论文部分内容阅读
目的:调查巴西圣埃斯皮里图州维多利亚市肺结核发病空间模型及其与社会经济状况的关系。设计:2002—2006年在维多利亚进行为期4年的、基于地区的新发肺结核患者监测回顾性研究,并使用空间聚集分析[Anselin局部空间相关指标(LISA)和Getis-Ord Gi*统计值]、平滑化后的经验贝叶斯估计和模型预测的发病率3种方法来比较疾病发病的空间模型。并用空间泊松模型检验结核发病和社会经济状况的关系。结果:78个社区共报告了651例结核患者,其报告率从0到129/10万不等。Moran’sI表明发病率中有很强的空间自相关(0.399,P<0.0001),LISA和Gi*统计值发现了4个高发病区域。平滑化后经验贝叶斯估计展示了其中的2个区域,发病率由70/10万至90/10万变化,而另外2个地区变化范围为40~70/10万。结核发病和社会经济有显著性曲线相关关系(P=0.02)。结论:空间统计工具提取的数据可以帮助结核病控制规划将结核病资源分配到高发病危险的人群和结核病控制规划需要加以关注的目标地区。
Objective: To investigate the spatial pattern of tuberculosis incidence in Victoria, the city of Santo Espírito Santo, Brazil and its relationship with socioeconomic status. Design: A four-year, region-based surveillance of patients with newly diagnosed pulmonary tuberculosis in Victoria from 2002 to 2006 and a retrospective study using spatial clustering analysis [Anselin LISA and Getis-Ord Gi * statistics] Three methods of smoothed empirical Bayes estimation and model predictive morbidity are used to compare spatial models of disease onset. Poisson model was used to test the relationship between tuberculosis incidence and socioeconomic status. Results: A total of 651 TB patients were reported in 78 communities, with rates ranging from 0 to 129 / 100,000. Moran’s showed that there was a strong spatial autocorrelation (0.399, P <0.0001) in the incidence, and four high-incidence areas were found on the LISA and Gi * statistics. Experienced Bayesian estimation of smoothening showed two of these regions, with the incidence varying from 70/10 to 90/10 million, while the other two regions varied from 40-70 / 100,000. Tuberculosis incidence and socio-economic significant curve correlation (P = 0.02). Conclusions: Data extracted from spatial statistical tools can help TB control programs allocate TB resources to high-risk populations and target areas for tuberculosis control programs that require attention.