Pooled testing is basically grouping samples together and administering a single test to each group. The idea is that only those groups that test positive undergo a second round of testing of each individual sample, and the test is designed to be sensitive enough to come up positive whenever at least one sample in the group is positive - thus using less tests which both reduces the costs and increases our testing capacity.
The above one minute long video gives a clear idea on how this testing works and how it would be easy and effective to use for safe and swift reopening of institutions around the world.
Pooled testing has been implemented for Covid-19 in a few places: the Hadassah Medical Center in Jerusalem (group size 8), the Nebraska Public Health Laboratory (group size 8), and most recently in Wuhan (group size 5-10).
A related idea of Pooled Screening is the right tool for widespread and affordable early detection systems. Pooled screening is the same process as Pooled Testing, except using larger group sizes. The trade-off for the larger group size is a loss in sensitivity-- a sample with sufficiently low viral load will be missed by the group screen. Whereas, the results of a negative pooled test will be reported as a negative test for the individuals in the group, the results of a negative pooled screen are not.
At first glance, this trade-off is not helpful, however for early detection, the added group size allows for both broader and more frequent testing: the sensitivity of the individual test goes down but the sensitivity of the early detection system goes up.
Members of an early detection surveillance system would be screened twice a week in a pooling screen of size 16. A self-sampling system would be used (either shallow nasal swab or saliva based). Assuming the sensitivity of the size 16 pooling screen being at least 70%, this would result in an early detection scheme that:
Detects each contagious person within 3 days (on average) (see here for analysis)
Costs under 5 dollars a day per person
Uses 1 test per day for every 48 people
Implementing an early detection system broadly would provide a cornerstone for controlling the spread of COVID-19 throughout the country. It is an inexpensive model (under $5 a day per person) using small amounts of testing resources (1 test a day for every 56 people) that could be replicated all over the country. Local labs capable of running 1000 tests a day could be providing ongoing surveillance for 56,000 local community members. It could provide early detection in places like nursing homes, jails, factories, and schools.