Mathematical formation and analysis of COVID-19 pool tests strategies

Abstract

Abstract Objectives The excessive spread of the pandemic COVID-19 around the globe has put mankind at risk. The medical infrastructure and resources are frazzled, even for the world's top economies, due to the large COVID-19 infection. To cope up with this situation, countries are exploring the pool test strategies. In this paper, a detailed analysis has been done to explore the efficient pooling strategies. Given a population and the known fact that the percentage of people infected by the virus, the minimum number of tests to identify COVID-19 positive cases from the entire population are found. In this paper, the problem is formulated with an objective to find a minimum number of tests in the worst case where exactly one positive sample is there in a pool which can happen considering the fact that the groups are formed by choosing samples randomly. Therefore, the thrust stress is on minimizing the total number of tests by finding varying pool sizes at different levels (not necessarily same size at all levels), although levels can also be controlled. Methods Initially the problem is formulated as an optimization problem and there is no constraint on the number of levels upto which pooling can be done. Finding an analytical solution of the problem was challenging and thus the approximate solution was obtained and analyzed. Further, it is observed that many times it is pertinent to put a constraint on the number of levels upto which pooling can be done and thus optimizing with such a constraint is also done using genetic algorithm. Results An empirical evaluation on both realistic and synthetic examples is done to show the efficiency of the procedures and for lower values of percentage infection, the total number of tests are very much less than the population size. Further, the findings of this study show that the general COVID-19 pool test gives the better solution for a small infection while as the value of infection becomes significant the single COVID-19 pool test gives better results. Conclusions This paper illustrates the formation and analysis of polling strategies, which can be opted for the better utilization of the resources. Two different pooling strategies are proposed and these strategies yield accurate insight considering the worst case scenario. The analysis finds that the proposed bounds can be efficiently exploited to ascertain the pool testing in view of the COVID-19 infection rate.