Dataset used in the research: Spatiotemporal patterns and integrated approach to prioritise areas for surveillance and control of visceral leishmaniasis in a large metropolitan area in Brazil.

We aimed to analyze the spatial and spatiotemporal patterns of VL occurrence and to identify priority risk areas for surveillance and control in the metropolitan region of Belo Horizonte-MG, Brazil. An ecological study was conducted considering all cases of VL in humans confirmed from 2006 to 2017, reported in Brazil’s national notification database (Notifiable Diseases Information System [SINAN]). We used databases from two different versions of SINAN: Windows version 2006 and TabNet version 2007–2017 (Ministry of Health, 2019). Therefore, we aggregated these two databases encompassing the entire period (2006–2017).
Incidence rates were calculated annually and for 3-year periods for all MRBH municipalities. To calculate the incidence rates, each case was aggregated by municipality of residence (analytical unit). The population estimates were set considering projections of the Federal Court of Audit (Tribunal de Contas da União [TCU]) calculated yearly for each municipality, based on data from the IBGE (2018).
The incidence rates were aggregated in 3 years, as follows: 1st triennium, 2006–2008; 2nd triennium, 2009–2011; 3rd triennium, 2012–2014; and 4th triennium, 2015–2017. The incidence rates per 100,000 inhabitants were calculated for each triennium.
Cumulative incidence rates were re-estimated for the geographic analytical units and for each triennium using empirical Bayesian space smoothing. Calculations were performed using GeoDa version 1.10 software (Arizona State University / Center for Geospatial Analysis and Computation, n.d.).
To perform the empirical Bayesian space smoothing analysis, the Moran’s Global Test and the Local Indicators of Spatial Association (LISA) were used to create the first order neighborhood matrix (Queen).
The units of analysis that presented at p ≤ 0.05 in the LISA were considered statistically significant. In order to view the priority municipalities for surveillance, we made choropleth maps. The Moran’s Global and LISA were calculated using GeoDa software version 1.10 (Arizona State University / Center for Geospatial Analysis and Computation, n.d.), and maps were constructed using QGIS software version 2.18 (QGIS project, 2019).
To identify spatiotemporal clusters, the SaTScan ™ 9.4.4 (Kulldorff, 2015) software scanning statistics was used. Statistical significance was considered when p <0.05 (Kulldorff and Nagarwalla, 1995). The spatial scan statistical analyzes for this study were performed using the case data set, population, and location. To perform the tests, the information regarding each municipality was inserted in the software: 1) number of cases, 2) year of infection, 3) population average of the 3 years that make up the 3-year period studied, and 4) geocode of each municipality. All this information was entered for the entire period.

本研究旨在分析巴西米纳斯吉拉斯州贝洛奥里藏特大都市区（Metropolitan Region of Belo Horizonte，MRBH）内脏利什曼病（Visceral Leishmaniasis，VL）的空间及时空分布模式，并识别优先监测与防控的高风险区域。
本研究为生态学研究，纳入了2006年至2017年巴西国家法定传染病报告数据库（Notifiable Diseases Information System，SINAN）上报的全部确诊人类VL病例。本次研究使用了两个不同版本的SINAN数据库：2006年Windows版与2007-2017年TabNet版（巴西卫生部，2019年），并将覆盖完整研究周期（2006-2017年）的两个数据库进行合并。
针对贝洛奥里藏特大都市区的所有辖区，本研究分别计算了年度发病率与3年周期发病率。发病率计算以病例所属常住辖区作为分析单元，对每一例病例进行归属汇总。人口估计数据采用巴西联邦审计法院（Tribunal de Contas da União，TCU）基于巴西地理与统计研究所（IBGE，2018年）数据编制的各辖区年度人口预测值。
发病率按3年周期进行聚合分组，具体如下：第一周期（2006-2008年）、第二周期（2009-2011年）、第三周期（2012-2014年）与第四周期（2015-2017年）。各3年周期的发病率均以每10万人口为单位进行计算。
针对各地理分析单元与每个3年周期，采用经验贝叶斯空间平滑法重新估算累积发病率。计算过程使用GeoDa 1.10版本软件完成（亚利桑那州立大学/空间分析与计算中心，无出版日期）。
为开展经验贝叶斯空间平滑分析，本研究使用全局莫兰检验（Moran’s Global Test）与空间关联局部指标（Local Indicators of Spatial Association，LISA）构建了一阶邻接矩阵（Queen邻接）。LISA检验中p值≤0.05的分析单元被认为具有统计学显著性。为可视化展示优先监测辖区，本研究绘制了分级统计图（choropleth maps）。全局莫兰检验与空间关联局部指标的计算仍使用GeoDa 1.10版本软件完成，分级统计图则通过QGIS 2.18版本软件（QGIS项目组，2019年）绘制。
为识别时空聚集区，本研究使用SaTScan™ 9.4.4软件的扫描统计功能（Kulldorff，2015年）。当p<0.05时认为结果具有统计学显著性（Kulldorff与Nagarwalla，1995年）。本研究的空间扫描统计分析基于病例数据集、人口数据与地理位置信息开展。软件录入的各辖区信息包括：1）病例数；2）感染年份；3）研究对应3年周期的年均人口数；4）各辖区的地理编码。上述全部信息均覆盖完整研究周期录入。