Physicians often obtain the same test over and over again. This articl
e first considers screening tests, such as mammographic screening for
breast cancer, and describes several proposals for finding the interva
l that best balances cost, harms, and benefits. The same principles ap
ply to monitoring a patient for recurrence of a disease, such as bladd
er cancer. The article closes with a theoretical consideration of the
circumstances for repeating a test once to check a surprising value.