An analysis of the value of additional diagnostic procedures.

When several diagnostic procedures are performed on a patient, the probability of his having or not having a specific ailment will change as the result of each procedure is known. This paper describes how such a change is readily calculated using Bayes' theorem. Information theory is also used to find the increment in information from each successive test. Numerical estimates are made using data pertinent to lung X-ray examinations. Finally, some implications for diagnostic examinations in general are drawn.