Infectiousness of places: The impact of human settlement and activity space in the transmission of COVID-19

Places are fundamental factors in the spread of epidemics, as they are where people agglomerate and interact. This paper explores how different types of places, activity spaces at micro-level and human settlements at macro-level, impact the transmission of infections using evidences from COVID-19. We examine eleven types of activity spaces and find heterogeneous impacts across countries, yet we also find that non-essential activity spaces tend to have larger impacts than essential ones. Contrary to common beliefs, settlement size and density are not positively associated with reproduction numbers. Further, the impacts of closing activity spaces vary with settlement types and are consistently lower in larger settlements in all sample countries, suggesting more complex pattern of virus transmission in large settlements. This work takes first steps in systematically evaluating the epistemological risks of places at multiple scales, which contributes to knowledge in urban resilience, health and livability.

infections using evidences from COVID-19. We examine eleven types of activity spaces 23 and find heterogeneous impacts across countries, yet we also find that non-essential 24 activity spaces tend to have larger impacts than essential ones. Contrary to common 25 beliefs, settlement size and density are not positively associated with reproduction 26 numbers. Further, the impacts of closing activity spaces vary with settlement types and are 27 consistently lower in larger settlements in all sample countries, suggesting more complex 28 pattern of virus transmission in large settlements. This work takes first steps in 29 systematically evaluating the epistemological risks of places at multiple scales, which 30 contributes to knowledge in urban resilience, health and livability. Humans continue to migrate to large, dense urban settlements in the past century. The consequent 41 growth of cities brings benefits such as economies of scale and knowledge spillovers, but also 42 increases the vulnerability of human society to risks related to people's agglomeration and 43 interaction such as infectious disease, pollution and crime (1, 2), for which the on-going outbreak 44 of COVID-19 is a prominent example. In tackling these risks, places are important aspects, as the 45 human interactions giving rise to various risks are all associated with certain physical places. To 46 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted September 7, 2021. ; https://doi.org/10.1101/2021.09.02.21263012 doi: medRxiv preprint NOTE: This preprint reports new research that has not been certified by peer review and should not be used to guide clinical practice.

of 19
understand the role of different types of places in the spread of risks is important for targeted 47 policy making to contain the risks and enhance the resilience of cities. 48 In terms of places' epistemological risks, the COVID-19 pandemic provides worldwide data with 49 natural experiments to investigate the spread of infectious disease at different types of places. 50 Place characteristics at multiple scales could have an impact. At the macro-level, for instance, 51 dense settlements lead to physical proximity among residents which is likely to generate more 52 contacts, and large settlements connect more people-both might increase the dissemination of 53 diseases (3-5). At the micro-level, different activity spaces, such as restaurants, museums, sports 54 venues, are likely to be associated with different risks of disease spread, affected by the 55 socioeconomic interactions at these locations. 56 Despite of a large number of researches on the spread of COVID-19, the potentially varying 57 transmission risks at different places are still not systematically investigated. Macro-scale place 58 characteristics are rarely examined and evidences are mixed (3,4,6). For micro-scale places, the 59 most relevant body of research is those on the efficacy of non-pharmaceutical interventions on 60 COVID-19 which include the closure of various activity spaces (7-17). However, these works 61 mostly use broad categories when estimating the impact of activity spaces, such as "non-essential 62 businesses", "venues" or "lock down", which can involve a number of distinct types of spaces (7, 63 10, 14, 18, 19). Since the closure of any activity space could affect the daily life of certain groups, 64 the effect estimates on broad categories are insufficient for governments to make cost-effective 65 intervention policies. As COVID-19 persists and countries have to lock down repeatedly, it is 66 critical that we understand which types of settlements and activity spaces need more rigorous 67 interventions than other, thus refining policies in the on-going COVID-19 and similar crisis in the 68 future. 69 In this work, we take the case of COVID-19 and investigate how different types of macro-level 70 settlement characteristics and micro-level activity spaces impact the spread of infections, as well 71 as how they interact. We examine two basic characteristics of settlements-population size and 72 density-which have been found to affect many social quantities (5,20), and eleven types of 73 activity spaces commonly included in government interventions, which are schools, childcare 74 centers, offices, non-essential retails, restaurants, bars, entertainment venues, cultural venues, 75 religious venues, indoor sports venues and outdoor sports grounds (detailed descriptions in Table   76 S1). To perform the analysis, we combine data from a variety of sources including COVID-19 77 infection case data from government data portals and open data repositories, government 78 intervention data from national and state-level government websites, and socioeconomic 79 characteristics of settlements from various official statistics (see Section S1 for details). Four 80 countries from different continents which are strongly hit by the pandemic are chosen as study 81 cases, which are Japan in Asia, the United Kingdom in Europe, the United States in North 82 America and Brazil in South America. We take settlements with population above 100,000 as 83 samples, since many smaller settlements do have enough cases to derive reliable estimates of 84 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted September 7, 2021. ; https://doi.org/10.1101/2021.09.02.21263012 doi: medRxiv preprint 3 of 19 instantaneous reproduction number, which is the outcome of concern in our analysis. We use data 85 from the first pandemic wave, that is, from March to August 2020, since there could be more 86 uncertainties confounding the analysis in later periods of the pandemic including the so-called 87 "lockdown fatigue", virus variants, and vaccination (9). however, infection case data can only be consistently acquired at this level in Japan (21). 98 Nonetheless, we prove that the results are not likely to be affected by the issue (Section S3.3). 99 The spatial units in each country with population above 100,000 are taken as sample, leading to 100 45 spatial units in Japan, 234 in the United Kingdom, 308 in the United States and 319 in Brazil 101 after cleaning missing data. 102 Our methodology is based on an econometric approach called difference-in-differences (DiD), 103 which is widely used in examining causal relationship in social processes (22). We first estimate 104 the causal impacts of activity space closures on the course of the epidemic, which can also be 105 interpreted as the risks of virus transmission associated with respective spaces. This is 106 implemented by modelling the relationship between instantaneous reproduction numbers (R t ) in 107 the spatial units and the corresponding status of activity spaces, controlling for other government 108 interventions including stay-at-home-orders and gathering bans. The DiD method estimates causal 109 impact by subtracting the course of R t in spatial units where a certain type of activity space get 110 closed or reopened with the course of R t in spatial units where the same type of activity space 111 remain unchanged, given that R t in the two groups should move in parallel trend absence of the 112 change. By subtracting the trends, the method can rule out the impact of common behavioral 113 changes shared by all spatial units such as increased self-protection, which could otherwise be 114 falsely attributed to activity spaces thus inflate the estimates (10). We estimate separate models on 115 each country to allow for heterogeneous impacts of activity spaces across countries, which might 116 be influenced by factors including the lifestyle, culture, urban form and building design. 117 The fact that governments might close or reopen multiple types of activity spaces together poses 118 obstacles for identifying the impact of individual types of activity spaces (14). But as time went, 119 there were more timing differences, especially in the reopening stage, to facilitate more nuanced 120 analysis. Correlation analysis shows that the Kendall's correlation coefficients between the status 121 of activity spaces are mostly smaller than 0.8 in our study period (Fig. 1). We merge activity 122 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted September 7, 2021. ; https://doi.org/10.1101/2021.09.02.21263012 doi: medRxiv preprint 4 of 19 space closures with correlation coefficients larger than 0.95 as one intervention, after which there 123 are at least 180 unit-day differences between any pair of interventions in Japan (due to only 45 124 spatial units) and 699 in the other countries. We further verify that the estimated impacts are not 125 sensitive to removing intervention variables, suggesting manageable collinearity (Section S3.2). 126 The DiD estimation is implemented through a two-way fixed effect model with fixed effects of 127 days and spatial units, which is a widely used modelling method to implement DiD analysis (23). infections related to the respective activity spaces, which may either happen inside these places or 155 on the way travelling to these places. 156 Most of the activity space closures satisfy the parallel trend assumption, meaning that the 157 estimates are not biased by potentially different pre-trends of R t in areas that close or reopen an 158 activity space and those that do not (detailed methodology and results of the parallel trend test in 159 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted September 7, 2021. ; https://doi.org/10.1101/2021.09.02.21263012 doi: medRxiv preprint 5 of 19 Section S2.2 and Table S3). The estimates are also generally robust to a number of alternative 160 settings in the analysis, including withholding spatial units from the sample and increasing or 161 decreasing confounding variables in the model, suggesting that they are not likely to be affected 162 by individual influential spatial units and the correlation among variables (detailed methodology 163 and results in Section S3.1 and S3.2). 164 Considering that governments often need to devise strategies on closing a set of activity spaces in 165 an epidemic, we further estimate the combined effects of multiple activity spaces. We compute 166 the effects and uncertainties of closing all possible combinations of activity spaces in each 167 country based on the modelling results in the last step. The full results are provided on this 168 project's Github repository https://github.com/lunliu454/infect_place for readers to explore. Here 169 we present the maximum reduction in R t that can be achieved by closing a given number of types 170 of activity spaces (Fig. 3). Our analysis suggests that the largest reductions in R t are achieved by 171 closing two to six types of activity spaces, while more closures do not further bring reproduction 172 numbers down. Governments could resort to this kind of analysis when making cost-effective 173 intervention strategies. 174 The combinations that generate the maximum effects are: closing schools, childcare centers, 175 entertainment, religious, indoor sports venues and outdoor sports grounds in Japan (67%, 46 ~ 176 80%); closing non-essential retails, restaurants, cultural venues and indoor sports venues in the 177 United Kingdom (plus banning indoor gatherings whose effect is inseparable, 75%, 45 ~ 88%); 178 closing schools, restaurants, entertainment and indoor sports venues in the United States (29%, 9 179 ~ 45%); and non-essential retails and outdoor sports grounds (27%, 13 ~ 38%) in Brazil. Closing 180 these sets of activity spaces could bring R t below 1 for previous R t up to 3.0 (1.9 ~ 5.0) in Japan, Brazil. Note that the few activity space closures not satisfying the parallel trend assumption are 183 excluded from this joint effect analysis since their impact estimates are not reliable (Table S3), 184 but they may actually be able to further contribute to the reduction of R t .

185
Impact of macro-level settlement characteristics 186 The results on the relationship between settlements' population size and density and the fixed 187 effects of spatial units are fairly consistent across the four countries (Table 1) and R t might include better health infrastructures in large cities and people's stronger awareness 196 of the risk thus more cautious behavior (24, 25). 197 Interaction between settlement characteristics and activity space closures 198 We also examine the interaction between macro-level settlement characteristics and activity space CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted September 7, 2021. ; https://doi.org/10.1101/2021.09.02.21263012 doi: medRxiv preprint 7 of 19 virus transmission. To our knowledge, this study is the first to attempt to estimate the impacts of 232 closing individual types of activity spaces and to identify the varying effects of activity space 233 closures in different types of settlements. 234 Though the risk of virus transmission at different types of activity spaces can also be evaluated 235 with mechanistic modelling (27, 28), actual human behavior could be more complicated than 236 experimental settings and our work provides data-driven evaluations. Our results demonstrate that 237 closing a selected set of activity spaces could reduce R t by 27-75% in our sample countries, 238 without imposing a full lock down. Actually, the stay-at-home order does not demonstrate a 239 statistically significant impact in reducing R t in three of the four countries after controlling for 240 other interventions and day and unit fixed effects (Table S2). This contradicts some previous 241 findings, which however either do not rule out simultaneous voluntary behavioral changes or omit 242 certain confounders (14,15,19). The magnitudes of the impacts are heterogeneous across demonstrate statistically significant effects in reducing R t and non-essential activity spaces tend to 253 be more effective, suggesting that governments could consider closing non-essential activity 254 spaces as cost-effective interventions. 255 At the macro-level, our findings on the impacts of settlement size and density on R t contradict the 256 common belief that large and densely populated cities are more vulnerable to infectious disease 257 (31). This could either because the seemingly increased connectivity and proximity among people 258 do not actually enhance the chance for one infected person to transmit the virus to others, or such 259 effect does exist but is offset by other positive factors of large and dense cities such as more 260 healthcare resources driven by economy of scale and more cautious behavior of people. The exact 261 causal chain could also involve the demography, economy even partisanship in different types of 262 settlements (25, 32), which is subject to further study. Either way, these results lend more 263 confidence to encouraging agglomeration of people and high-density development, as in the end, 264 they are not associated with higher epistemological risks. 265 The finding that specific activity spaces account for a smaller proportion of transmission in 266 relatively large settlements suggests that the pattern of virus transmission in these settlements is 267 more complicated, which might be related to longer travel, more contacts on public transit and 268 streets, home transmission in crowded residences and so on. It indicates that governments might 269 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted September 7, 2021. need to take extra measures other than locking down to contain the pandemic in large cities, such 270 as contact tracing or providing assistance to those living in poor conditions. 271 There are a number of limitations associated with our methodology. In terms of the causal 272 identification strategy, the DiD method requires both parallel trend and exogeneity of the 273 treatment. While the parallel trend assumption is examined with an event-study design, the 274 exogeneity assumption could be challenged by unobserved confounders that affect both R t and 275 activity space closures. Though we are able to rule out a number of confounders by including a Third, we assume linear relationship between R t and the independent variables in the entire 288 analysis, which is a convenient assumption made by many studies on intervention effects in 289 10,12,16,18). However, the impact of closing one type of activity space may rely is no longer only about confined areas such as hospitals, residences or the water supply system, 300 but also the entire urban space. Understanding the linkage between places, human activities and 301 diseases would be important for long-and short-term policy making in public health, urban 302 planning, urban economy and other relevant fields.

306
. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted September 7, 2021. ;https://doi.org/10.1101https://doi.org/10. /2021 9 of 19 Data. We curate a data set combining daily infection cases, government interventions (including 307 activity space closures, stay-at-home orders and gathering bans) and the spatial, demographic and 308 economic characteristics of the spatial units in our study, from the onset of the pandemic till 309 August 15 2020. The infection case data are sourced from Japan Broadcasting Corporation's case 310 reports, the UK government, Johns Hopkins University and the Brazilian Ministry of Health. The 311 timetable of government interventions is manually collected from the websites of national and 312 state-level governments, which are the main levels of government making decisions on 313 interventions. The settlement-related information is gathered from a number of official websites. 314 More details on data sources are provided in Section S1.  Table S1); α c,i and τ c,t denote the unit and time fixed effects, respectively; 328 and ε c,i,t denotes the error term. For the uncertainty over the parameters, we estimate robust 329 standard errors allowing for ε c,i,t to cluster at the unit level, to account for heterogeneity in the 330 treatment effects (38).

342
. CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

(which was not certified by peer review)
The copyright holder for this preprint this version posted September 7, 2021. ; https://doi.org/10.1101/2021.09.02.21263012 doi: medRxiv preprint 10 of 19 Estimating impacts of macro-level settlement characteristics. We take the unit fixed effects 343 estimated by Eq. 1, which can be interpreted as the intrinsic reproduction number in each spatial 344 unit, and model their relationship with the size and density of settlements while controlling for the 345 proportion of elder population (over 65 or 60 years old depending on data availability), proportion 346 of Black and Asian (in the United Kingdom and the United States only), the average income of 347 residents and the per capita gross domestic product, using simple linear regression. is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted September 7, 2021. ; https://doi.org/10.1101/2021.09.02.21263012 doi: medRxiv preprint   Table S4). The variance inflation factors are all below 7. 582 583 584 . CC-BY-NC-ND 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) The copyright holder for this preprint this version posted September 7, 2021. ; https://doi.org/10.1101/2021.09.02.21263012 doi: medRxiv preprint