Optimizing the diagnostic capacity for COVID-19 PCR testing for low resource and high demand settings: The development of information-dependent pooling protocol

Aim: To compare different pooling methods in an attempt to improve the COVID-19 PCR diagnostic capacities. Method: We developed a novel information-dependent pooling protocol (indept), based on transmission of less informative sequential pools on to the next pooling cycle to maximize savings. We then compared it to the halving, generalized halving, splitting and hypercube protocols in a simulation study, across variety of scenarios. Results: All five methods yielded various amount of test savings, which mostly depended on the virus prevalence in the population. In situations of low prevalence (up to 5%), indept had the best performance, requiring on average 20% of tests needed for singular testing across scenarios that were analyzed. Nevertheless, this comes at the expense of speed, with the worst-case scenario of indept protocol requiring up to twice the time needed to test the same number of samples in comparison to the hypercube protocol. In order to offset this, we developed a faster version of the protocol (indeptSp), which minimizes the number of terminal pools and manages to retain savings compared to other protocols, despite marginally longer processing times. Conclusion: The increasing demand for more testing globally can benefit from application of pooling, especially in resource-restrained situations of the low- and middle-income countries or situations of high testing demand. Singular testing in situations of low prevalence should be systematically discouraged.