Modelling the impact of rapid tests, tracing and distancing in lower-income countries suggest optimal policies varies with rural-urban settings
Jiang X., Gong W., Gao Y., Fraser C., Holmes C.
AbstractBackgroundNon-pharmaceutical intervention (NPI) remains the most reliable COVID-19 containment tool for low- and middle-income countries (LMICs) given the inequality of vaccine distribution and their vulnerable healthcare systems. We aimed to develop mathematical models that capture LMIC demographic characteristics such as young population and large household size and assess NPI effectiveness in rural and urban communities.MethodsWe constructed synthetic populations for rural, non-slum urban and slum settings to capture LMIC demographic characteristics that vary across communities. We integrated age mixing and household structure into contact networks for each community setting and simulated COVID-19 outbreaks within the networks. Using this agent-based model, we evaluate NPIs including testing and isolation, tracing and quarantine, and physical distancing. We explored the optimal containment strategies for rural and urban communities by designing and simulating setting-specific strategies that deploy rapid diagnostic test, symptom screening, contact tracing and physical distancing. We performed extensive simulations to capture the uncertainty of outbreak trajectories and the impact of varying model parameters.FindingsWe found the impact of testing, tracing and distancing varies with rural-urban settings. In rural communities, we found implementing either high quality (sensitivity > 50%) antigen rapid diagnostic tests or moderate physical distancing could contain the transmission. Additionally, antibody rapid diagnostic tests and symptom-based diagnosis could be useful for mitigating the transmission. In non-slum urban communities, we demonstrated that both physical distancing and case finding are essential for containing COVID-19 (average infection rate < 10%). In slum communities, we found physical distancing is less effective compared to rural and non-slum urban communities. Lastly, for all communities considered, we demonstrated contact quarantine is essential for effective containment and is effective at a low compliance rate (30%).InterpretationOur findings could guide setting-specific strategy design for different communities in LMICs. Our assessments also have implications on applying rapid diagnostic tests and symptom-based diagnosis for case finding, tracing and distancing in lower-income communities.