Chen Xi, Ba Jianbo, Liu Yuanhua, Huang Jiaqi, Li Ke, Yin Yun, Shi Jin, Xu Jiayao, Yuan Rui, Ward Michael P, Tu Wei, Yu Lili, Wang Quanyi, Wang Xiaoli, Chang Zhaorui, Zhang Zhijie
Department of Epidemiology and Health Statistics, School of Public Health, Fudan University, Shanghai, China; Key Laboratory of Public Health Safety, Ministry of Education, Shanghai, China.
Naval Medical Center, Naval Medical University, No.880 Xiangyin Road, Yangpu District, Shanghai, China.
Epidemiol Infect. 2025 Apr 16;153:e61. doi: 10.1017/S0950268824001080.
Hand, foot, and mouth disease (HFMD) shows spatiotemporal heterogeneity in China. A spatiotemporal filtering model was constructed and applied to HFMD data to explore the underlying spatiotemporal structure of the disease and determine the impact of different spatiotemporal weight matrices on the results. HFMD cases and covariate data in East China were collected between 2009 and 2015. The different spatiotemporal weight matrices formed by Rook, K-nearest neighbour (KNN; = 1), distance, and second-order spatial weight matrices () with first-order temporal weight matrices in contemporaneous and lagged forms were decomposed, and spatiotemporal filtering model was constructed by selecting eigenvectors according to and the We used , standard deviation of the regression coefficients, and five indices (, , , , and ) to compare the spatiotemporal filtering model with a Bayesian spatiotemporal model. The eigenvectors effectively removed spatial correlation in the model residuals (Moran's < 0.2, > 0.05). The Bayesian spatiotemporal model's Rook weight matrix outperformed others. The spatiotemporal filtering model with was superior, as shown by lower (92,029.60), (92,681.20), and (418,022.7) values, and higher (0.56) value. All spatiotemporal contemporaneous structures outperformed the lagged structures. Additionally, eigenvector maps from the Rook and closely resembled incidence patterns of HFMD.
手足口病（HFMD）在中国呈现出时空异质性。构建了一个时空滤波模型并将其应用于手足口病数据，以探索该疾病潜在的时空结构，并确定不同时空权重矩阵对结果的影响。收集了2009年至2015年中国东部地区的手足口病病例及协变量数据。分解了由Rook、K近邻（KNN；k = 1）、距离以及二阶空间权重矩阵（W）与同期和滞后形式的一阶时间权重矩阵形成的不同时空权重矩阵，并根据特征值和赤池信息准则（AIC）选择特征向量构建时空滤波模型。我们使用偏差、回归系数的标准差以及五个指标（AIC、BIC、离差、均方误差和平均绝对误差）将时空滤波模型与贝叶斯时空模型进行比较。特征向量有效地消除了模型残差中的空间相关性（莫兰指数I < 0.2，p > 0.05）。贝叶斯时空模型的Rook权重矩阵表现优于其他矩阵。具有W的时空滤波模型更优，其AIC（92,029.60）、BIC（92,681.20）和离差（418,022.7）值更低，平均绝对误差（0.56）值更高。所有时空同期结构均优于滞后结构。此外，Rook和W的特征向量图与手足口病的发病模式非常相似。
