Pooled testing for COVID-19 can maintain the accuracy of test results while reducing the use of testing resources by as much as 76%, according to two new studies of nasopharyngeal test samples.
Widespread testing is critical to manage the ongoing COVID-19 pandemic, as it allows authorities to trace the spread of SARS-CoV-2 and implement effective mitigation measures. However, testing large populations with individual tests is resource-intensive, and breakdowns in supply chains and limitations on testing speeds have severely held back widespread testing worldwide, especially in resource-constrained regions.
One solution involves pooled testing, where researchers combine individual samples into large batches before testing each batch for SARS-CoV-2. If a batch turns up negative, then the individual samples are also assumed to be negative; if it’s positive, then researchers go back and test the samples further. Pooled testing allows scientists to test more people with the same resources, but concerns remain about whether current pooled protocols might be less sensitive than individual tests.
Brian Cleary and colleagues harnessed mathematical models and data on viral loads and dynamics to assess the sensitivity of pooled testing designs, depending on variations in factors such as the population’s SARS-CoV-2 prevalence and at what time during an individual’s course of infection a nasal swab is taken.
They estimated that with the same budget, even simple pooling designs could accurately gauge the prevalence of COVID-19 and identified as many as 20 times more positive cases than individual tests. The scientists also validated their projections with 2,304 nasopharyngeal samples, finding that only 48 pooled tests accurately estimated the prevalence of COVID-19 infections at 1%.
In the second study, Netta Barak and colleagues tested 133,816 nasopharyngeal samples in pools and 121,929 individual samples collected in and near Jerusalem, Israel between March and mid-September of 2020. The team discovered that the pooled strategy outperformed their theoretical predictions and reduced the use of PCR tests and RNA extraction by 76%, remaining efficient even as the prevalence of infections rose from less than 1% to 6% during the testing period.
Furthermore, the pooled results showed no significant decline in sensitivity compared with individual tests. Barak et al. further improved the efficiency of the pooled approach by pre-grouping samples from family members and roommates who may be more likely to test positive for SARS-CoV-2 jointly, highlighting the utility of leveraging pre-existing knowledge about incoming samples when assessing testing protocols.