论文部分内容阅读
近几年,空间流行病学已经成为流行病学中一个研究热点,其相关理论与技术也在快速发展。空间流行病学旨在发掘和展示疾病在地理单元上的发病趋势,对于公共卫生决策具有重要意义。随着计算机技术高速发展和马尔科夫蒙特卡罗方法的成熟,贝叶斯时空模型已经成为空间流行病学的重要方法。贝叶斯时空模型不仅能够发掘疾病的时空变化规律,而且可以通过整合协变量提高模型估计的精度。贝叶斯时空模型已经发展出很多衍生类型。本文首先介绍了BYM模型,BYM模型是目前应用最广泛的贝叶斯时空模型。其次,介绍了在多水平空间数据结构中具有很高应用价值的分层贝叶斯时空模型。最后,介绍了最近几年在BYM模型基础上新研发的FBM模型。针对不同类型疾病时空数据特点,需要拟合相应的贝叶斯时空模型。本文在综合查阅文献的基础上对三种贝叶斯时空模型的研究进展进行阐述。
In recent years, space epidemiology has become a research hotspot in epidemiology, and its related theories and technologies are also rapidly developing. Spatial epidemiology aims to discover and display the trend of the disease on geographical units, which is of great significance to public health decision-making. With the rapid development of computer technology and the maturity of Markov Monte Carlo method, Bayesian spatio-temporal model has become an important method of space epidemiology. Bayesian spatio-temporal model can not only find out the temporal and spatial variation of disease, but also improve the accuracy of model estimation by integrating covariates. Bayesian space-time models have developed many derived types. This article first introduces the BYM model, which is the most widely used Bayesian spatio-temporal model. Secondly, the layered Bayesian spatio-temporal model with high application value in multi-level spatial data structure is introduced. Finally, the FBM model newly developed based on the BYM model is introduced in recent years. According to the characteristics of spatio-temporal data of different types of diseases, it is necessary to fit the corresponding Bayesian spatio-temporal model. This article elaborates the research progress of the three Bayesian spatio-temporal models based on a comprehensive review of the literature.