Cheong Yoon Ling, Ghazali Sumarni Mohd, Che Ibrahim Mohd Khairuddin Bin, Kee Chee Cheong, Md Iderus Nuur Hafizah, Ruslan Qistina Binti, Gill Balvinder Singh, Lee Florence Chi Hiong, Lim Kuang Hock
Institute for Medical Research, National Institutes of Health, Ministry of Health Malaysia, Kuala Lumpur, Malaysia.
Sector for Biostatistics and Data Repository, National Institutes of Health, Ministry of Health Malaysia, Shah Alam, Malaysia.
Front Public Health. 2022 Mar 4;10:836358. doi: 10.3389/fpubh.2022.836358. eCollection 2022.
The unprecedented COVID-19 pandemic has greatly affected human health and socioeconomic backgrounds. This study examined the spatiotemporal spread pattern of the COVID-19 pandemic in Malaysia from the index case to 291,774 cases in 13 months, emphasizing on the spatial autocorrelation of the high-risk cluster events and the spatial scan clustering pattern of transmission.
We obtained the confirmed cases and deaths of COVID-19 in Malaysia from the official GitHub repository of Malaysia's Ministry of Health from January 25, 2020 to February 24, 2021, 1 day before the national vaccination program was initiated. All analyses were based on the daily cumulated cases, which are derived from the sum of retrospective 7 days and the current day for smoothing purposes. We examined the daily global, local spatial autocorrelation and scan statistics of COVID-19 cases at district level using Moran's I and SaTScan™.
At the initial stage of the outbreak, Moran's I index > 0.5 ( < 0.05) was observed. Local Moran's I depicted the high-high cluster risk expanded from west to east of Malaysia. The cases surged exponentially after September 2020, with the high-high cluster in Sabah, from Kinabatangan on September 1 (cumulative cases = 9,354; Moran's I = 0.34; < 0.05), to 11 districts on October 19 (cumulative cases = 21,363, Moran's = 0.52, < 0.05). The most likely cluster identified from space-time scanning was centered in Jasin, Melaka (RR = 11.93; < 0.001) which encompassed 36 districts with a radius of 178.8 km, from November 24, 2020 to February 24, 2021, followed by the Sabah cluster.
Both analyses complemented each other in depicting underlying spatiotemporal clustering risk, giving detailed space-time spread information at district level. This daily analysis could be valuable insight into real-time reporting of transmission intensity, and alert for the public to avoid visiting the high-risk areas during the pandemic. The spatiotemporal transmission risk pattern could be used to monitor the spread of the pandemic.
史无前例的新冠疫情极大地影响了人类健康和社会经济背景。本研究考察了马来西亚新冠疫情从首例病例到13个月内291,774例的时空传播模式，重点关注高风险聚集事件的空间自相关性以及传播的空间扫描聚类模式。
我们从马来西亚卫生部的官方GitHub库中获取了2020年1月25日至2021年2月24日（即国家疫苗接种计划启动前一天）马来西亚新冠确诊病例和死亡数据。所有分析均基于每日累计病例，这些病例是为了平滑处理而由回顾性7天病例数与当日病例数之和得出。我们使用莫兰指数（Moran's I）和SaTScan™软件，在地区层面考察了新冠病例的每日全局、局部空间自相关性以及扫描统计数据。
在疫情爆发初期，观察到莫兰指数（Moran's I）> 0.5（p < 0.05）。局部莫兰指数显示高风险聚集从马来西亚西部扩展到东部。2020年9月之后病例呈指数级激增，沙巴州出现高风险聚集，从9月1日的京那巴当岸（累计病例 = 9354；莫兰指数 = 0.34；p < 0.05），到10月19日扩展至11个地区（累计病例 = 21363，莫兰指数 = 0.52，p < 0.05）。从时空扫描中确定的最可能聚集区域以马六甲的麻坡为中心（相对风险 = 11.93；p < 0.001），该区域在2020年11月24日至2021年2月24日期间涵盖36个地区，半径为178.8公里，其次是沙巴聚集区。
两种分析方法在描绘潜在的时空聚集风险方面相互补充，提供了地区层面详细的时空传播信息。这种每日分析对于实时报告传播强度可能具有重要价值，并可提醒公众在疫情期间避免前往高风险地区。时空传播风险模式可用于监测疫情的蔓延。
