Summary
How should we best conduct pool testing to keep an eye on the next pandemic while making best use of expensive tests?
When testing people for a disease, we often use individual tests, where each person gets their own test and their own result. An alternative strategy is “pooled testing” where samples are mixed together before testing: a negative test means every individual in the pool are disease-free, but a positive test means at least one of individuals has the disease and further investigation is required.
What is the optimal way to design such pooled tests? This mathematical question can be addressed in a number of ways, using ideas from probability, combinatorics, information theory, statistics, or computer science. There are purely theoretical questions, questions that require the running of computer simulations, and many questions that require a mixture of the two.
For example: What if tests that should be negative are contaminated? What if tests that should be positive are accidentally diluted? How do we deal with correlation, where people in the same family or workplace are likely to infect each other?
