When testing for infections, the standard method is to test each subject individually. If testing methodology is such that samples from multiple subjects can be efficiently combined and tested at once, yielding a positive results if any one subject in the subgroup is positive, then one can often identify the infected sub-population with a considerably lower number of tests compared to the number of test subjects. We present two such methods that allow an increase in testing efficiency (in terms of total number of test performed) by a factor of $approx$ 10 if population infection rate is $10^{-2}$ and a factor of $approx$50 when it is $10^{-3}$. Such methods could be useful when testing large fractions of the total population, as will be perhaps required during the current coronavirus pandemic.
Download