In pandemics or epidemics, public health authorities need to rapidly test a large number of individuals, both to determine the line of treatment as well as assess the spread of infection to plan containment, mitigation and future responses. However, lack of adequate testing kits could be a bottleneck, especially in case of unanticipated new diseases, such as COVID-19, where testing technology, manufacturing capability, distribution, trained manpower and laboratory equipment may be unavailable or be in short supply.
In addition, the cost of test kits might be prohibitive for poorer patients or for governments in low to middle-income countries. This bottleneck can be addressed by examining a test methodology that pools samples from two (or more) patients in a single test (Dorfman, 1943).
The key insight with pooled testing is that a negative result from a pooled sample likely implies negative infection for all individual patients and thereby rules out the need for further tests. This protocol, therefore, requires significantly fewer tests. In the context of diagnosing SARS-CoV–2, pooled testing has been advocated with few or no caveats (Gossner & Gollier, 2020a; Kaul, 2020; Ray, 2020).
Drawing on this analysis, we suggest situations where pooled testing can be an effective strategy for identifying and ruling out infections, permitting economic and social activity without insisting on extreme social distancing.
Pooled testing was first proposed by Dorfman (1943) as a methodology for identifying syphilis infection among US soldiers. Since then, pooled testing has been used to diagnose malaria (Zhou et al., 2014), HIV (Ming Tu et al., 1995) or infertility (Bilder & Tebbs, 2012). The method has also been adopted in other fields such as computer science (Tanenbaum & Wetherall, 2011) and information science (Aldridge et al., 2019).
Sinnott-Armstrong et al. (2020) report results from the first COVID-19 laboratory test using pooled samples. Having tested with three different pool sizes, they report significant improvement upon naive testing in all three cases.
Pooled testing performed better for a particular infection prevalence. Gossner and Gollier (2020b) analyse the value of group testing with a binary search protocol (also see additional protocols proposed by Scarlett  and Hahn-Klimroth and Loick ). In contrast, we propose a simpler protocol while emphasising trade-offs resulting from pooled testing, namely savings in testing kits, increased turnaround time, and increased false negative rates.