key: cord-0866535-97my2qve authors: Lacson, R.; Veldkamp, P.; Zapanta, C. title: Assessing the Impact of Mask Usage on COVID-19 Transmission Using a Computer Simulation date: 2021-06-09 journal: nan DOI: 10.1101/2021.06.08.21258593 sha: 5118413c3f746becbad90293f96e8eb04fda5c80 doc_id: 866535 cord_uid: 97my2qve Background: COVID-19, caused by SARS-CoV-2, is highly contagious and causes substantial morbidity and mortality. Mask usage has been advocated by health professionals to minimize its spread. Thus, it is important to develop a simulation that models SARS-CoV-2 spread in indoor environments to evaluate mask usage effectiveness. Methods: A visual computer simulation was developed with Pygame in Python 3. A virtual indoor supermarket is simulated by a given flow of customers with an initial infection percentage and mask usage percentage who enter, move around, and exit a supermarket with shelves, tables and cashiers to demonstrate a systems dynamic complexity, i.e. nonlinear interactions of system elements over time. A supermarket was simulated with initial infection rates of 5%, 10%, and 20% and mask use percentages of 0%, 25%, 50% 75%, and 100%. The environmental settings (e.g. shelf number and location) and total customers (N=200) were kept constant. Results: The number of infected customers increased as the percentage of mask usage decreased (p < 0.01). At 5% initial infection, almost no infections were observed at 50% mask usage, with a logarithmic best-fit model (R2 = 0.947). At 10% initial infection, the association between mask usage and decrease in number of infections was best fit with a linear model (R2 = 0.924). For 20% initial infection, a quadratic model was the best fit (R2 = 0.934). While a linear model suggests proportional decreases in infection, the quadratic model suggests more significant reductions in infections at higher rates of mask use (i.e. increasing mask usage from 5% to 10% is less impactful than from 65% to 70%). Conclusion: The results suggest that mask usage has a significant impact on decreasing COVID-19 transmission. Ideally, mask usage should be as high as possible to achieve more significant reductions in COVID-19 infections. Various parameters can be adjusted during simulation as we learn more about SARS-CoV-2 to guide policies for minimizing COVID-19 transmission. A new type of coronavirus called SARS-CoV-2 was identified in December 2019, which led to the novel COVID-19 infection. 1, 2 This was classified by the World Health Organization (WHO) in January 2020 as a public health emergency of international concern, 3 and in March 2020 as a pandemic. 4 COVID-19 has impacted human lives, suffering, and behavior, specifically health-related behavior. The disease has spread to over 100 countries resulting in substantial mortality and morbidity. [4] [5] [6] As the pandemic has progressed throughout the United States, mask use has become a divisive political issue, despite the large body of evidence supporting its effectiveness to reduce SARS-CoV-2 transmission. 7 With the induction of President Joe Biden as the 46th President of the United States, Biden has signed an executive order to mandate mask use on federal grounds and will advocate for a "100-day mask challenge" to depoliticize mask use. 8 Thus far, simulations and models to predict SARS-CoV-2 transmission have been developed using the Susceptible-Exposed-Infective-Recovered (SEIR) model showing viral transmission as a function of time using factors such as population, average number of interactions, the basic reproduction constant, and incubation period. [9] [10] [11] However, no simulation has been designed based on physical and spatial interactions between a population and the viral particles emitted by infectious individuals. Furthermore, no simulation attempts to model SARS-CoV-2 transmission in indoor environments, which are enclosed, leading to higher concentrations of airborne droplets. 12 . 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 June 9, 2021. ; https://doi.org/10.1101/2021.06.08.21258593 doi: medRxiv preprint Several strategies have been proposed to minimize viral transmission, especially indoors. In supermarkets, restricting occupancy, one way traffic along aisles, physical distancing and mask usage have been implemented. 11 The impact of these strategies, however, have not been rigorously assessed. The purpose of this study is to evaluate the effectiveness of masks in indoor environments through a computer simulation that uses digital representations of viral particle spread. The infection count per simulation was tracked and modelled to determine how increasing mask usage affects viral transmission. A visual computer simulation was developed with Pygame in Python 3 to model the spread of SARS-CoV-2 in a supermarket. This simulation relied primarily on the physical interactions between customers and viral particles. In the visual simulation, red dots represent approximately 10 viral particles. When red dots come in contact with a customer, a counter is added to an individual if they do not have a mask. If a customer has a mask on, however, they received 0.8 counters as per the approximately 20% filter effectiveness for airborne particles observed for a common cloth mask. 13 When 28 counters are accumulated, a customer becomes infected. This was based on the proposed infectious dose of SARS-CoV-2 being approximately 280 viral particles. 14 The respiratory rate of all customers was 20 breaths per minute. Each customer exhaled between 10-20 particles per exhalation, but this number is reduced on average by 80% for customers with masks, as masks filter approximately 80% of exhaled particles. 13 Red dots . 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. The copyright holder for this preprint this version posted June 9, 2021. ; https://doi.org/10.1101/2021.06.08.21258593 doi: medRxiv preprint travelled in a straight parallel path in front of the customer, the normal movement pattern for particles being exhaled or sneezed. 15 Customers entered at random intervals with a predetermined maximum occupancy of 60 people (plus 3 store employees). The supermarket has 8 store shelves arranged in parallel, and three cashiers (i.e. store clerks) at the front of the store. Each customer had a random number of shelf locations to visit (between 1 and 6), and once all locations were visited, they approached a cashier. Each customer moved at 2 pixels per frame. An A* path-finding algorithm was implemented to optimize customer movement from location to location without walking through shelves and tables. The A* algorithm is a dynamic search optimization algorithm commonly used in graph theory. 16 Figure 1 shows a sample screenshot of the simulation. . 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. The copyright holder for this preprint this version posted June 9, 2021. ; https://doi.org/10.1101/2021.06.08.21258593 doi: medRxiv preprint The number of new infections are included in Table 1 below. There is a strong negative correlation between mask use percentage and number of new infections. This supports the current knowledge suggesting mask use as an effective measure to reduce SARS-CoV-2 transmission. To fit a graph to the above data, the correlation coefficient R 2 for each manipulation of mask use percentage is included in Table 2 . The model with best fit (highest correlation coefficient) was logarithmic for 5% initial infection (R 2 = 0.947), linear for 10% initial infection (R 2 = 0.924) and quadratic for 20% initial infection (R 2 = 0.934). There was a statistically significant association between mask usage and number of new infections (p < 0.01). The logarithmic, linear, and quadratic models for 5%, 10%, and 20% initial infection respectively, are graphed in Figure 2 . . 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. The copyright holder for this preprint this version posted June 9, 2021. ; https://doi.org/10.1101/2021.06.08.21258593 doi: medRxiv preprint . 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. The copyright holder for this preprint this version posted June 9, 2021. ; https://doi.org/10.1101/2021.06.08.21258593 doi: medRxiv preprint Figure 2b : Graph of linear model over the raw data for 10% initial infection. . 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. The copyright holder for this preprint this version posted June 9, 2021. ; https://doi.org/10.1101/2021.06.08.21258593 doi: medRxiv preprint As mask use increased, the number of customer infections decreased, almost entirely diminished at 50% mask use at 5% initial infection, and according to a logarithmic model at 5% initial infection, a linear model at 10% initial infection, and a quadratic model at 20% initial infection. The negative association between mask use and viral transmission was to be expected given that mask use both provides slight protection to users and an even larger protection to those around an infected individual. Furthermore, this conclusion agrees with the common idea that mask use is an effective measure to reduce SARS-CoV-2 transmission, even in indoor environments. In previous studies, [17] [18] [19] mask usage was an independent factor that was associated with an increased odds of transmission control with an odds ratio of 3.5. 19 Based on the significant reduction in infections at 50% mask usage for 5% initial infection and the linear model for 10% initial infection, mask use should be encouraged due to the direct effect it has on reducing new infections. However, based on the quadratic model for 20% initial infection, it is important to note that more significant reductions in infections can be observed at higher rates of mask use (i.e. increasing mask usage from 5% to 10% is less impactful than from 65% to 70%). This suggests that in order to have a 50% reduction in infections, the required mask usage should be higher. Based on the quadratic model in Figure 3 , this happens at approximately 66% mask usage with more significant decreases in viral transmission occurring at these higher mask use percentages. . 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. The copyright holder for this preprint this version posted June 9, 2021. ; https://doi.org/10.1101/2021.06.08.21258593 doi: medRxiv preprint While this simulation provides a solid foundation for modelling the spread of SARS-CoV-2 in indoor environments, there are many additional factors that could be added to further improve the accuracy of this simulation. One such factor is re-circulation which could circulate viral particles in indoor environments. This could lead to transmission from infected individuals to others outside the range of exhalation. 12 Another factor to include is the incubation period. Under this simulation, infected individuals immediately begin exhaling viral particles when there is an incubation period before such spread occurs significantly. 20 Under simulated conditions, mask usage was demonstrated to be effective in reducing viral spread in indoor environments. Ideally, mask usage should be as high as possible to achieve more significant reductions in COVID-19 infections. Additionally, in order to improve the accuracy of this simulation, adjustments could be made to account for ventilation, incubation period, and specific environment factors. Once the adjustments are included, further simulation . 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 June 9, 2021. ; https://doi.org/10.1101/2021.06.08.21258593 doi: medRxiv preprint Asymptomatic cases in a family cluster with SARS-CoV-2 infection Temporal dynamics in viral shedding and transmissibility of COVID-19 Timeline: WHO's COVID-19 response Fair Allocation of Scarce Medical Resources in the Time of Covid-19 Impact of COVID-19 on the Cardiovascular System: A Review of Available Reports. Cardiovasc Drugs Ther Asymptomatic and Presymptomatic SARS-CoV-2 Infections in Residents of a Long-Term Care Skilled Nursing Facility The Battle Between The Masked And The Masked-Nots Unveils Political Rifts Inside Joe Biden's Agenda for His First 100 Days To mask or not to mask: Modeling the potential for face mask use by the general public to curtail the COVID-19 pandemic A Simulation on Potential Secondary Spread of Novel Coronavirus in an Exported Country Using a Stochastic Epidemic SEIR Model Simulation of coronavirus disease 2019 (COVID-19) scenarios with possibility of reinfection How can airborne transmission of COVID-19 indoors be minimised? Aerosol Filtration Efficiency of Common Fabrics Used in Respiratory Cloth Masks Development of a Dose-Response Model for SARS Coronavirus Airflow Dynamics of Human Jets: Sneezing and Breathing -Potential Sources of Infectious Aerosols Generalized best-first search strategies and the optimality of A* Face coverings and respiratory tract droplet . 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 An evidence review of face masks against COVID-19 Mask-wearing and control of SARS-CoV-2 transmission in the USA: a cross-sectional study. Lancet Digit Health The Incubation Period of Coronavirus Disease 2019 (COVID-19) From Publicly Reported Confirmed Cases: Estimation and Application Investigation of novel SARS-CoV-2 variant: Variant of Concern 202012/01 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 can be run to test the effectiveness of other preventative measures, both in isolation and in combination with other measures. It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. The copyright holder for this preprint this version posted June 9, 2021. ; https://doi.org/10.1101/2021.06.08.21258593 doi: medRxiv preprintThe copyright holder for this preprint this version posted June 9, 2021. ; https://doi.org/10.1101/2021.06.08.21258593 doi: medRxiv preprint