Can Coronavirus Testing Be Optimized?

As far as I can tell, each test is applied to a single sample from one individual. Would it be possible to combine two samples together and apply a single test to them to determine if one is positive for Coronavirus? If so, in circumstances with low positive rates, fewer tests would need to be used.

For example, in a group with a 1% positive rate, testing 100 people in groups of 2, and then testing each individual again when the group tests positive would essentially double testing capacity.