湖北省疟疾疫情时空分布特征及预测研究

发布时间：2018-05-23 14:08

本文选题：疟疾 + 时空分析　；参考：《华中科技大学》2015年博士论文

【摘要】：目的：分析湖北省本地疟疾发病的时空分布特征,识别疟疾流行的高危风险区域和高危时段。评价气象因素在湖北省本地疟疾流行中的作用,建立疟疾流行的气象预测模型。探讨ARIMA模型预测湖北省本地疟疾发病率的可行性,预测疟疾发病趋势。本研究可以为湖北省本地疟疾疫情的监测和预测提供决策支持,最终为指导疟疾的防控,合理的分配卫生资源提供科学的参考依据和理论指导。方法：(1)采用Cochran-Armitage趋势检验分析2004～2011年湖北省本地疟疾疫情发展变化的长期趋势。绘制2004～2011年湖北省各县(市、区)本地疟疾年度发病率分布图。(2)采用ArcGIS10.1软件的全局Moran'sI空间自相关分析整个研究区域疟疾发病是否存在空间自相关。采用ArcGIS10.1局部Moran's/空间自相关分析方法和单纯空间扫描统计方法两种局部空间聚集性研究方法来确定2004～2011年湖北省本地疟疾流行的高风险区域。利用单纯时间扫描,时空扫描研究疟疾发病的时间和时空分布特点,确定发病的高风险时段和高风险区域。(3)运用Spearman等级相关分析不同区域尺度疟疾发病率与气象因素相关性。采用多元回归逐步回归分析法筛选影响疟疾发病率气象因素,利用气象因素对疟疾发病变化进行回归拟合。(4)本研究将枣阳市疟疾流行程度按月病人数占全年病人总数的构成比,分为低发月、中发月、高发月。应用逐步判别分析方法利用气象因素对枣阳市未来疟疾流行程度进行判别。(5)采用2004～2009年湖北省本地疟疾发病率构建ARIMA模型,2010年1-12月的数据用于检验模型,并评价模型的拟合及预测效果。结果：(1)湖北省2004～2011年不同年份疟疾发病率整体上呈显著下降的趋势。全局空间自相关分析结果提示湖北省的本地疟疾发病具有一定的空间聚集性。2004～2011年,局部Moran's I空间自相关发现11个疟疾发病高危县,高危县疟疾发病率的中位数从2004年58.1/10万降至2011年0.79/10万；单纯空间扫描分析结果发现2004～2011年湖北省有11个聚集地区,其中一级聚集区8个,二级聚集区3个；时空聚集性分析结果发现一级聚集区域,其中一级聚集区域13个县,聚集时段从2004年4月-2007年11月。(2)湖北省与枣阳市两种区域尺度上2004～2009年的疟疾发病率周期性变化与年中气候的周期变动明显相关,与气温相关指标和降雨量的相关性较为显著,相关系数多在0.7左右。湖北省每个月疟疾发病率的变动78.1%可归因于当月和前一个月平均气温(MeanT-01)和之前2个月的平均最低气温(MinT-2);枣阳市每个月疟疾发病率的变动67.8%可归因于当月和前一个月平均最高气温(MaxT-01)。(3)本研究利用气象因素建立判别函数对枣阳市疟疾流行程度进行判别,首先对32个气象因子进行逐步判别,最终引入判别方程的气象因子有MinT0、MaxT0和D-012。判别函数的准确率为73.61%,具有一定的判别效果。(4)利用湖北省2004～2009年的每月本地疟疾发病率建立模型,结果显示ARIMA(1,1,1)(1,1,0)12模型拟合效果相对最优,预测发病变动趋势与实际发病趋势完全一致,实际值均在预测值的95%可信区间范围,表明模型预测值与实际情况基本一致,拟合效果好。结论：(1)2004～2011年,研究发现湖北省本地疟疾发病率呈显著下降的趋势。湖北省的疟疾发病具有空间聚集性,高危风险地区仍然存在,高危风险地区主要位于嗜人按蚊和中华按蚊复媒疟区。(2)利用气象因素拟合全省和疟疾高发县的疟疾发病率取得较好的效果,构建的判别函数能准确地判断枣阳市疟疾高、中、低发月份的出现时间。(3)构建的ARIMA模型对湖北省疟疾发病情况的拟合结果满意,预测效果良好,可用于预测湖北省未来疟疾的变动趋势。
[Abstract]:Objective: to analyze the temporal and spatial distribution characteristics of malaria in Hubei Province, identify the high-risk area and high risk period of malaria, evaluate the role of meteorological factors in the local malaria epidemic in Hubei Province, establish the meteorological prediction model of malaria epidemic, and discuss the feasibility of the ARIMA model to predict the local malaria incidence in Hubei Province, and predict the malaria incidence. This study can provide decision-making support for the monitoring and prediction of local malaria epidemic in Hubei province. Finally, it provides scientific reference and theoretical guidance for guiding malaria control and rational distribution of health resources.
Methods: (1) Cochran-Armitage trend test was used to analyze the long-term trend of the development and changes of local malaria epidemic in Hubei province for 2004~2011 years. The annual distribution map of annual incidence of malaria in each county (city, district) of Hubei province was drawn. (2) the global Moran'sI spatial autocorrelation analysis was used to analyze the incidence of malaria in the whole study area by the ArcGIS10.1 software. The spatial autocorrelation is not existed. Two local spatial aggregation methods are used to determine the high risk area of the local malaria epidemic in Hubei province in 2004~2011 years by using the ArcGIS10.1 local Moran's/ spatial autocorrelation analysis method and the simple spatial scanning statistics method. The high risk period and high risk area of the disease were determined. (3) the correlation between the incidence of malaria and the meteorological factors at different regional scales was analyzed by Spearman level correlation analysis. The meteorological factors affecting the incidence of malaria were screened by multiple regression stepwise regression analysis, and the change of malaria incidence was fitted with the weather factors. (4) the study was carried out. The incidence of malaria prevalence in Jujube was divided into the proportion of the number of patients in the whole year of the total number of patients in the whole year, which were divided into low hair month, mid month and high month. Using the stepwise discriminant analysis method, the epidemic degree of malaria in the future of Jujube was judged by meteorological factors. (5) the ARIMA model was constructed by the local malaria incidence of 2004~2009 years in Hubei Province, 1-12 in 2010. The monthly data are used to test the model and evaluate the fitting and prediction effect of the model.
Results: (1) the incidence of malaria in different years in Hubei province was significantly decreased in 2004~2011 years. The results of global spatial autocorrelation analysis suggested that the local malaria incidence in Hubei province had a certain spatial aggregation.2004 to 2011, and the local Moran's I spatial autocorrelation found 11 high risk counties of malaria, and the incidence of malaria in high-risk counties. The median from 58.1/10 million in 2004 to 0.79/10 million in 2011; the results of simple spatial scanning analysis found that there were 11 aggregated areas in Hubei province in 2004~2011 years, of which 8 were gathered in the first class and 3 in the two level, and the results of spatial and temporal aggregation analysis found the first order aggregation area, including 13 counties in the first class gathering area, and the aggregation period from April 2004 -2 In November, 007 years. (2) the periodic changes of the incidence of malaria in 2004~2009 years in Hubei and Jujube were significantly related to the cycle changes of climate during the year. The correlation with the temperature related indexes and rainfall was more significant, and the correlation coefficient was about 0.7. The change of malaria incidence rate of 78.1% in Hubei province was attributable to the month of the month. The average temperature (MeanT-01) and the average minimum temperature of the previous 2 months (MinT-2) were used in the previous month, and the change of malaria incidence in Jujube was 67.8% due to the average maximum temperature of the month and the previous month (MaxT-01). (3) a discriminant function of meteorological factors was used to distinguish the malaria prevalence in Jujube, and the first to 32 gases. The meteorological factors of the image factor were gradually discriminate, and the meteorological factors that finally introduced the discriminant equation were MinT0, the accuracy rate of the MaxT0 and D-012. discriminant functions was 73.61%. (4) the model was established by using the monthly local malaria incidence rate of 2004~2009 years in Hubei province. The results showed that the fitting effect of ARIMA (1,1,1) (1,1,0) 12 model was relatively optimal and predicted. The trend of the incidence of the disease is exactly the same as the actual incidence trend. The actual values are in the range of 95% confidence interval of the predicted value, which shows that the model prediction value is basically the same as the actual situation, and the fitting effect is good.
Conclusions: (1) 2004~2011 years, the incidence of local malaria in Hubei province was found to be significantly decreased. The incidence of malaria in Hubei province was spatially aggregated, high-risk areas still existed, and high-risk areas were mainly located in Anopheles Anopheles and Anopheles sinensis. (2) the use of meteorological factors to fit malaria and malaria in high incidence counties of malaria. The disease incidence rate has achieved good results. The constructed discriminant function can accurately determine the occurrence time of the high, middle and low onset months of malaria in Jujube. (3) the constructed ARIMA model is satisfied with the fitting results of malaria incidence in Hubei Province, and the prediction effect is good, which can be used to predict the trend of the change of the future malaria in Hubei province.
【学位授予单位】：华中科技大学
【学位级别】：博士
【学位授予年份】：2015
【分类号】：R531.3

【相似文献】