Methods Used in the Spatial and Spatiotemporal Analysis of COVID-19 Epidemiology: A Systematic Review

Abstract

The spread of the COVID-19 pandemic was spatially heterogeneous around the world; the transmission of the disease is driven by complex spatial and temporal variations in socioenvironmental factors. Spatial tools are useful in supporting COVID-19 control programs. A substantive review of the merits of the methodological approaches used to understand the spatial epidemiology of the disease is hardly undertaken. In this study, we reviewed the methodological approaches used to identify the spatial and spatiotemporal variations of COVID-19 and the socioeconomic, demographic and climatic drivers of such variations. We conducted a systematic literature search of spatial studies of COVID-19 published in English from Embase, Scopus, Medline, and Web of Science databases from 1 January 2019 to 7 September 2021. Methodological quality assessments were also performed using the Joanna Briggs Institute (JBI) risk of bias tool. A total of 154 studies met the inclusion criteria that used frequentist (85%) and Bayesian (15%) modelling approaches to identify spatial clusters and the associated risk factors. Bayesian models in the studies incorporated various spatial, temporal and spatiotemporal effects into the modelling schemes. This review highlighted the need for more local-level advanced Bayesian spatiotemporal modelling through the multi-level framework for COVID-19 prevention and control strategies.

Adaptive Gaussian Markov random field spatiotemporal models for infectious disease mapping and forecasting

“…Bayesian disease mapping methodology provides important tools for research into spatial epidemiology and ecology. The methodological applications are seen in the large and still growing literature on spatial and spatiotemporal analysis of COVID-19 infection and related health outcomes, and on applications of AR, RW, and CAR models in COVID-19 related research; see Nazia et al (2022) for a systematic review. As we broaden the scope of Bayesian disease mapping from modelling the risks of non-communicable diseases to the risks of communicable ones such as the COVID-19 infection and its outbreaks, and from disease risks prediction and inference to forecasting disease occurrences and spread, we also encounter real world data that necessitate increasingly flexible models that enable the characterisation of complex dynamics of spatial and spatiotemporal interactions, dependencies, heterogeneities, and discontinuities.…”

Waves in time, but not in space – An analysis of pandemic severity of COVID-19 in Germany based on spatio-temporal clustering

“…Studies focusing on the tempo-spatial aspects of the COVID-19 pandemic rely on the case incidence of patients who tested positive for COVID-19 almost exclusively as an indicator (Nazia et al 2022). This is somewhat problematic, especially since the testing regime changed throughout the pandemic and potentially also varied regionally.…”

The COVID-19 Mortality Rate Is Associated with Illiteracy, Age, and Air Pollution in Urban Neighborhoods: A Spatiotemporal Cross-Sectional Analysis

“…Accordingly, concentrations of microscopic matter, such as PM 10 and PM 2.5 , and various noxious air gases decreased significantly during the lockdown, resulting in improved air quality [ 19 , 20 ]. Demographic, socioeconomic, and climatic factors have been identified as influencing COVID-19 incidence and death rates [ 21 , 22 ]. The COVID-19 mortality rates have been associated with air pollution, highlighting the importance of investigating COVID-19 spread and mortality in relation to air quality [ 23 ].…”

“…Due to the limitations of the COVID-19 and risk factor data, the application of geospatial techniques was initially limited to cluster analysis via global and local Moran’s I, hotspot analysis, interpolation, and space-time scan statistics [ 26 , 27 ]. However, according to a recent systematic review by Nazia et al [ 21 ], a variety of spatial analytic techniques have been used to study COVID-19 in association with various risk factors, ranging from commonly used descriptive methods (85%) to Bayesian methods (15%). While the traditional frequentist method uses the likelihood function to derive parameter estimates, the Bayesian approach incorporates probability to measure uncertainties in estimates, prediction, or inference on posterior distributions by specifying priors [ 21 ].…”

“…Compared to other methods, spatial statistical modeling of geographically weighted regression (GWR) has been widely used to identify the drivers of COVID-19 spatial variations [ 21 ]. For example, ordinary least squares (OLS) regressions ignore spatial autocorrelation and heterogeneity.…”