Ranking non-pharmaceutical interventions against Covid-19 global pandemic using global sensitivity analysis—Effect on number of deaths

Abstract

In this study, we use Global Sensitivity Analysis (GSA) to rank four non-pharmaceutical interventions (NPIs) in a deterministic compartmental model that might control Covid-19 related deaths in the United States. The NPIs are social distancing, isolation of infected individuals, identifying asymptomatically infected individuals through testing, and the use of face masks. The model uses a fear-based behavioral model that leads unmasked susceptible individuals to wear masks. The model parameters are estimated from the reported deaths for the United States of America from March 1, 2020 to November 26, 2020. Two GSA tools, the Sobol’ sesntivity indices and Partial Rank Correlation Coefficient are used to obtain the rankings of the input parameters at different stages of the disease propagation. We found that social distancing and outward mask efficiency alone decreases the output uncertainty by 25-45%. Sobol’ second order indices show that the combined effect of social distancing with increased mask usage and identifying and isolating asymptomatically infected individuals decreases uncertainty an additional 10%.

Learning about COVID-19: sources of information, public trust, and contact tracing during the pandemic

“…Michigan was designated a hotspot in the first months of the pandemic [2]. Prior to the availability of an effective vaccine, other public health interventions like mask wearing, social distancing, isolation, quarantine, and contact tracing were primarily used to control the pandemic [3,4]. Contact tracing, a long-standing public health tool, involves identifying the network of individuals that encounter an infected person and screening them for symptoms of the disease.…”

Exploring how ecological and epidemiological processes shape multi-host disease dynamics using global sensitivity analysis

“…Global sensitivity analysis is a well developed methodology that has been widely applied in engineering settings where understanding reliability is very important (Saltelli, 2002;Saltelli et al, 2005;Wu, 1994). Within the past few decades, sensitivity analysis has begun to be used to assess model predictions, experimental design, and robustness in a variety of biological applications (Blower and Dowlatabadi, 1994;Jarrett et al, 2015Jarrett et al, , 2017bRenardy et al, 2019;Aggarwal et al, 2019Aggarwal et al, , 2021Marino et al, 2008;Wentworth et al, 2016;Hanthanan Arachchilage and Hussaini, 2021). One of the differences between more traditional applications and biological applications is in the interpretation of the rankings.…”

“…Our metric for computing global sensitivities is Partial Rank Correlation Coefficients (PRCC) (Marino et al, 2008; Hanthanan Arachchilage and Hussaini, 2021; Wentworth et al, 2016; Jarrett et al, 2017b; Hamby, 1995; Blower and Dowlatabadi, 1994). PRCC quantifies how the variations in the model parameters affect the values of the quantities of interest (QoIs).…”

“…This type of sensitivity analysis is referred to as ‘local’ in the sense that one parameter is varied at a time. Other widely used ‘global’ measures (which simultaneously vary multiple parameters) include Partial Rank Correlation Coefficient, Sobol’ measures, importance measures and screening methods (Sobol, 2001; Marino et al, 2008; Salelli et al, 2004; Saltelli et al, 2019; Jansen, 1999; Jarrett et al, 2017a; Hanthanan Arachchilage and Hussaini, 2021). Global approaches rely on quantifying the relationship between parameter and the QoI based on statistical relationships.…”

Stress and Stressors: The Impact of the COVID-19 Pandemic on Students, Faculty and Staff at a Historically Black College/University

“…On a daily basis, people face challenges that can be stressful, overwhelming, and cause strong emotions in adults and children. Since the beginning of COVID-19 in December 2019, public health actions, such as social distancing, became critically necessary to reduce the spread of the disease [ 2 , 3 ], but these measures can make people feel isolated and lonely and can increase stress and anxiety. The non-availability of treatment for COVID-19 during the early stages of its onset created much panic across the world.…”