Xia Shang, Xue Jing-Bo, Zhang Xia, Hu He-Hua, Abe Eniola Michael, Rollinson David, Bergquist Robert, Zhou Yibiao, Li Shi-Zhu, Zhou Xiao-Nong
National Institute of Parasitic Diseases, Chinese Center for Disease Control and Prevention, Key Laboratory of Parasite and Vector Biology, Ministry of Health, WHO Collaborating Center for Tropical Diseases, Shanghai, 200025, People's Republic of China.
Jiangling Institute of Schistosomiasis Control and Prevention, Jiangling, 434100, People's Republic of China.
Infect Dis Poverty. 2017 Apr 26;6(1):91. doi: 10.1186/s40249-017-0303-5.
The prevalence of schistosomiasis remains a key public health issue in China. Jiangling County in Hubei Province is a typical lake and marshland endemic area. The pattern analysis of schistosomiasis prevalence in Jiangling County is of significant importance for promoting schistosomiasis surveillance and control in the similar endemic areas.
The dataset was constructed based on the annual schistosomiasis surveillance as well the socio-economic data in Jiangling County covering the years from 2009 to 2013. A village clustering method modified from the K-mean algorithm was used to identify different types of endemic villages. For these identified village clusters, a matrix-based predictive model was developed by means of exploring the one-step backward temporal correlation inference algorithm aiming to estimate the predicative correlations of schistosomiasis prevalence among different years. Field sampling of faeces from domestic animals, as an indicator of potential schistosomiasis prevalence, was carried out and the results were used to validate the results of proposed models and methods.
The prevalence of schistosomiasis in Jiangling County declined year by year. The total of 198 endemic villages in Jiangling County can be divided into four clusters with reference to the 5 years' occurrences of schistosomiasis in human, cattle and snail populations. For each identified village cluster, a predictive matrix was generated to characterize the relationships of schistosomiasis prevalence with the historic infection level as well as their associated impact factors. Furthermore, the results of sampling faeces from the front field agreed with the results of the identified clusters of endemic villages.
The results of village clusters and the predictive matrix can be regard as the basis to conduct targeted measures for schistosomiasis surveillance and control. Furthermore, the proposed models and methods can be modified to investigate the schistosomiasis prevalence in other regions as well as be used for investigating other parasitic diseases.
血吸虫病的流行在中国仍然是一个关键的公共卫生问题。湖北省江陵县是典型的湖沼型血吸虫病流行区。江陵县血吸虫病流行模式分析对于推动类似流行区的血吸虫病监测与控制具有重要意义。
基于2009年至2013年江陵县的年度血吸虫病监测数据以及社会经济数据构建数据集。采用一种从K均值算法改进而来的村庄聚类方法来识别不同类型的流行村庄。对于这些识别出的村庄聚类，通过探索一步向后时间相关性推断算法开发了一种基于矩阵的预测模型，旨在估计不同年份间血吸虫病流行率的预测相关性。对家畜粪便进行现场采样，作为潜在血吸虫病流行情况的指标，并将结果用于验证所提出模型和方法的结果。
江陵县血吸虫病流行率逐年下降。参照人类、牛和螺类群体5年的血吸虫病发病情况，江陵县的198个流行村庄可分为四类。对于每个识别出的村庄聚类，生成了一个预测矩阵，以表征血吸虫病流行率与历史感染水平及其相关影响因素之间的关系。此外，现场粪便采样结果与识别出的流行村庄聚类结果相符。
村庄聚类结果和预测矩阵可作为开展血吸虫病监测与控制针对性措施的依据。此外，所提出的模型和方法可进行修改，以调查其他地区的血吸虫病流行情况，也可用于调查其他寄生虫病。
