青海省肺结核发病与气象因素的时间序列分析

梁达; 商越; 王兆芬; 马斌忠

doi:10.16462/j.cnki.zhjbkz.2021.10.013

青海省肺结核发病与气象因素的时间序列分析

doi: 10.16462/j.cnki.zhjbkz.2021.10.013

梁达¹,
商越¹,
王兆芬¹,
马斌忠^{1, 2, ,}

1.
810001 西宁，青海大学医学院公共卫生系流行病学教研室
2.
810007 西宁，青海省疾病预防控制中心传染病预防控制所

基金项目:

国家自然科学基金 81860593

青海省高端创新人才千人计划 2017-2

青海省卫生健康委员会适宜技术推广 2020-wjtg-02

详细信息

通讯作者:
马斌忠，E-mail：qhxnmbz@126.com

中图分类号: R181
计量
- 文章访问数: 351
- HTML全文浏览量: 242
- PDF下载量: 80
- 被引次数: 0
出版历程
- 收稿日期: 2021-01-14
- 修回日期: 2021-03-20
- 网络出版日期: 2021-11-17
- 刊出日期: 2021-10-10

The relationship between incidence of pulmonary tuberculosis and meteorological factors in Qinghai Province and multivariate time series analysis

1.
Epidemiology Teaching and Research Office, Department of Public Health, Medical College, Qinghai University, Xining 810001, China
2.
Institute for Prevention and Control of Infectious Diseases, Qinghai Center for Disease Control and Prevention, Xining 810007, China

Funds:

National Natural Science Foundations of China 81860593

Qinghai Provincial Thousand Talents Program for High-level Innovative Talents 2017-2

Appropriate Technology Promotion Project of Health Commission of Qinghai Province 2020-wjtg-02

More Information

Corresponding author: MA Bin-zhong, E-mail: qhxnmbz@126.com

摘要

摘要: 目的探索气象因素与青海省肺结核发病之间的关系，并构建ARIMAX(autoregressive integrated moving average model-X, ARIMAX)模型对肺结核病例数进行短期预测。方法利用地理加权回归(geographically weighted regression, GWR)分析气象因素对肺结核发病的影响。以2014―2018年肺结核月发病数作为响应序列，气象因素作为输入序列，通过互相关函数(cross-correlatioan function, CCF)图确定与肺结核发病有关的气象因素并构建ARIMAX模型，分别回代拟合和预测2014―2018年肺结核月发病数和2019年肺结核月发病数，并与实际发病数相比较。结果降水量和相对湿度对肺结核发病存在正向影响，气压、气温和日照时数为负向影响，风速既存在正向影响，也存在负向影响；通过CCF图确定平均气温和平均风速与肺结核发病存在相关，并最终建立纳入2个协变量(平均气温3阶滞后和平均风速2阶滞后)的ARIMAX(0, 1, 2)×(0, 1, 0)₁₂最优模型，通过回代拟合和预测得出该模型拟合优度(R²)为0.71，平均绝对百分比误差(mean absolute percentage error, MAPE)为24.91%。结论气象因素在不同程度上影响着青海省肺结核的发病，本研究建立的最优ARIMAX模型可用于预测肺结核发病。
- 肺结核 /
- 地理加权回归 /
- 气象因素 /
- ARIMAX模型 /
- 时间序列分析
Abstract: Objective To explore the relationship between meteorological factors and the incidence of pulmonary tuberculosis (PTB) in Qinghai Province, and establish a autoregressive integrated moving average model-X (ARIMAX) model to make a short-term prediction of the number of PTB cases. Methods Geographically weighted regression (GWR) was applied to analyze the influence of meteorological factors on the incidence of PTB. The monthly number of PTB cases in Qinghai Province from 2014 to 2018 was used as the response sequence and meteorological factors as the input sequence, the meteorological factors related to the incidence of PTB were determined by the cross-correlation function (CCF) diagram. The ARIMAX model was established to fit and predict the monthly number of PTB cases from 2014 to 2018 and 2019 respectively, and compared with the actual monthly cases. Results Precipitation and relative humidity had a positive effect on the incidence of PTB, air pressure, temperature and sunshine hours had a negative effect, while the wind speed had both positive and negative effects. The correlation between the average temperature, wind speed and the incidence of PTB was determined by CCF diagram. The optimal model established was ARIMAX(0, 1, 2)×(0, 1, 0)₁₂ with two covariables (the third order lag of average temperature and the second order lag of average wind speed). The model has a goodness of fit (R²) of 0.71 and mean absolute percentage error (MAPE) was 24.91%. Conclusions Meteorological factors affected the incidence of PTB to different degrees. An optimal ARIMAX model was established to predict the incidence of PTB.
- Pulmonary tuberculosis /
- Geographically weighted regression /
- Meteorological factors /
- ARIMAX model /
- Time series analysis

HTML全文

图 1 2014―2019年青海省肺结核年平均报告发病率地理分布

Figure 1. Geographical distribution diagram of annual average reported incidence of PTB in Qinghai Province from 2014 to 2019

下载: 全尺寸图片幻灯片

图 2 2014―2019年青海省肺结核病例时间分布

Figure 2. Temporal distribution of PTB cases in Qinghai Province from 2014 to 2019

下载: 全尺寸图片幻灯片

图 3 2014―2019年青海省各气象因素月均值雷达图

Figure 3. Radar diagram of monthly mean of meteorological factors in Qinghai Province from 2014 to 2019

下载: 全尺寸图片幻灯片

图 4 2014―2019年青海省各气象因素回归系数分布图

Figure 4. Distribution diagram of regression coefficients of meteorological factors in Qinghai Province from 2014 to 2019

下载: 全尺寸图片幻灯片

图 5 2014―2019年青海省各气象因素与肺结核发病互相关函数图

Figure 5. Distribution diagram of regression coefficients of meteorological factors in Qinghai Province from 2014 to 2019

下载: 全尺寸图片幻灯片

图 6 最优ARIMAX模型拟合和预测图

Figure 6. Optimal ARIMAX model fitting and prediction diagram

下载: 全尺寸图片幻灯片

表 1 2014―2019年青海省肺结核空间自相关分析

Table 1. Spatial autocorrelation analysis of PTB in Qinghai Province from 2014 to 2019

年份	Moran’s I值	Z值	P值
2014	0.107 8	4.012 7	＜0.001
2015	0.071 7	2.894 7	0.003
2016	0.130 3	4.704 2	＜0.001
2017	0.193 0	6.565 4	＜0.001
2018	0.149 4	5.302 1	＜0.001
2019	0.298 3	9.741 8	＜0.001

下载: 导出CSV

表 2 ARIMAX模型参数估计、检验和模型诊断

Table 2. Parameter estimation, test and model diagnosis of ARIMAX model

模型	气象因素		β值	s_x值	t值	P值	AIC值	BIC值	残差序列白噪声检验(P值)
模型	变量	滞后阶数(lag)	β值	s_x值	t值	P值	AIC值	BIC值	残差序列白噪声检验(P值)
ARIMA(0, 1, 2)×(0, 1, 0)₁₂	平均气温	3	2.791	0.959	2.912	0.003	636.92	645.16	0.788
ARIMA(0, 1, 2)×(0, 1, 0)₁₂	平均气温	9	2.736	1.011	2.707	0.004	637.90	646.14	0.804
ARIMA(0, 1, 2)×(0, 1, 0)₁₂	平均风速	2	74.647	30.146	2.476	0.008	638.52	646.76	0.802
ARIMA(0, 1, 2)×(0, 1, 0)₁₂	平均气温	3	2.260	0.844	2.679	0.005	634.29	644.59	0.807
ARIMA(0, 1, 2)×(0, 1, 0)₁₂	平均风速	2	72.383	27.074	2.674	0.005	634.29	644.59	0.807
ARIMA(0, 1, 2)×(0, 1, 0)₁₂	平均气温	9	2.237	0.841	2.659	0.005	634.47	644.77	0.826
ARIMA(0, 1, 2)×(0, 1, 0)₁₂	平均风速	2	74.854	26.928	2.780	0.004	634.47	644.77	0.826

下载: 导出CSV

参考文献(16)

[1]	Greer A, Ng V, Fisman D. Climate change and infectious diseases in North America: the road ahead [J]. CMAJ, 2008, 178(6): 715-722. DOI: 10.1503/cmaj.081325.
[2]	李金娟, 裴凌云, 赵翊, 等. 兰州市肺结核病发病与气象因素关系研究[J]. 疾病预防控制通报, 2019, 34(4): 10-13. DOI: 10.13215/j.cnki.jbyfkztb.1905015. Li JJ, Pei LY, Zhao Y, et al. Study on relationship between incidence of pulmonary tuberculosis and meteorological factors in Lanzhou [J]. Bull Dis Control Prev, 2019, 34(4): 10-13. DOI: 10.13215/j.cnki.jbyfkztb.1905015.
[3]	江伟. 昆山市肺结核流行病学特征及其与气象因素关系分析[D]. 苏州: 苏州大学, 2014. http://cdmd.cnki.com.cn/article/cdmd-10285-1015536834.htm Jiang W. Epidemiological characteristics of pulmonary tuberculosis and its relationships with the meteorological factors in Kunshan City [D]. Soochow: Soochow University, 2014. http://cdmd.cnki.com.cn/article/cdmd-10285-1015536834.htm
[4]	郑能雄, 林云钦, 郑高, 等. 福州市主要传染病与气象因素关系的研究[J]. 中国公共卫生管理, 2004, 20(4): 343-346. DOI: 10.19568/j.cnki.23-1318.2004.04.034. Zheng NX, Lin YQ, Zheng G, et al. Study on the relationship between main infectious diseases and meteorological factors in Fuzhou [J]. Chin J Public Heal Manag, 2004, 20(4): 343-346. DOI: 10.19568/j.cnki.23-1318.2004.04.034.
[5]	王朝才, 刘燕, 晁秀珍, 等. 2006-2015年青海省肺结核流行特征分析[J]. 现代预防医学, 2017, 44(3): 397-400. https://www.cnki.com.cn/Article/CJFDTOTAL-XDYF201703004.htm Wang ZC, Liu Y, Chao XZ, et al. Epidemiological characteristics of pulmonary tuberculosis of Qinghai, 2006-2015 [J]. Mod Prev Med, 2017, 44(3): 397-400. https://www.cnki.com.cn/Article/CJFDTOTAL-XDYF201703004.htm
[6]	张松林, 张昆. 全局空间自相关Moran指数和G系数对比研究[J]. 中山大学学报(自然科学版), 2007, 46(4): 93-97. DOI: 10.3321/j.issn:0529-6579.2007.04.021. Zhang SL, Zhang K. A comparative study of Moran index and G coefficient of global spatial autocorrelation [J]. Acta Sci Nat Univ Sunyatseni, 2007, 46(4): 93-97. DOI: 10.3321/j.issn:0529-6579.2007.04.021.
[7]	时涛, 李士雪. 地理加权回归模型在卫生经济领域的应用举例: 基于ArcGIS 10.4软件介绍[J]. 中国卫生经济, 2018, 37(10): 10-14. DOI: 10.7664/CHE20181002. Shi T, Li SX. Application of geographically weighted regression model in the economics of health and health care: based on the introduction of ArcGIS 10.4 [J]. Chin Heal Econ, 2018, 37(10): 10-14. DOI: 10.7664/CHE20181002.
[8]	王前, 李涛, 杜昕, 等. 2015-2019年全国肺结核报告发病情况分析[J]. 中国防痨杂志, 2021, 43(2): 107-112. DOI: 10.3969/j.issn.1000-6621.2021.02.002. Wang Q, Li T, Du X, et al. The analysis of national tuberculosis reported incidence and mortality, 2015-2019 [J]. Chin J Antituberc, 2021, 43(2): 107-112. DOI: 10.3969/j.issn.1000-6621.2021.02.002.
[9]	李世雄, 王玉明, 董继元. 嘉峪关市肺结核病与气象相关影响因素研究[J]. 疾病预防控制通报, 2019, 34(1): 5-9, 41. DOI: 10.13215/j.cnki.jbyfkztb.1811002. Li SX, Wang YM, Dong JY. Study on influential factors of meteorology correlated to tuberculosis in Jiayuguan City [J]. Bull Dis Control Prev, 2019, 34(1): 5-9, 41. DOI: 10.13215/j.cnki.jbyfkztb.1811002.
[10]	邓斌, 周志刚, 马泽粦, 等. 肺结核病与气象因素关系的BP神经网络模型研究[J]. 国际医药卫生导报, 2008, 14(1): 17-20. DOI: 10.3760/cma.j.issn.1007-1245.2008.01.004. Deng B, Zhou ZG, Ma ZL, et al. The model of back-propagation neural network about the relationship between meterological factors and pulmonary tuberculosis [J]. Int Med Heal Guid News, 2008, 14(1): 17-20. DOI: 10.3760/cma.j.issn.1007-1245.2008.01.004.
[11]	李林林. 湖南省肺结核发病率及气象因素分析[D]. 济南: 山东大学, 2013. http://cdmd.cnki.com.cn/Article/CDMD-10422-1013219716.htm Li LL. The incidence of tuberculosis and the influence of related meteorological factors in Hunan Province [D]. Jinan: Shandong University, 2013. http://cdmd.cnki.com.cn/Article/CDMD-10422-1013219716.htm
[12]	李盛, 马汉平, 王金玉, 等. 兰州市气温和风速对肺结核发病的影响研究[J]. 环境卫生学杂志, 2018, 8(2): 91-98. DOI: 10.13421/j.cnki.hjwsxzz.2018.02.003. Li S, Ma HP, Wang JY, et al. Effect of air temperature and wind speed on the incidence of tuberculosis in Lanzhou [J]. J Environ Hyg, 2018, 8(2): 91-98. DOI: 10.13421/j.cnki.hjwsxzz.2018.02.003.
[13]	龚风云, 王凯, 樊旭成, 等. 乌鲁木齐市流感样病例与气象因素的ARIMAX模型预测分析[J]. 公共卫生与预防医学, 2020, 31(2): 4-8. DOI: 10.3969/j.issn.1006-2483.2020.02.002. Gong FY, Wang K, Fan XC, et al. Prediction and analysis of influenza-like illness and meteorological factors by ARIMAX model in Urumqi [J]. J Pub Health and Prev Med, 2020, 31(2): 4-8. DOI: 10.3969/j.issn.1006-2483.2020.02.002.
[14]	陈佳, 谢娜, 邓晟, 等. 新疆喀什百日咳与气象因素的多元时间序列分析[J]. 职业与健康, 2018, 34(13): 1834-1839. DOI: 10.13329/j.cnki.zyyjk.2018.0507. Chen J, Xie N, Deng S, et al. Multivariate time series analysis of pertussis and meteorological factors in Kashi of Xinjiang [J]. Occup and Health, 2018, 34(13): 1834-1839. DOI: 10.13329/j.cnki.zyyjk.2018.0507.
[15]	李虎玲. 新疆喀什地区结核病动力学模型与时间序列模型的建立与分析[D]. 乌鲁木齐: 新疆医科大学, 2019. Li HL. Establishment and analysis of tuberculosis dynamics model and time series model in Kashgar, Xinjiang [D]. Urumqi: Xinjiang Medical University, 2019.
[16]	妥小青, 张占林, 龚政, 等. 基于ARIMAX模型的乌鲁木齐市流感样病例预测分析[J]. 中华疾病控制杂志, 2018, 22(6): 590-593. DOI: 10.16462/j.cnki.zhjbkz.2018.06.012. Tuo XQ, Zhang ZL, Gong Z, et al. Forecasting influenza like illness in Urumqi based on ARIMAX model [J]. Chin J Dis Control Prev, 2018, 22(6): 590-593. DOI: 10.16462/j.cnki.zhjbkz.2018.06.012.