广西1996～2012年艾滋病流行特征及流行地区数理判别模型研究

发布时间：2018-09-03 15:03

【摘要】：第一部分广西1996～2012年艾滋病流行特征研究目的分析广西1996～2012年艾滋病的流行特征，探讨广西艾滋病流行强度的地区差异。为广西艾滋病针对性防控策略和措施的制定提供科学依据。方法收集广西1996～2012年广西艾滋病防治信息管理系统中首次报告的确证HIV抗体阳性的感染者（包括HIV感染者和AIDS病人），，用各种统计图表对广西艾滋病流行特征进行描述性分析；并将感染者以县（市、区）为基本统计单位，通过GIS软件的自然断点分级法划分累积感染率（CPI）、累积患病率(CP)、累积死亡率(CMR)、累积病死率(CFR)4项流行指标（下称4项流行指标）高、中、低流行区，对比分析各县（市、区）4项流行指标的特征。结果 1.广西艾滋病流行特征（1）传染源累积数量大。1996～2012年累计报告广西籍HIV感染者和AIDS病人83241例，其中HIV感染者47094例，所占比例为56.58%、；36147例感染者在首次确证感染时即为AIDS病人,占43.42%。（2）感染趋势先慢后快。2003年之前，感染者数量增长缓慢，2003年以后，数量呈快速增长，但2012年有下降趋势。（3）感染人群特征明显。以男性为主，占72.34%；年龄集中在15-49岁，占68.94%；以汉族（61.11%）、已婚（57.05%）、初中及以下文化程度（79.11%）、职业为农民（含民工）和工人（60.39%）为主。（4）异性性传播为主要感染途径。66.31%的HIV感染者通过异性性传播途径感染；（5）非婚异性性行为是主要高危行为。感染者中存在非婚异性性行为（至少1个非婚异性性伴）的比例占42.56%，其中80.59%有2个及以上非婚异性性伴。（6）临床就诊为发现HIV感染者的主要途径。感染者主要以临床就诊检查发现为主，占65.68%。 2.广西艾滋病流行区的划分及特点以GIS软件的自然断点分级法将CPI、CP、CMR、CFR分成三组，对应为高、中、低流行区。CPI高、中、低流行区所包括的县（市、区）分别为15、45、49个，相应区间分别为2.99‰～6.91‰、1.21‰～2.98‰、0.18‰～1.20‰。CP高、中、低流行区所包括的县（市、区）分别为13、41、65个，相应区间分别为1.54‰～4.24‰、0.55‰～1.53‰、0.06‰～0.54‰。CMR高、中、低流行区所辖县（市、区）分别为4、20、85个，相应区间为0.58‰～1.26‰、0.17‰～0.57‰、0.00‰～0.16‰。CFR高、中、低流行区所包括的县（市、区）分别为18、45、46个，相应区间分别为27.25%～50.00%、13.28%～27.24%、0.00%～13.27%。 CPI和CP高、中、低流行区分布一致性较高，以CPI为参照，CP高、中、低流行区与CPI高、中、低流行区的符合率分别为86.67%、84.44%、93.88%,总体符合率达到88.99%。CPI、CP均高的县（市、区）艾滋病流行较早，主要分布在边境城市、柳州市城区及周围。CPI、CP均居中的县（市、区）主要分为两大聚集丛：以柳州片区的高流行县为中心向周围辐射的聚集丛和以边境城市的高流行县为起始，向内部蔓延的聚集丛。CPI、CP均低的县（市、区）则主要分布高、中流行区的外围。直观地图可发现，CPI、CP高、中流行区主要为自东北向西南以约45°角倾斜走向的广西经济相对发达，交通较为便利的中部地带。CMR高、中流行区所包括的县（市、区）与CPI、CP一致。然而CFR高流行区所辖的县（市、区）中，77.78%CPI、CP、CMR均低。结论 1.广西艾滋病传染源累积基数较大，1996～2012年艾滋病的流行特点呈现先慢后快趋势，但在2012年开始下降。广西艾滋病存在43.42%的HIV阳性感染者在首次确证感染时即为AIDS病人，感染途径主要为异性接触传播，主要高危行为是非婚异性性行为等特征。因此，针对性的防控策略和措施，如推广简便、快捷的HIV检测方法，加强对文化程度较低人群的艾滋病相关健康教育，进一步推广性行为、特别是非婚性行为中安全套的使用等措施对预防广西艾滋病通过性途径的进一步蔓延具有重大的意义。 2.利用GIS自然断点分级法划分的高、中、低流行区具有较高的符合率，为直观广西不同县（市、区）艾滋病流行状况提供了便利，但需要重点考虑社会、行为等因素的影响。 3.针对不同县（市、区）4项流行指标的等级差异，广西艾滋病防控重点可以考虑CPI与CP的高、中流行区域，而CPI与CP的低流行区域（即CFR高流行区）的干预重点是降低HIV/AIDS的病死率。第二部分广西艾滋病流行地区数理判别模型研究目的以CPI、CP、CMR（简称为3项流行指标，下同）为主体指标进行分层分析，分析不同流行地区的艾滋病流行模式和趋势，并探讨影响广西艾滋病3项流行指标的经济社会学因素，建立影响CPI的地区经济社会发展相关因素的多重线性回归模型以及3项流行指标高、中、低流行地区类别的数理判别模型，为科学制定艾滋病防控策略提供依据。方法收集影响高、中、低流行区的经济社会学发展指标资料，利用Spearman简单相关分析它们与CPI、CP、CMR的关联性；用单因素方差分析方法比较不同流行区经济社会发展指标的差异性；用逐步回归分析方法建立影响CPI的多重线性回归模型；用Fisher逐步判别法建立广西艾滋病不同流行地区类别的数理判别模型，并进行效果评价。结果 1.经济社会学发展因素与CPI、CP、CMR的关联性：8项经济社会学相关指标（人口密度X1、非农业人口占总人口比重X2、人口自然增长率X3、人均地区生产总值X4、城镇居民人均可支配收入X5、农村居民人均纯收入X6、受教育人口九年义务教育完成率X7、AIDS病人累积治疗率X8与3项流行指标进行Spearman简单相关分析发现，X1、 X2、X4、X5、X6、X7与CPI、CP相关（P0.10）；X2、X4、X6、X7与CMR相关（P0.10）。 2.影响CPI、CP、CMR的经济社会学发展因素：（1）影响CPI的地区经济社会学因素为X2、X4、X5、X6、X7，多因素模型影响因素为X7，判别模型影响因素为X4、X7；影响CP的地区经济社会学因素为X2、X4、X6、X7，判别模型影响因素为X4、X7，多因素模型影响因素为X7；影响CMR的地区经济社会学因素为X2、X4、X7，多因素模型影响因素为X4，判别模型影响因素为X4。X4是3项流行指标地区类别判别模型的共同影响因素，均呈正相关关系；而X7是CPI和CP的地区类别判别模型的共同影响因素，同样呈正相关关系。 3. CPI、CP、CMR不同流行地区数理判别模型的构建：结合单因素分析有统计学意义的经济社会学发展因素以及专业知识，将与艾滋病流行密切相关的因素纳入Fisher逐步判别模型分析。建立相应的地区类别判别模型结果如下： 3.1CPI判别模型 CPI高流行区：YH=-19.36+2.39×10-4X4+0.62X7 CPI中流行区：YM=-19.97+4.68×10-4X4+0.69X7 CPI低流行区：YL=-15.72+4.27×10-4X4+0.61X7 3.2CP判别模型 CP高流行区：YH=-19.32+1.80×10-4X4+0.61X7 CP中流行区：YM=-20.31+4.43×10-4X4+0.69X7 CP低流行区：YL=-15.97+4.29×10-4X4+0.62X7 3.3CMR判别模型 CMR高流行区：YH=-3.78+3.78×10-4X4（由于自然断点法划分的CMR高流行区包括的研究单位较少，因此在建模过程中将高、中流行区合并为高流行区） CMR低流行区：YL=-2.54+2.92×10-4X4 4.判别模型效果评价：用回代法检验各判别函数的符合率：（1）CPI高、中、低流行地区的判别符合率分别为56.25%、58.18%、53.06%，总的判别符合率为51.38%。(2)CP高、中、低流行地区的判别符合率分别为53.84%、48.78%、52.4%，总体判别符合率为54.13%。(3)CMR高、低流行地区的判别符合率分别为48.00%、69.05%，总体判别符合率为64.22%。结论 1.CPI、CP、CMR、CFR能够全面反映广西艾滋病的流行强度。 2.人均地区生产总值（X4）、受教育人口九年义务教育完成率（X7）是促进广西艾滋病流行的重要经济社会学发展因素。 3.对广西艾滋病流行区的划分可考虑构建数理判别模型，但其判别符合率的提高仍需要结合考虑多样、复杂的行为、生物学等因素的影响。
[Abstract]:The first part is about the epidemiological characteristics of AIDS in Guangxi in the past 1996~2012 years.
objective
To analyze the epidemic characteristics of AIDS in Guangxi from 1996 to 2012, and to explore the regional differences of the epidemic intensity of AIDS in Guangxi.
Method
To collect the first reported HIV-positive infections (including HIV-infected persons and AIDS patients) in Guangxi AIDS prevention and control information management system from 1996 to 2012, and analyze the epidemic characteristics of AIDS in Guangxi with various statistical charts and charts. However, the breakpoint classification method was used to classify the four epidemic indicators (CPI, CP, CMR and CFR) into high, middle and low epidemic areas. The characteristics of the four epidemic indicators in each county (city, district) were compared and analyzed.
Result
1. epidemiological characteristics of AIDS in Guangxi
(1) The cumulative number of sources of infection is large. From 1996 to 2012, 83 241 cases of HIV-infected persons and AIDS patients in Guangxi were reported, of which 47 094 were infected with HIV, accounting for 56.58%; 36 147 were infected with AIDS at the time of first confirmation of infection, accounting for 43.42%. (2) The trend of infection was slow and fast. Before 2003, the number of infected persons increased slowly, and in 2003, the number of infected persons increased slowly. After that, the number showed a rapid growth, but there was a downward trend in 2012. (3) The characteristics of the infected population were obvious. (3) Male dominated, accounting for 72.34%; age concentrated in 15-49 years old, accounting for 68.94%; Han (61.11%), married (57.05%), junior high school and below education level (79.11%), occupation for farmers (including migrant workers) and workers (60.39%) mainly heterosexual transmission. (4) heterosexual transmission was the main route of infection. 66.31% of HIV-infected people were infected through heterosexual transmission; (5) heterosexual non-marital sex was the main high-risk behavior. 42.56% of HIV-infected people had heterosexual non-marital sex (at least one heterosexual partner), of which 80.59% had two or more heterosexual partners. (6) Clinic visits were the main way to find HIV-infected people. The patients were mainly found by clinical examination, accounting for 65.68%.
2. the classification and characteristics of Guangxi AIDS epidemic area
CPI, CP, CMR and CFR were divided into three groups according to the natural breakpoint grading method of GIS software, which corresponded to high, middle and low prevalence areas. The corresponding ranges were 1.54 8240to 4.24 82, 0.55 82to 1.53 82, 0.55 82to 1.53 82, 0.06 82to 0.54 82. CMR was high, 4, 20, 85 count (city, district) in the middle and low endemic areas were 4, 20, 85, the corresponding ranges were 0.58 82to 1.26 82, 0.17 82to 1.57 82, 0.17 82to 0.57 82, 0.00 82to 0.16 82 82. CFR was high, the middle and low endemic areas were 4, 20, 20 and 85 respectively, the corresponding ranges were 0.58 82 82to 1.26 8218,45,46, respectively 27.25% ~ 50%, 13.28% ~ 27.24%, 0% ~ 13.27%.
The coincidence rates of CPI and CP were 86.67%, 84.44%, 93.88% and 88.99% respectively. The counties (cities and districts) with higher CPI and CP had earlier AIDS epidemic, mainly distributed in border cities, Liuzhou city and its surrounding areas. The counties (cities, districts) in the middle are mainly divided into two clusters: the cluster with the high prevalence counties in the Liuzhou area as the center and the cluster with the high prevalence counties in the border cities as the beginning, and the cluster spreading inward. It is mainly from the northeast to southwest to about 45 degrees inclined to the relatively developed economy, traffic is more convenient in the central zone. CMR is high, the counties (cities, districts) included in the epidemic areas are consistent with CPI, CP. However, in the counties (cities, districts) under the jurisdiction of the high prevalence areas of CFR, 77.78% of the CPI, CP, CMR are low.
conclusion
1. The cumulative base of HIV infection sources in Guangxi is large. The epidemic characteristics of AIDS from 1996 to 2012 show a slow and fast trend, but it began to decline in 2012. 43.42% of HIV-positive people in Guangxi were infected with AIDS when they were first confirmed. The main route of infection is heterosexual contact transmission, and the main high-risk behavior is heterosexual non-marriage. Therefore, targeted prevention and control strategies and measures, such as the promotion of simple and fast HIV testing methods, to strengthen AIDS-related health education for the less educated population, to further promote sexual behavior, especially the use of condoms in non-marital sexual behavior, and other measures to prevent the further spread of AIDS through sexual channels in Guangxi. Of great significance.
2. The high, middle and low prevalence areas classified by the natural breakpoint classification method of GIS have higher coincidence rate, which provides convenience for visualizing the epidemic situation of AIDS in different counties (cities and districts) in Guangxi, but the influence of social and behavioral factors should be considered.
3. Aids prevention and control in Guangxi should focus on the high and middle prevalence areas of CPI and CP, while the low prevalence areas of CPI and CP (i.e. high prevalence areas of CFR) should focus on reducing the mortality of HIV/AIDS.
The second part is the mathematical discriminant model of the epidemic area in Guangxi.
objective
Taking CPI, CP, CMR (hereinafter referred to as three epidemic indicators) as the main indicators for stratified analysis, the epidemic patterns and trends of AIDS in different epidemic areas were analyzed, and the economic and sociological factors affecting the three epidemic indicators of AIDS in Guangxi were explored. The multiple linear regression model and 3. The mathematical discriminant model of the high, middle and low epidemic areas provides the basis for scientific formulation of AIDS prevention and control strategies.
Method
The data of economic and social development indicators in high, middle and low prevalence areas were collected, and the correlation between them and CPI, CP and CMR was analyzed by Spearman simple correlation method. Fisher stepwise discriminant method was used to establish mathematic discriminant models for different types of AIDS epidemic areas in Guangxi.
Result
1. Correlation between economic sociology development factors and CPI, CP, CMR: 8 economic sociology related indicators (population density X1, non-agricultural population proportion X2, natural population growth rate X3, per capita GDP X4, per capita disposable income X5 of urban residents, per capita net income X6 of rural residents, nine-year compulsory education completion rate of educated population) Spearman correlation analysis showed that X1, X2, X4, X5, X6, X7 were correlated with CPI and CP (P 0.10), X2, X4, X6, X7 were correlated with CMR (P 0.10).
2. Economic and sociological development factors affecting CPI, CP and CMR: (1) The regional economic and sociological factors affecting CPI are X2, X4, X5, X6, X7, the multi-factor model influencing factor is X7, the discriminant model influencing factor is X4, X7; the regional economic and sociological factors influencing CP are X2, X4, X6, X7, the discriminant model influencing factor is X4, X7, and the multi-factor model influencing factor is X7. X2, X4, X7 were the regional socioeconomic factors influencing CMR, X4 was the multifactor model influencing factor, and X4.X4 was the common influencing factor of the three epidemic index regional classification discriminant models. X7 was the common influencing factor of CPI and CP regional classification discriminant models, and also had a positive correlation.
3. Establishment of mathematical discriminant models for different epidemic areas of CPI, CP and CMR: Combining with single factor analysis of statistically significant economic and social development factors and professional knowledge, the factors closely related to AIDS epidemic will be incorporated into Fisher stepwise discriminant model analysis.
3.1CPI discriminant model
CPI high endemic area: YH=-19.36+2.39 * 10-4X4+0.62X7
CPI epidemic area: YM=-19.97+4.68 * 10-4X4+0.69X7
CPI low epidemic area: YL=-15.72+4.27 * 10-4X4+0.61X7
3.2CP discriminant model
CP high endemic area: YH=-19.32+1.80 * 10-4X4+0.61X7
CP epidemic area: YM=-20.31+4.43 * 10-4X4+0.69X7
CP low epidemic area: YL=-15.97+4.29 * 10-4X4+0.62X7
3.3CMR discriminant model
High prevalence area of CMR: YH = - 3.78 + 3.78 *10-4X4 (because there are fewer research units in the high prevalence area of CMR divided by natural breakpoint method, the high and medium prevalence areas are merged into high prevalence areas in the modeling process)
CMR low epidemic area: YL=-2.54+2.92 * 10-4X4
4. Effect evaluation of discriminant model: The coincidence rate of each discriminant function was tested by substitution method: (1) The coincidence rate of discriminant function was 56.25%, 58.18%, 53.06% in high, medium and low prevalence areas respectively, and the total coincidence rate was 51.38%. (2) The coincidence rate of discriminant in high, middle and low prevalence areas was 53.84%, 48.78%, 52.4%, and the overall coincidence rate was 54.13%. (3) CMR. The discriminative coincidence rates in high and low endemic areas were 48% and 69.05% respectively, and the overall discriminant accordance rate was 64.22%.
conclusion
1.CPI, CP, CMR and CFR can fully reflect the epidemic intensity of AIDS in Guangxi.
2. The per capita GDP (X4) and the nine-year compulsory education completion rate (X7) of the educated population are important economic and sociological factors to promote the AIDS epidemic in Guangxi.
3. Establishing mathematical discriminant model can be considered in the division of AIDS epidemic areas in Guangxi, but the improvement of the discriminant coincidence rate still needs to be combined with the influence of diverse, complex behavior, biology and other factors.
【学位授予单位】：广西医科大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：R512.91

【参考文献】