Estimating COVID-19 Prevalence in Sri Lanka: A Dynamic Sampling Model Approach

Abstract

Over history, human has had to face various crises and diseases, and among them, the COVID 19 virus stands out as one of the most deathful diseases ever. It has brought numerous challenges to all the fields worldwide. Despite efforts to control its spread, the virus persists globally with varying intensity. Addressing this challenge requires an effective and precise control measure. The progression of the virus in different sub-regions is influenced by factors such as population density, public mobility, and healthcare infrastructure. Consequently, the prevalence of the virus varies across sub-regions. This study proposes an adaptive sampling design that modifies the stratified sampling technique to capture the changing prevalence of COVID-19, considering the dynamic nature of infected populations. This adaptation is essential as the increase of infected cases boosts the virus spread, and the standard sampling techniques do not address such dynamic population conditions in determining the sample size. The study aims to narrow the gap between reported and actual daily infections, providing more accurate estimates of virus distribution. The weighted allocation method incorporates the skewed pattern of coronavirus progression, with weights determined based on the first derivative of reported infected cases. This derivative information is based on the recent dynamics of the infected cases. Thereby larger weights were assigned when the virus progression increased, and smaller weights were assigned when the virus progression decreased. The resulting sample sizes for each sub-region are calculated using the modified stratified sampling method. Further, to illustrate the accuracy of the sampling design, simulated data from different epidemic scenarios, such as community spread, cluster spread, and border spread was used. This simulation allowed us to test the robustness of the techniques for the different states of the virus progression based on the infected cases. The sample size obtained through this dynamic sampling technique exhibits a direct correlation with the fluctuations in the number of infected cases, increasing as the infection cases rise and decreasing as they decline. In conclusion, the study introduces a novel sampling technique that accommodates the dynamic nature of population sizes, and it can be straightforwardly applied for the real-world data as well. Thus, this modified stratified sampling method emerges as a precise approach for capturing the actual prevalence of COVID-19.