【作者】 王丽萍；

【导师】 金水高； 曹务春；

【作者基本信息】 中国疾病预防控制中心 ， 流行病与卫生统计学， 2008， 博士

【摘要】 背景:疟疾是一种由疟原虫引起的、经按蚊叮咬而传播的重要的寄生虫病。全球仍有107个国家和地区,约32亿人口受到疟疾的威胁,每年有3亿多人受感染,病死者超过100万。上世纪50、60年代,我国疟疾流行严重,经过半个多世纪的努力,我国疟疾防治工作取得巨大成效,20世纪末全国疟疾疫情下降到最低,但近年来疟疾发病数呈明显回升。2000年以来以安徽省为代表的中部地区在北纬32°以北的单一中华按蚊区陆续出现疟疾疫情回升、小暴发点或局部暴发流行。影响疟疾流行的因素复杂,寄生虫、蚊媒、人类宿主和环境等因素及其交互作用决定了疟疾传播和感染的风险。近年来基于地理信息系统的3S已成为疟疾等自然疫源性传染病研究新的技术手段与方法,而将统计学方法与空间技术相结合也成为一个新的研究方向。目的:通过对20世纪90年代以来安徽省疟疾疫情时空分析及影响因素研究,为该省及类似地区的疟疾防控提供科学依据。通过尝试3S空间分析技术与统计分析方法在疟疾研究领域的综合应用,为类似的研究提供方法学参考。方法:收集整理1990～2006年安徽省疟疾疫情监测数据,安徽省1:1 000 000县区级数字区,并加工处理安徽省1:5万的乡镇边界图。采用空间分析技术提取安徽省各乡镇温度、降雨量等气象监测数据,海拔、NDVI、湿度指数、水体等环境数据以及人口、GDP等数据,建立乡镇疟疾流行的地理信息系统数据库。对淮北地区高发自然村庄采用抽样调查及GPS定位的方法,收集居民生产生活及行为因素、疟疾病例诊治及水体定位等相关信息,建立村庄尺度的研究数据库。综合应用时空扫描聚类分析方法、时间序列分析方法、主成分分析及logistic回归模型、Poisson回归模型等统计分析方法对资料进行分析处理。所采用的软件包括Arc GIS 9.0软件、SaTScan7.0空间聚类分析软件、SAS9.1及SPSS13.0统计分析软件。结果:①“2004～2006年淮北地区”是安徽省20世纪90年代以来疟疾流行新“热点”,淮北地区疟疾传播季节延长。②时间纵向上,温度、降雨量与疟疾发病率序列呈现显著的互相关关系,即“月平均气温”升高则“月疟疾发病率”上升,呈现正相关关系,在滞后0～3个月的相关性均显著;“月降雨量”增加则“月疟疾发病率”上升,在滞后1～3个月的相关性显著,不同的是,其相关性出现在至少滞后1个月,表明降雨量与疟疾发病间的关系相对缓慢一些。③地区横向上,安徽省各乡镇的疟疾发病率与温度、降雨量、NDVI和海拔因素有关,即“冬季/年最低气温”、“年降雨总量”、“海拔”和“NDVI”4个因素与安徽乡镇是否发生疟疾有关。在控制模型中其它自变量不变的情况下,前3个因素数值增加,则乡镇疟疾发生的可能性越小,而NDVI增加,则乡镇发生疟疾的可能性越大。④淮北地区居民不良生活生产行为使该地区疟疾发病风险增加,分析结果显示,村民露宿习惯比例每增加1%,村发生疟疾病例的风险增加18%,而耕种豆类农作物也增加了村民发生疟疾的危险。⑤采用2000～2007年5月安徽省疟疾月发病率建立的ARIMA时间序列模型对2007年6月份的发病率的预测值为5.437/10万,95%可信区间为[2.308/10万,12.808/0万],实际监测发病率为5.334/10万,预测的相对误差为1.9%,预测效果良好。结论:本课题研究重点回答了2000年以来安徽省疟疾疫情回升的新的时空热点地区,并重点回答了安徽省南北地区疟疾疫情存在差异的主要影响因素。淮北地区除当地地形地貌、自然环境等特点影响其疟疾流行外,当地居民不良的生产生活行为因素加大了蚊媒接触机率,增加了疟疾发生的机率,研究结果具有很现实的指导意义。研究中虽然没有直接得出病例诊治及时性对该地区疟疾疫情的实际影响,但从控制传染源的理论出发,疟疾病人的及时发现和治疗无疑也是控制疟疾疫情流行的重要方面。课题研究结果将为安徽等我国中部疟疾流行地区建立疟疾的早期预警预测系统奠定基础。此外,时间序列模型可以很好地拟合疟疾发病率在时间序列上的变动趋势,在人群免疫状态、人口流动、防制措施等人群疟疾易感性指标没有发生大幅度变化的情况下,可以用来对未来的疟疾发病率进行预测,为疟疾防治工作提供服务。更多还原

【Abstract】 Background:Malaria is one of the important verminosis caused by a parasite that is transmitted to humans by a female mosquito’s bite.Some 3.2 billion people live in areas at risk of malaria transmission in 107 countries.Each year 300 million are diagnosed with malaria and more the one million people die of malaria.There were serious malaria epidemics in our country during 1950’s and 1960’s.Huge progress has been made in preventing and treating malaria after efforts for over half a century.The incidence rate of malaria reached the lowest level at the end of last century, and started to increase again at the central Anopheles sinensis area represented by Anhui province located 32°north latitude since 2000.Some small and large outbreaks at local areas were also reported.The complexity of malaria,its transmission and infection risks is determined by parasite,human host,environment and the interactions between them.The GIS,GPS and RS(abbreviated as 3S technology) have become the new methods and technical measures to study malaria and other zoonotic and vector-borne diseases.The combination of statistics and space technology has become the new research direction.Objectives:The study explores the spatio-temporal characters of malaria epidemic situation and its determinants in Anhui province since 1990 in order that provide empirical evidence for the prevention and control policy of malaria’s in the Anhui and similar districts. In addition to it also pave a way for the synthetic technology for malaria research field by using spatial statistics and 3S technologies,which can be used in the other similar study.Methods:The surveillance data of malaria epidemic was collected and digital maps at the county(1:1 000 000),township(1:50 000) levels were processed.The township level data for the temperature,precipitations,altitude,NDVI,humidity and water environment,also including the population and GDP was extracted through the spatial analysis technology, and the corresponding GIS database was created.Then collecting village level data for the life styles,behavior factors,and disease diagnosis and treatment information of local residents in the high incidence rate area in Huaibei districts by the sample field survey, and water body locations can be obtained by GPS localizer.The village level analysis database was also created accordingly.Spatio-temporal scanning cluster analysis, time-series analysis,principal component analysis and logistic regression model,and Poisson regression model was used and the software used includes Arc GIS 9.0, SaTScan7.0,SAS9.1 and SPSS13.0.Results:1.The Huaibei area in Anhui province from 2004 to 2006 is the hot spot of malaria epidemic since 1990s’,the duration of malaria transmission season was extended in Huaibei area.2.Longitudinally significant correlation was found between malaria incidence rate and environmental factors(temperature and precipitation).The positive correlation existed between the ’increase’ of monthly average temperature and monthly malaria incidence rates when the data were staggered to allow a lag of 0-3 month.The increasing monthly precipitation was accompanied by increase of monthly malaria incidence rate when they were transformed by 1-3 month lag.The correlations were significant for the malaria incidence rates with at least one month lag.This indicated that the relationship between precipitation and malaria incidence was relatively slow.3.The incidence of malaria was related to temperature,precipitation,NDVI and altitude at township level analysis in Anhui province.Winter/the lowest temperature,annual precipitation,altitude and NDVI are the influencing factors for malaria.When other variables in the model are controlled,the higher the value of former 3 factors is the less chance of the malaria occurs.However,the increase of NDVI increases the chances to have malaria.4.The poor life style and local behavior of residents in this area put them in higher risk to get infected.The results show that if the percentage of outdoor sleeping increases by 1%, the relative ratio of infected malaria will increase 8%.The cultivation of bean crop will increase the risk of infected malaria. 5.The ARIMA time series model established based upon the incidence data of Jan.2000 to May.2007,forecast that the incidence of malaria in June 2007 is 5.437/100 000(CI 2.308/100 000,12.808/100 000).The real incidence is 5.334/100000.The relative forecast error is 1.9%.Conclusions:This study mainly answered the reasons of the rebound of malaria epidemics in the new spatio-temporal hot spot at Anhui province since 2000.The combination of 3S spatial analysis techniques and statistics were used to identify the main factors explaining the differences of incidence rate between southern and northern areas.In addition to the terrain and natural environment in Huaibei area,the tillage style and poor lifestyle behavior factors increased contact with mosquito,so this increased the chance to be infected for the residents.The results nail down the focus of malaria prevention and control,which is of momentous current significance for the anti-malaria.The results did not support that the effects of the timeliness of malaria diagnosis,but based upon the theory of controlling infectious source,it is very important to find and treat patient timely in order to control the malaria.The research will lay a solid foundation for early warning and early forecast in the central areas such as Anhui.Moreover,time series model can be used to fit the changing trends of malaria incidence rate at different times.If the population immunization status,population mobility, prevention measures and other susceptibility indicators are relative stable,it can be used to forecast the incidence of malaria in the future,and to provide services for prevention and control of malaria epidemics.更多还原