In estimating the cost-effectiveness of diagnostic procedures, it is helpful to treat diagnostic information as a commodity with a unit price. The amount of useful information provided by a test result can be measured in binary units (bits), and the unit price of the information produced by the test result can be expressed in dollars per bit in much the same way that the price of gold is given in dollars per ounce. This allows comparison of the unit prices of various diagnostic tests, examination of the effect of multiple testing, and calculation of the most cost-effective conditions for screening tests.