To the Editor.—
The conclusion that decision analysis shows that additional knowledge can be harmful is misleading. Decision analysis in the pancreatic cancer example would show that the attractive-sounding test is useless, not harmful. The test appears harmful in the example because it is misused.The statistics reported for the hypothetical test for pancreatic cancer indicate that 80% of patients who have cancer are detected, while positive results are obtained for only 5% of noncancerous patients. The encouraging impression given by these figures is reversed when Bayes theorem is used to compute the probability that a patient with positive results has cancer:P(C/T) is the probability of cancer if the test result is positive, P(C) is the probability of cancer, P(NC) is the probability of no cancer, P(T/C) is the probability of a positive test if cancer exists, and P(T/NC) is the probability that the test is positive if no
Stewart TR. Decision Analysis. JAMA. 1977;237(7):642–643. doi:10.1001/jama.1977.03270340027008
Customize your JAMA Network experience by selecting one or more topics from the list below.
Create a personal account or sign in to: