Now that many countries and regions have applied lockdown strategies to mitigate the effects of COVID-19, researchers are wondering if better or more targeted strategies could have been devised. In light of the lessons learned, spatial modeling is emerging as one area that may provide better answers and strategies the next time we are faced with a major pandemic.
Researchers studying infectious diseases have typically applied susceptible-infected-recovered (SIR) models to look at how likely infectious diseases spread in communities. Many of these models have lacked clear spatial dimensions or at least a consistent use of spatial coverage. Individual-based, or agent-based (ABMs), models have been shown to represent bottom-up processes, using individual behaviors and actions, as a basis for spatially-explicit models that can potentially be used to predict likely outcomes of infection spread. Other models apply top-down or aggregate approaches, such as system dynamic, Monte Carlo, or finite element models. These different forms of models have their advantages, with ABMs being far more detailed and able to demonstrate how individual behaviors can affect disease spread, while aggregate models are generally easier to test. Regardless of the approach, what is critical is that relatively small spatial units, which can act independent or explicit from other regions, are the best approach to modeling pandemic spread, as it allows circumstances within small areas to best affect modeling. This enables models to also be applicable at small spatial scales, that is sub-national scales, so that lockdown measures or strategies could be applied more specifically to an area, such as cities, neighborhoods, or small regions to see how effective they might be.
Where studies and empirical data have been applied using spatially-explicit models on given countries, such as Italy, it has been shown transmission of COVID-19 has significantly dropped when overall mobility is dropped (in the case of Italy over 40% reduction in transmission). In this case, a network-based susceptible–exposed–infected–recovered (SEIR) connectivity model that looks at provinces and sub-regions connected based on mobility pattens can be tunned to look at mobility rates between areas. The model was able to show that measures used in given regions to restrict movement helps to drastically limit infection rates by simply limiting mobility. Nevertheless, the model is not very fine-scaled, which limits its utility at urban or specific community levels. One approach used spatial regression modeling to show that demographics, rather than only mobility, have a major impact on mortality with COVID-19. Regions that had much older populations tended to have far higher death rates in Europe. Thus, measures that also targeted vulnerable populations, as well as restrict mobility in given regions, are also more successful in limiting overall deaths. Another spatial model showed that distance to testing sites for COVID-19 could be a factor in transmission. Areas with low population density or poor testing access could have disproportional higher COVID-19 rates based on the fact testing was not as frequent and community transmission became high without adequate testing to determine who could infect others. By placing testing sites more evenly across areas, irrespective of population density and healthcare facilities, then infection rates could have been lower and spread could have been more limited in parts of the United States. This study highlights how unequal healthcare is also a major spatially-based factor in the spread of COVID-19, as minority and poorer communities were also found to have generally lower access to testing despite the fact that studies have suggested these populations could be potentially more vulnerable to death from infection than other demographic groups.
Overall, spatially-explicit models highlight how COVID-19 has various spatial factors that can affect its spread and death rates. Regional connectivity, mobility, demographic patterns, and areas where tests are available have been shown to affect the rate of infection and death. Creating spatially-explicit models will mean that for future pandemics these and other factors should be represented in models in order to increase the efficacy of SEIR/SIR modeling. Models also need to be more fine-scale and able to handle how specific circumstances in regions, such as demographic differences, could affect overall infection and death rates.
 For more on spatially explicit models using ABM approaches and discussion on other methods, see: O’Sullivan, D., Gahegan, M., Exeter, D., Adams, B., 2020. Spatially‐explicit models for exploring COVID‐19 lockdown strategies. Transactions in GIS tgis.12660. https://doi.org/10.1111/tgis.12660.
 For more on the network-based spatial model in COVID-19 spread in Italy, see: Gatto, M., Bertuzzo, E., Mari, L., Miccoli, S., Carraro, L., Casagrandi, R., Rinaldo, A., 2020. Spread and dynamics of the COVID-19 epidemic in Italy: Effects of emergency containment measures. Proc Natl Acad Sci USA 117, 10484–10491. https://doi.org/10.1073/pnas.2004978117.
 For more on demographic factors affecting morality in spatial models of COVID-19, see: Sannigrahi, S., Pilla, F., Basu, B., Basu, A.S., 2020. The overall mortality caused by COVID-19 in the European region is highly associated with demographic composition: A spatial regression-based approach. arXiv:2005.04029 [q-bio].
 For more on the location of testing sites for COVID-19 and the effect of unequal testing, see: Rader, B., Astley, C.M., Sy, K.T.L., Sewalk, K., Hswen, Y., Brownstein, J.S., Kraemer, M.U.G., 2020. Geographic access to United States SARS-CoV-2 testing sites highlights healthcare disparities and may bias transmission estimates. Journal of Travel Medicine taaa076. https://doi.org/10.1093/jtm/taaa076.