@KNOWITALL The issue is that you never want to get a test that has no data to support it or could lead to harm. Let’s say you get a test and it finds an adrenal mass. The chance that that adrenal mass is benign is very high and the chance that it’s cancer is very low. But what do you do? Let’s say you decide to have the adrenal mass operated on. You do, and you get a complication of a kidney bleed, leading to hypovolemic shock and a prolonged ICU stay. Was it worth it for your peace of mind? A test should never be done without asking two questions. 1) Is there any data to show that getting this test will improve my life in some way. and 2) What will you do with a result that is bad?
In the case of mammograms, the answer is clear. You do it because a positive test will lead to a possible cancer diagnosis and save your life. But did you ever wonder why we don’t just start mammograms at, say, age 35? It’s because at that age, a positive test is most likely a false positive and leads to an unnecessary surgery.
This goes into Bayesian analysis which is best described as asking the question “What is the prior probability of a positive result?” and then attaching that prior probability to the test.
Here is an xkcd that makes fun of the issue (it may or may not help)
Okay, cancer in blood tests. What, precisely are you testing for? If you test, say, for a blood count, you’re counting blood cells. A potassium level, you’re measuring potassium level. There is no good “cancer” blood test, at least from a screening exam point of view. That’s why you can’t just get a blood test.
hth.