制中心, 2017.Chen QL. Spatio-temporal analysis of brucellosis epidemic in China and study on health-related quality of life of patients[D]. Beijing: China Center for Disease Control and Prevention, 2017. [11] 朱奕奕, 冯玮, 赵琦, 等. ARIMA乘积季节模型在上海市甲肝发病预测中的应用[J]. 复旦学报(医学版), 2012, 39(5): 460-464. DOI: 10.3969/j.issn.1672-8467.2012.05.004.Zhu YY, Feng W, Zhao Q, et al. Application of ARIMA product seasonal model in prediction of hepatitis A incidence in Shanghai[J]. Fudan Journal (Medical Edition), 2012, 39(5): 460-464. DOI: 10.3969/j.issn.1672-8467.2012.05.004. [12] 王永斌, 李向文, 柴峰, 等. ARIMA模型在我国梅毒发病率预测中的应用[J]. 现代预防医学, 2015, 42(3): 385-388, 417. https://www.cnki.com.cn/Article/CJFDTOTAL-XDYF201503001.htmWang YB, Li XW, Chai F, et al. Application of ARIMA model in predicting syphilis incidence in China[J]. Modern Prev Med, 2015, 42(3): 385-388, 417. https://www.cnki.com.cn/Article/CJFDTOTAL-XDYF201503001.htm [13] 陈秋兰, 杨维中. 内蒙古及相邻省份2010-2014年布鲁氏菌病流行特征和时空聚集性分析[J]. 中华疾病控制杂志, 2017, 21(1): 30-34. DOI: 10.16462/j.cnki.zhjbkz.2017.01.007.Chen QL, Yang WZ, Inner Mongolia and adjacent provinces: epidemiological characteristics and spatial-temporal aggregation analysis of brucellosis from 2010 to 2014[J]. Chin J Dis Control Prev, 2017, 21(1): 30-34. DOI: 10.16462/j.cnki.zhjbkz.2017.01.007. [14] 马有祥, 解光峰, 李静, 等. 2006-2013年东营市人间布鲁氏菌病流行特征分析[J]. 现代预防医学, 2015, 42(14): 2511-2513, 2516. https://www.cnki.com.cn/Article/CJFDTOTAL-XDYF201514007.htmMa YX, Xie GF, Li J, et al. Analysis of epidemiological characteristics of human brucellosis in Dongying City from 2006 to 2013[J]. Modern Prev Med, 2015, 42(14): 2511-2513, 2516. https://www.cnki.com.cn/Article/CJFDTOTAL-XDYF201514007.htm [15] 王水根, 何水渊, 罗学辉. 布鲁氏菌病高危职业人群血清抗体检测结果分析[J]. 浙江预防医学, 2011, 23(9): 53-56. DOI: 10.3969/j.issn.1007-0931.2011.09.020.Wang SG, He SY, Luo XH. Analysis of serum antibody test results in high-risk occupational population of Brucellosis[J]. Zhejiang Preventive Medicine, 2011, 23(9): 53-56. DOI: 10.3969/j.issn.1007-0931.2011.09.020.