The importance of widespread testing for COVID-19 pandemic: systems thinking for drive-through testing sites

ABSTRACT

On 11 March 2020, the World Health Organisation (WHO) declared COVID-19 a pandemic. Early epidemiological estimates show that COVID-19 is highly transmissible, infecting populations across the globe in a short amount of time. WHO has recommended widespread clinical testing in order to contain COVID-19. However, mass testing in emergency department (ED) settings may result in crowded EDs and increase transmission risk for healthcare staff and other ED patients. Drive-through COVID-19 testing sites are an effective solution to quickly collect samples from suspected cases with minimal risk to healthcare personnel and other patients. Nevertheless, there are many logistical and operational challenges, such as shortages of testing kits, limited numbers of healthcare staff and long delays for collecting samples. Solving these problems requires an understanding of disease dynamics and epidemiology, as well as the logistics of mass distribution. In this position paper, we provide a conceptual framework for addressing these challenges, as well as some insights from prior literature and experience on developing decision support tools for public health departments.

1. Introduction

Coronavirus Disease 2019 (COVID-19) is caused by a virus strain, SARS-CoV-2 (CDCa: Centers for Disease Control and Prevention, 2020). It was first detected by the end of December 2019 in Wuhan, China, and shortly after, it has spread to nearly every country. By early April 2020, more than 1.5 million people have been infected with COVID-19 and it has caused over 88 thousand deaths worldwide. These numbers continue to rise every day (Johns Hopkins, 2020). The main mode of transmission is through respiratory droplets expelled from the mouth or nose of an infected person and inhaled by a healthy person (CDCb: Centers for Disease Control and Prevention, 2020).

The exponential growth of cases with COVID-19 is due to its high transmissibility. The basic reproductive number (R₀) of COVID-19, which is the average number of secondary cases generated from a single infectious case in a completely susceptible population, was estimated to be 2.2 in early studies (Li et al., 2020). In order to suppress community spread, WHO recommends countries develop targeted measures by following six steps: expand, train and deploy the public health workforce; implement a system to find every suspected case; ramp up testing capacity and availability; identify and adapt key facilities that will be used to treat and isolate patients; develop a clear plan to quarantine contacts; and refocus national policy on suppression and containing of COVID-19 (WHO: World Health Organization, 2020). Therefore, widespread and accurate COVID-19 testing is crucial to slowdown the infection rate and relieve congestion for Emergency Departments (EDs) and Intensive Care Units (ICUs) in hospitals.

Drive-through testing sites can be an effective option to quickly collect samples from suspected cases and have been successfully implemented in some countries, e.g., in South Korea, and the United States has also been expanding their implementation (Cohen & Kupferschmidt, 2020; Dyer, 2020). The effectiveness of testing sites to quickly identify suspected cases and confirm the disease depends on allocating resources (testing kits and healthcare staff) at the right place, with the right amount, and at the right time. Effective supply chain management and execution of these operations can influence epidemic dynamics and reduce the burden of the disease in the communities. Nevertheless, drive-through testing sites are not isolated from the emergency response system. Moreover, the deployment of these sites requires resources that must be obtained from other healthcare services (e.g., labs, EDs, hospitals). Consequently, a systemic perspective is needed in order to adequately optimise the scarce public health resources and increase the impact of the mitigation plan.

In this paper, we discuss key findings in the deployment and operationalisation of drive-through testing sites based on prior research and propose a systematic framework that integrates dynamic mathematical epidemiology models with healthcare delivery operations to improve public health emergency response efforts.

2. Drive-through sites in a public health emergency

The testing process in a drive-through testing site consists of two sequential steps: registration and sample collection, as shown in Figure 1. At the registration station, the patient receives instructions and answers a questionnaire soliciting important information. During sample collection, the patient provides the sample for the test.

Figure 1.

Drive-through testing process flow.

There are potentially challenging issues to be addressed in the deployment process of drive-through testing sites. These include determining the number of sites to open, the location of these sites and allocation of staff to each site, and safety of staff. The solutions to these issues can affect the performance of the response plans in terms of quickly identifying infectious COVID-19 cases for isolation and treatment, then contact tracing. Former public health emergency preparedness exercises and related literature provide important information for response efforts during this time of crisis (Lee, Chen et al., 2009; Lee, Smalley et al., 2009; Nelson et al., 2008; Whitworth, 2006).

Some of these studies suggest considering not only the size of the population to be served at each site, but also their socio-demographic characteristics. Differences in socio-demographic and health status across communities, e.g., age distribution and proportion of the population with underlying medical conditions, may affect the variability of testing time in sites and, consequently, boost waiting times (Ramirez‐Nafarrate et al., 2015). Racial disparities have become increasingly evident as the socio-demographic profiles of hospitalisation cases and deaths are reported (NPR, 2020). Therefore, using a standard configuration of sites, or assuming all sites will have equal demand and demand profiles may lead to inefficient and delayed processes. Figure 2 shows a visualisation of site distributions in the Phoenix metropolitan area in Arizona used for mass dispensing operations for a public health preparedness exercise. These type of visualisations assist decision-makers to identify areas of special concern either due to population density to avoid unacceptable service times in sites, or to sociodemographic and health characteristics of the communities (e.g., communities formed by people with limited English proficiency, hearing and visual impairments and other special needs or high prevalence of chronic complications). Thus, by identifying community vulnerabilities decision-makers can dynamically adjust the amount and type of resources in the region to meet public health needs. The CDC Vulnerability Index mapping tool is one example of a resource that can help planners identify communities that may need support in preparing for, or recovering from, disasters (CDCd: Centers for Disease Control and Prevention, 2020).

Figure 2.

Visualisation of site allocation in Phoenix, Arizona. Left figure shows a restricted area to allocate sites (circle) and the colour of the sites shows congestion in the site (Ramirez‐Nafarrate et al., 2015). Right figure shows the allocation of sites throughout the city, size of the marker shows the number of staff allocated (Ramirez‐Nafarrate et al., 2019).

In addition, the arrival pattern of patients to sites also affects testing system performance. If the system is designed by using the expected arrival rate to determine staffing and resource allocation across sites, the actual waiting times would be underestimated. Therefore, it is important to predict specific site demand, and dynamically add and re-allocate staff based on the needs of the sites over time. Additionally, the response plan should consider if individuals will be allowed to select which site to get their test (Allocation Reaction Only-AR), or if they will be assigned to a specific site (No Consumer Reaction-NR) (Berman & Krass, 2015). In the AR model, individuals maximise a utility function incorporating travel and waiting times across sites. Hence, individuals may select a site geographically further away if it is less crowded than a nearer one, which seems to be the most feasible and appropriate case for the COVID-19 testing operations. In the NR model, individuals would be assigned to a specific site. Current information and communication technologies could simplify the appropriate assignment of people to sites. Furthermore, the assignment could also specify a time period during which to visit the site in order to manage the utilisation of the site over time. Although NR can be seen as more effective for the public health practitioners to execute, it may raise some political and social concerns in the community.

In addition to healthcare staff, the rapid diagnostic kits for COVID-19 are still a scarce resource for public health systems in many parts of the world (Pang et al., 2020). Several existing diagnostic kits are for research purposes only, and the sensitivity and specificity of these tests are still under investigation. However, these rapid diagnostic kits may confirm a suspected contagious case in less than a day. The test accuracy and speed are important factors that can influence the effectiveness of the public health mitigation policy; therefore, modellers should consider these factors in their analyses as well.

These operating principles may help to anticipate the needs for other healthcare resources, such as ventilators, ED and ICU beds, as well as designing interventions to minimise the spread of the disease. The next section presents a framework to combine all these decision-making opportunities in a systematic way.

3. A systematic framework for public health emergency response

Dealing with a public health emergency that threatens the lives of millions, such as the current COVID-19 pandemic, requires an integrated and systematic framework for a quick and effective response. Epidemiological simulation models can help predict and assess disease progression in a population. Understanding the disease spread patterns may help to design effective and timely interventions and policies to mitigate risk and improve outcomes. These types of models also have been used to evaluate the impact of policies, such as social distancing and vaccine distribution (Araz, Damien et al., 2012; Araz, Galvani et al., 2012; Araz et al., 2011, 2013). This pandemic has been a challenge for public health practitioners because of COVID-19’s uncertain disease progression characteristics (e.g., asymptomatic transmission, longer incubation period) as well as unforeseen supply chain challenges. China used to make half of the world’s surgical masks before COVID-19 emerged there, and one of the major manufacturers that produces the nasal swabs needed to test patients are produced in Lombardy, Italy, another area hit hard by the virus. The reliance of the US on these external supply chains has resulted in national shortages of personal protective equipment (PPE) for health care workers and testing kits for patients. Therefore, we propose a systematic framework that incorporates multiple types of models to assess public health risks associated with the pandemic and to assist decision-making regarding the supply management and logistics of case identification, isolation for pandemic suppression and mitigation. Figure 3 represents the integration of these models. Disease transmission models generate projections on infections, hospitalisations, deaths, etc., using the force of infection which is a function of average number of contacts in the population, transmission probability of the virus strain and the proportion of infectious individuals in the population. Physical contacts between individuals is a behavioural input to these models which can be controlled with public policies, e.g., social distancing measures, and these policies can be effective in controlling transmission dynamics as presented in Figure 3. However, in a case of lacking proper vaccines to immunise the population these measures would only delay the surge of cases and inevitable demand on the healthcare system after the social distancing policies are relaxed. On the other hand, using mass testing campaigns for case identification, contact tracing and isolation can help achieve effective suppression of the epidemic (McClellan et al., 2020). This would also help earlier removal of the social distancing measures. Therefore, these public health policies and mitigation efforts are interrelated as shown in Figure 3. The dashed line between social distancing and public health supply management problems is representing this indirect relation of the interventions. We group the stockpiling of critical resources, public accessibility of testing kits and emergency care delivery under public health supply management problems as the effectiveness of medical mitigation interventions are strongly dependent of efficiency of the solutions provided to these supply management issues in the health care delivery system. Full epidemic control requires an understanding of this interrelationship together with other population-specific characteristics.

Figure 3.

Integrated systems approach on response planning for COVID-19.

This proposed framework recognises these relations. For instance, an emergency preparedness and response plans should have optimised allocation of the stockpiles of medical resources, such as testing kits, vaccines and ventilators (Huang et al., 2017). This would certainly speed up timely case identification, contact tracing, isolation, and would improve the accessibility to effective treatment; thus, ultimately slowing community spread and decrease the burden of the pandemic.

Deployment of resources to effectively collect samples for testing is an important task in the whole system of public health response and it is strongly emphasised in this proposed system. Supply chain optimisation models can help decision-makers allocate resources based on the socio-demographic characteristics and needs of a specific population (Araz et al., 2014; Araz, Galvani et al., 2012; Ramirez‐Nafarrate et al., 2019, 2015). Using drive-through testing sites is a strategy to rapidly and safely collect samples from suspected individuals that may lead to improved accuracy of the information about the spread dynamics of the disease. Having more accurate information would lead to improved decision-making for pandemic containment.

This presented framework can be developed and embedded into a simulation -visualisation architecture which can then be used to test emergency scenarios, design plans and train decision-makers in the planning process (Araz, Jehn et al., 2012). These types of exercises not only improve public health decision-making, but also increase confidence in performing key capabilities required to address a pandemic (Araz & Jehn, 2013).

4. Conclusions

Increasing our understanding of disease spread dynamics during a pandemic requires collection and processing of accurate and timely information. As suggested by the WHO, widespread diagnostic testing is crucial in suppressing the community spread of COVID-19. However, using EDs to undertake mass testing will result in crowded EDs and increase transmission risk to healthcare staff and other ED patients. Rapid deployment of drive-through testing sites presents an effective solution to these issues. Drive-through testing protocols have been used to test for COVID-19 in a limited number of countries and by some healthcare providers in the US. However, it is important to understand the practical and operational implications of drive-through testing sites in order to improve their efficiency for serving communities within a short time period with limited resources. In this paper, we discussed the key insights obtained from our previous experience and contributions to the literature. These insights underline the need to consider the socio-demographic and health characteristics of the affected population for an effective response to the pandemic.

The proposed framework and the systems approach not only helps support decision-making for large-scale sample collection or vaccine distribution operations, but also helps to design pandemic response plans by evaluating interventions and policies. The proposed systematic framework integrates models to dynamically assess evolving situations with public health data and optimise operational decisions for a response. The models and the decision support framework discussed in this paper, along with exercises and visualisations, are useful to gain insights, test scenarios and make informed decisions.

Disclosure statement

No potential conflict of interest was reported by the authors.